Introduction to WiMAX Technology: An Overview

Introduction to WiMAX Technology
Khalid Sheikh
January 2009
Presentation Overview
Introduction Overview of other wireless technologies Radio Architectures OFDM Basic OFDMA Basic WiMAX Architecture
PHY, MAC, Control & Management
System Performance WiMAX Validation WiMAX Interoperability WiMAX Network Summary Q&A
Page 2
Wireless Systems
Wireless communication has been around for over 100 years
Pioneered by Thomas Edison who did not take serious interest in this technology and sold his patent to Marconi for a single song Most non-broadcast typed applications geared to P2P & PMP Generally standalone operation based on proprietary architecture Incompatible over the air interface
Many lacks interoperability with other equipment Most failed to provide high data rate and mobility Some commonly used recent LAN applications include
Infra-Red, IR Bluetooth (WPAN) Mobile Phone WiFi (LAN) WiMAX (MAN)
Page 3
A Case, why we need it!

Wide use of internet resulting in an increased demand for convenient internet access and high speed data access. Increase demanded by new applications such as streaming video, on-line gaming, on-demand movie distribution, VoIP, video teleconferencing, telemedicine, serveillance & monitoring, etc. Fixed wireless offers several advantages over traditional wired solutions. These advantages include lower entry and deployment cost; faster and easier deployment and revenue realization; ability to build out the network as needed; lower operational costs for network maintenance, management, and operation; and independence from the incumbent carriers
Page 4
WPAN, Bluetooth
Wireless personal area networks based on IEEE802.15.1 FHSS operation using TDD Data rate Sync., connection oriented, 64 kbps Async., 433.9 kbps symmetric Async., 723.2 / 57.6 kbps asymmetric, 1 Mbps aggregate bit rate Ver2 to increase up to 3 Mbps Three PO classes Class 1, 1 to 100 mW, for 100 m range Class 2, 0.25 to 2.5 mW, for 10 m range Class 3, up to 1 mW, for 1 m range Shares data among up to 8 Bluetooth enabled devices
One of 79 channels in 2.402-2.480 GHz ISM-band 100 bytes long packet length BPSK for Ver1, DQPSK & 8-DPSK for Ver2 GFSK with mod index of h = 0.28 Limited QoS Guarantees, ARQ/FEC Connection setup time
Depends on power mode 2.56s max, 0.64s average

Provides high level protocol support Security
Challenge/response (SAFER+) hopping sequence
Page 5
Bluetooth Protocol Stack

Middleware Layer to support:
L2CAP, logical link control and adaptation protocol SDP, service discovery protocol
Audio Apps Management Apps Phone Apps Networking Apps vCard Application Layer
AT SDP Control Audio L2CAP TCS
BNEP RF Comm
OBEX
Middleware Layer
HCI, Host Control Interface Link Manager Baseband Radio

Required
Transport Layer
Optional
Not part of Bluetooth standard

Page 6
WiFi, Wireless Fidelity

Low cost & very popular due to standardized architecture & mutual IOP Short distance coverage, 100 m in open area No QoS offered, best effort only Operation in License Exempt Band IEEE802.11a/b/g/n, Standard first adopted in 1997 Delivers services previously found in wired networks Relatively high throughput Highly reliable against interference by applying fragmentation technique Continuous connection Every station reacts to every frame it receives Requires participation of all stations Low power operation to prolong battery life Authentication services Architecture includes IR, FHSS, DSSS, OFDM
Introduction to WiMAX Technology Page 7
IR, Infra-Red
Uses a near visible light as a transmission media
Typically Line of Sight operation or reflected from object Restricted to indoor applications PSDU
Symbol Mapping 16-PPM 4-PPM Modulator LED Driver
Can not pass through walls

Data rate 1-2 Mbps
Uses 4-PPM for 2 Mbps
PSDU
Uses 16-PPM modulation for 1 Mbps
Symbol De-mapping
16-PPM 4-PPM Demodulator
LED Detector
IR PMD XCVR
PPM is a modulation technique that keeps the amplitude and pulse width constant and varies the position of the pulse in time. Each position represents a different symbol in time.
Operates at base band Inexpensive system based on IEEE802.15.1 Short range, 1m

FHSS, Frequency Hop Spread Spectrum

Data rate 1-2 Mbps, TDD (2-level GFSK for 1 Mbps, 4-level GFSK for 2 Mbps) A set of hop sequence defined in 802.11 Channels are evenly spaced across a span of 83.5 MHz 78 (75 min) frequencies, each occupying 1 MHz BW channel (total of 79 Ch.) Hopping at least every 400 ms, then resync before resuming data transmission Predetermined pseudo random pattern Used in licensed exempt 2.4 GHz band (2.4 to 2.4835 GHz) Hopping sequence: 3 set of 26 channels (min hop distance of 6 Ch.) Interference tolerant Echo resistance Simpler to install than the DSSS Less expensive than the DSSS More vendors/selections than the DSSS (not true anymore) No spreading gain, No SNR improvement
Page 9
FHSS, Frequency Hop Spread Spectrum

80
Frequency Slots
60
1 MHz FHSS Packets
Time
D E B A C
2.40 2.41 2.42 2.43 2.44 2.45
40
DSSS Packet
20
DSSS Packet
DSSS Packet
22 MHz
0 1 2 3 4 5 6
Time
7
Freq (GHz)
Hopping Pattern: C A B E D
DSSS, Discrete Sequence Spread Spectrum

Regulated per IEEE802.11b
1 or 2 Mbps using 11-bit Barker spreading code
Spreading yields processing gain at receiver
Requires channel linearity over 11 MHz
Operates in license exempt 2.4 GHz ISM band (2.4 to 2.4835 GHz)
3 non-interfering 25 MHz apart channels Extremely crowded band
Page 11
DSSS, (1)
Fairly effective at low data rate
Bandwidth requirement becomes too large at higher data rate Not practical to implement due to increased cost, power, size & technical difficulties
Processing gain = 10*Log (chip rate / bit rate) = 10.4 dB for 11 chips
Feasible to achieve negative SNR at lower modulation in equivalent BW
1 bit
Original Data Barker Sequence Spread Data
1 chip
Narrowband Signal Interfere Noisefloor SNR
Page 12
DSSS, (2)
Frequency, BW & PO are regulated worldwide
PO of 100 mW nominal
Data rate up to 11 Mbps

operates at 11 Mbps and falls back to 1/2 Mbps as the legacy 802.11
DBPSK & DQPSK for 1&2 Mbps, CCK for 5.5 & 11 Mbps with enhanced 802.11b
Interference tolerant
Upgradeable to higher speed while operating in 2.4 GHz
Page 13
DSSS Spectrum
Regulated spectral mask
Signal occupies in about 20 MHz BW regardless of data rate (1, 2, 5.5, or
11 Mbps)
Spectral shape of the channel represents sin(x)/x function Spectral products to be filtered to -30 dBr from central frequency and all other 0 dBr products to be filtered to -50 dBr
-30 dBr -50 dBr
fc - 11MHz fc - 22MHz
fc
fc + 11MHz fc + 22MHz Minimum Channel spacing between center frequency
Transmit Channel Shape
25 MHz
25 MHz
2.400 GHz
2.412 GHz (channel 1)
2.483 GHz
Page 14
DSSS & FHSS Implementations
Page 15
DSSS vs. FHSS, Summary

DSSS
Short latency time Constant proc gain = a better SNR Quick lock-in as radios synchronize No dwell time No re-sync with other radio necessary Short indoor range Long outdoor range (40 km in LoS) Greater overall data throughput Noise immunity (high)
FHSS
Long latency time No processing gain Slow lock-in, must search a channel 400 ms dwell time Must re-sync with other radio after every hop Short indoor range Short outdoor range (10 km in LoS) Lower overall data throughput Noise immunity (low)
Multipath immunity (good)
Multipath immunity (none)
Page 16
802.11a, (1)
LAN standard revised and released in 1999
Multicarrier OFDM system Similar to ETSI Hiperlan-2, main difference resides in the convolution encoding method
Operates in 5 GHz UNII license-exempt band

Three 100 MHz bands in ANSI operation PO restricted per operating band 40 mW in 5.15-5.25 MHz, 200 mW in 5.25-5.35 MHz & 800 mW in 5.7255.825 MHz Antenna gain restricted to 6 dBi max The centers of the outmost channels shall be at a distance 30 MHz from bands edges for the lower and middle bands, 20 MHz for the upper bands Channel frequency numbers are defined by 802.11a Minimum sensitivity -82 to -65 dBm depending on the chosen data rate & Mod PER rate to be less than 10% at a physical sub-layer service data units of length 1000 bytes
802.11a, (2)
BPSK to 64 QAM modulation Different FEC rates, double encoding (inner & outer) & block interleaving Raw data rate up to 54 Mbps Multi-path fade tolerant Intended for short range, 100 m Best effort service (no QoS) Sync using a fixed training sequence lasting less than 16 us Data rate, FEC & Mod throttle up / down based on path conditions Encrypted security with enable/disable option
Page 18
802.11a, Receiver Performance Requirements

Required minimum threshold & tolerable interference level NF dependency per frequency band FEC & Mod dependency
Data rate (Mbps) Sensitivity (dBm), 1e- 6 Adjacent Channel Rejection (dB), 1dB Alternate Adjacent Channel Rej. (dB), 1dB
6 9 12 18 24 36 48
-82 -81 -79 -77 -74 -70 -66
16 15 13 11 8 4 0
32 31 29 27 24 20 16
54
-65
-1
15
Page 19
802.11a, Example Symbol

OFDM using 64-points IFFT/FFT
Data subcarriers: 48 Pilot subcarriers: 4 DC subcarrier: 1 Null subcarriers: (6+5) Subcarrier spacing: 20M/64 = 312.5 kHz Per channel spacing: 3.2 us Guard band: 0.8 us Symbol spacing: 3.2 + 0.8 = 4 us Symbol rate: 250 kHz Bit rate at 64QAM-3/4: Symbol * FEC * Active-Subcarriers * Symbol-rate = 6 * (3/4) * 48 * 250k = 54 Mbps
20 MHz BW
Page 20
802.11a, Example Frame

Frame format: PSDU PHY sub-layer service data unit, MPDU MAC protocol data unit, PLCP physical layer convergence procedure, PPDU preamble and header to create the PLCP protocol data unit.
Preamble (header) is always BPSK-1/2 for ruggedness
PLCP Header Rate 4 bits Reserved 1 bit Length 12 bits Parity 1 bit Tail 6 bits Service 16 bits PSDU Tail 6 bits Pad bits
Coded OFDM (rate is indicated in signal) PLCP Preamble 12 symbols Signal One OFDM symbol Data Variable number of OFDM symb
PPDU Frame Format

802.11g
Covers both 802.11a and 802.11b standards

Same MAC layers for all variants (802.11b, a & g)
Adds 802.11a equivalent operation in 2.4 GHz band

Higher data rate (54 Mbps) than the 802.11b OFDM based transmission
Page 22
802.11n
Uses either 20 or 40 MHz channel
Transmit diversity with multiple data streams Increased use of MIMO operation (SM, STBC, Tx Beam Forming) Increased throughput to 100 Mbps per stream (600 Mbps with all options)
Enhanced QoS & FEC

Selectable CP delays (400n or 800n) Backward compatibility for 802.11a/b/g
Increased data sub-carriers (48 to 52 in 20 MHz, double for 40 MHz)

LTE High Level Requirements

Standardization effort for LTE was launched in Nov 2004 Expected to complete in Oct,09 Peak data rate 100 Mbps in 20 M BW in the DL (with 2x2 MIMO) 50 Mbps in 20 M BW in the UL (without MIMO) Control plane latency Transition time from idle to active state 100ms Transition time from dormant to active state 50ms User plane latency Measured from UE to edge of RAN (one way) Shall be less than 5 ms for single user for small IP packet Control Plane Capacity At least 200 active voice calls / cell / 5 MHz Mobility Optimum performance at low speeds from 0 to 15 km/hr High performance at higher speeds from 15 to 120 km/hr Spectrum flexibility 1.25, 1.6, 2.5, 5, 10,15 and 20 MHz All IP network
All services in the packet switched domain No circuit switched domains

WLAN Standards, Perspective
Page 25
Super Heterodyne Architecture
Page 26
Zero-IF Architecture
Page 27
Radio Architectures
Direct Conversion
Advantages No off-chip IF filter Single synthesizer
Super-heterodyne
Advantages Low LO leakage Wide LO pulling range
Cheap
Low power consumption No image signal Disadvantages LO leakage LO pulling range High freq low PN requirements I/Q mismatch Quadrature LORF DC Offset
No quadrature LO
Design flexibility Superior I & Q matching at IF High performance Disadvantages Off-chip IF filter Two synthesizers Low integration High IF-RF separation to avoid Image signal
Page 28
SDR, Software Defined Radio

Advantages (with direct conversion)
S/W configurable in the field for specific conditions
Disadvantages
DC offset issue due to RF to LO isolation or 2nd order non-linearity Appropriate mixer design high IP2
S/W upgrade in the field

Reduced parts count Reduced die/board space Lower power consumption
2nd order nonlinearity in the LNA generates low frequency beat

Static (slow varying) DC offset due to LO to RF isolation DC offset cancellation at each received packet LO pulling VCO at multiple or sub-multiple of LO Fast synthesizer response LO emission LO leakage to mixer input or antenna
Lower over all cost

Simpler assembly Single integrated synthesizer Almost spurious free
Reduced RF filtering
Page 29
Disadvantages of Direct Conversion, (contid)
Gain control to move to RF section

Higher cost to pay for lower phase noise higher frequency LO
No possibility to reduce in-band noise & spurious generated by signal chain Suffers LO and side band leakage when DAC Synth a low IF Flicker noise
Low frequency noise in all active devices WLAN has a large modulation bandwidth, no energy at low frequency
Phase and gain mismatch the IQ-symbol

Difficult to correct imperfect quadrature error
Page 30
LAN, 3G & others vs. MAN, (1)

MAN
BS connected to public networks BS serves subscriber stations BS and stationary/mobile SS SS typically serves a building (business or residence) & mobile Provide SS with first mile access to public networks Multiple services (voice, data & multimedia) with different QoS priority simultaneously Robust security Many more users Much higher data rate Much longer distance Selectable bandwidth (1.25- 20 MHz) & data rate Adaptive modulation & coding Advanced antenna techniques TDD & FDD, symmetric & asymmetric rate Spectrally efficient Lower cost than 3G solution Link layer retransmission OFDMA for mobility, freq & multiuser diversity Support for fixed SC, MC & mobility IP based architecture
Page 31
LAN
Already covered under 802.11
Page 32

3G:
Fixed operating bandwidth Very difficult & expensive to spread higher data rate using CDMA Established infrastructures in process of upgrading GSM using UMTS and or HSDPA
DL only, 14.4 Mbps in 5M BW using 15 codes (specified but a challenging task) DL only, 3.6 Mbps in 5M BW using 5 codes but typically averages about 250 kbps DL only, 7.2 Mbps in 5M BW using 10 codes but typically averages about 750 kbps UL specified to support 384 kbps but typically averages about 40 to 100 kbps
CDMA using 1x EV-DO

Rev A DL, specified to support 2.4 Mbps in 1.25 M BW but typically 100 to 300 kbps Rev A UL, specified to support 1.8 Mbps in 1.25 M BW Rev A to provide low latency of 30 ms, VoIP, video, QoS, fast handoffs and broadcast applications Rev B is specified to support 73 Mbps DL, 27 Mbps UL in 20 M BW
Others
IEEE820.20
Standard under development For mobility up to 250 kmph 3.5 G band 4 Mbps DL, 1.2 Mbps UL
IEEE820.22
Standard under development Broadband access targeted for far reaching rural area
Using unused VHF & UHF bands
Page 34
WiFi vs. WiMAX

Parameter Licensed Band Operation AGC Range (dB), 64QAM OFDMA Advanced antenna Mesh antenna Power Output (dBm) 802.11 No N/A No, yes with 803.2a None Can introduce mesh topology, but not supported by Restricted in unlicensed bands Optimization centers around PHY and MAC layer for 100m range. Range can be extended by cranking up the power - but 10 31 N/A WiMAX Yes 50 Yes Standard supports advanced antenna technique Std supports mesh network topology 43, per local regulatory requirements PHY and MAC designed with multi-km (40) range in mind. Standard MAC 7 31 30 Channel bandwidths can be chosen by operator (e.g. for sectorization). 1.25 MHz to 20 MHz width channels. MAC designed for scalability independent of channel BW 10, 20 ~2.7 ~4.5 Designed to handle many users spread out over distance No "near-far" compensation Optimized for indoor non-line-of-sight (NLoS) Designed to tolerate greater multipath delay spread performance (signal reflection). Optimized for outdoor NLoS. MAC designed to support(CSMA/CA) no guaranteed MAC designed to support thousands of users Contention - based MAC 10s of users QoS Standard cannot currently gaurantee latency for Voice, Grant-request MAC Video Designed to support Voice and Video from ground up Standard does not allow for differentiated levels of Supports differentiated service levels: e.g. T1 for service on a per user basis business customers; best effort for residential TDD only - asymmetric (proposed) QoS is prioritization TDD/ FDD/HFDD symmetric or asymmetric No QoS today. 802.11e only Centrally enforced QoS Convolution RS & CC Existing standards is WEP. 802.11i in process of addressing security RSA (1024 bits) Optional RC4 Mandatory, triple-DES (128 bits)
Range NF, dB SNDR (dBc), TX Alternate channel Rej
Channel BW (MHz) Maximum bps/Hz User handling Multipath MAC capability MAC operation Latency Services Transmission QoS FEC Security Encryption
Page 35
Technology Throughput Comparison
Page 36
Spectral Efficiency Comparison
Page 37
Point to Multipoint WiMAX Applications
Page 38
WiMAX, (1)
Stands for Worldwide Interoperability for Microwave Access A cost effective alternative to wireline services especially in the developing countries where no existing wireline services available Operation in licensed & licensed exempt bands using 1.25-20 MHz BW QoS, advanced security & higher throughput than WiFi Supports QoS, VoIP, video distribution, on-line gaming & real time video conferencing Standards and interoperability is the key to its success GPS and IEEE 1588 over IP to synchronize the network from a master clock GPS may be more expensive and difficult to access open sky if in the basement IEEE 1588 requires a master source access in the network WiMAX network is entirely IP and there is no option of recovering timing signal (without embedded mechanism) as there is with the TDM application Dynamic frequency selection for operation in unlicensed bands Longer wavelength makes multipath more significant LOS not feasible in residential applications There may be cost associated with outdoor mounted antenna
Uses a very versatile configurable modulation schemes that adds complexity

WiMAX, (2)
Why we need it? Demand created by internet & mobile usage. Broadband access to residential, SOHO, SME, backhauling hotspot, long wait time for increased T1 services. Lack of land lines in some countries. Mobility to offer in two stages (portable-nomadic and seamless mobility)
Higher data rate Multiple levels of guaranteed QoS Stationary & Mobility Multipath tolerant by using multiple lower frequency carriers Switchable mode Concatenated FEC Low latency Security Ease of installation Lower cost deployment and operating solution Large system gain (about 150 dB) and coverage range
WiMAX, (3)
Mobility under 120 km/hr (target applications are handset, laptop). Standard released in Nov05.
Support for both LOS & NLOS
It is not mandated by standard but TDD is most likely mode of choice for mobile applications because it divides the entire frequency spectrum into upstream and down stream time slots (more efficient use of limited frequency)
Uses all IP backbone OFDMA-PHY with sub-channelization allows time & frequency resources to be dynamically allocated among multiple users across DL & UL subframe
WiMAX, (4)
Broadcast and Multicast support, low latency 100 ms, low to zero packet loss during handovers at speed 120 km/Hr or higher Simultaneous support of real time multimedia and isochronous applications like VoIP Simple self installed user station (SS/MS)
Automated management of IP connection with session persistence Automatic reestablishment following transitions between access points
Likely applications: single carrier for back haul, OFDM for fixed access in up to 28 MHz BW, scalable OFDMA is most versatile and preferred for mobile operation in 1.25 to 20 MHz BW Frequency inaccuracy of 1e-6 max for FDD and TDD Time accuracy: N/A for FDD but 5-25 us for TDD
Other Systems
Non-standard i-Burst from ArrayComm and FlashOFDM from Qualcomm

3G (UMTS, HSDPA by GSM operators, EV-DO by CDMA), WiFi (higher data rate than 3G due to 20 M BW, using inefficient CSMA protocol-carrier sense multiple access) WiFi standard 802.11n is being enhanced to support 100 Mbps, better QoS, transmit diversity and other enhanced techniques
Page 43
WiMAX Standards, Timeline

Evolving standard, work started under 802.16 in 1999 802.16a, Jan03 802.16d, July04, replaced 802.16, a & c. 895 pages 802.16e, MAC function to support higher layer handoff in under 6 GHz band, Dec05. 864 pages 802.16f, fixed WiMAX management information base. Added multi-hop functionality 802.16g, management procedures and interfaces for fixed and mobile 802.16 systems. Addresses efficient handover and further improves the QoS support. 802.16h, mechanisms, policies and MAC enhancements for coexistence in licensed exempt bands 802.16i, (with drawned). Mobile WiMAX management information base, merging into 802.16-2008 802.16j, multi-hop operation 802.16k, bridging amendment 802.16m, advanced air interface for next generations, higher data rate and higher speeds
Page 44
802.16s Comparison, (1)

Standards 802.16 Completed Dec01 Alignment Mode LOS only Spectrum 10-66 GHz 802.16a Jan03 LOS & NLOS 2-11 GHz, licensed & license exempt Up to 134 Mbps Up to 75 Mbps 28 MHz 20 MHz Single carrier Single carrier, only 256 OFDM or 2048 OFDM sub-carriers 802.16d June04 LOS & NLOS 2-11GHz 802.16e Dec05 NLOS 2-11GHz for fixed, 2-6GHz for mobility Up to 75 Mbps Up to 15 Mbps 20 MHz 5* MHz Single carrier, Single carrier, 256 OFDM or 256 OFDM or 2048 OFDM scalable sub-carriers OFDMA (128, 512, 1024, 2048 sub-carriers)
Bit Rate Bandwidth Modulation
Page 45
802.16s Comparison, (2)

Standards Modulation Mobility Bandwidth 802.16 4/16/64 QAM Fixed 802.16a 4/16/64 QAM, 256Q optional Fixed/Portable 802.16d 802.16e 4/16/64 QAM, 4/16/64 QAM, 256Q optional 256Q optional Fixed/Nomadic Mobile/Portable 1.25-25 MHz
20, 25, 28 MHz 1.25-20 MHz
MAC Architecture Applications
1.25-20 MHz, uplink to conserve Po PMP mesh, PMP mesh, PMP mesh, PMP mesh, TDD and FDD TDD and FDD TDD and FDD TDD and FDD E1/T1 services, E1/T1 services, Indoor Portable backhauling backhauling broadband broadband hot spots hot spots. access for access for Wireless DSL residential consumers. users (HSpeed Mobile internet. internet, VoIP) Always best connected
Page 46
WiMAX & Others

Parameter Standards FTT size User carriers Pilot carriers MIMO Guard period Multiple users over frequency (@1 symb time) Multiple users over time (@1 channel) Fixed WiMAX IEEE 802.16d 256, 2048 1680/1728 166/192 Yes 1/4, 1/8, 1/16, 1/32 Mobile WiMAX IEEE 802.16e 2048, 1024, 512, 128 various various Yes 1/4, 1/8, 1/16, 1/32 HSPA 3GPP, Release 6 N/A N/A N/A No N/A 1xEV-DO Rev A 3GPP2 N/A N/A N/A No N/A WiBRO IEEE 802.16e 1024 864/840 96 Yes 1/4,1/8,1/16,1/32 Wi-Fi 802.11a,b,g 64 52 4 No 1/4
Yes
Yes
No
No
Yes
No
Yes
Yes
No
No
Yes
No ~2.7 bps/Hz peak data rate, 54Mbps shared in 20MHz using 802.11a/g more than 100Mbps peak layer 2 throughput using 802.11n
Peak DL data rate
~4.5bps/Hz, 9.4Mbps in 3.5MHz with 3:1 DL to UL ratio TDD, 6.1 Mbps with 1:1 3.3Mbps in 3.5MHz with 3:1 DL to UL ratio TDD, 6.5 Mbps with 1:1
~4.5 bps/Hz, 46Mbps with 3:1 DL to UL ratio TDD, 32 Mbps with 1:1 7Mbps in 10MHz with 3:1 DL to UL ratio, 4 Mbps with 1:1
14.4 Mbps using all 15 codes, 7.2 Mbps with 10 codes
~4.5 bps/Hz, 38 Mbps with 3:1 3.1Mbps, Rev B DL to UL ratio will support TDD, 27 Mbps 4.9Mbps with 1:1
Peak UL data rate
1.4Mbps initially, 5.8Mbps later
1.8Mbps
5.9 Mbps with 3:1 DL to UL ratio, 3.4 Mbps with 1:1 same as above
Page 47
WiMAX & Others (contid)

1xEV-DO Rev A
Parameter
Bandwidth
Fixed WiMAX Mobile WiMAX 3.5MHz and 7MHz in 3.5GHz band, 3.5M, 7M, 5M, 10MHz in 5.8GHz 10M & 8.75M band, 28M max initially, 28M max QPSK, 16QAM, 64QAM TDM TDD, FDD 3.5G and 5.8G initially 3-5 miles Not applicable QPSK, 16QAM, 64QAM TDM/OFDMA TDD initially 2.3G, 2.5G &3.5G initially
HSPA
WiBRO
Wi-Fi
5M
1.25 x 2 M QPSK, 8PSK,16QAM TDM/CDMA FDD 800/900/1800 /1900M 1-3 miles High
8.75M QPSK, 16QAM, 64QAM TDM/OFDMA TDD 2.3G
Modulation Multiplexing Duplexing Frequency Coverage (typical) Mobility
QPSK, 16QAM TDM/CDMA FDD 800/900/1800 /1900/2100M 1-3 mile High
20M for 802.11 a/g, 20/40M for 802.11a, 11M for 11b BPSK, QPSK, 16QAM, 64QAM CSMA TDD
QoS
<2 miles Mid QoS designed in QoS designed in for for voice/video, voice/video, differentiated differentiated services. services. Grant Grant request MAC request MAC
DL only
DL only
2.4G, 5G <100ft indoor 1-3 miles <1000ft outdoor High Low QoS designed in No QoS for support.802.11e voice/video, working to differentiated standardize. services. Grant Contention based request MAC MAC
Page 48
OFDM, Orthogonal Freq Division Multiplexing

A combination of modulation and multiplexing technique Mapping of information on changing in the carrier phase, freq., amplitude or a combination Method of sharing bandwidth with other data channels Channel bandwidth divided by a number of sub-channels Aggregate data rate throughput is about the same but data rate on each sub-carrier is much lower Longer symbol time practically eliminates the effects of variable time delays Integer number of cycles to complete for each sub-channel OFDM bit rate is based on number of active data sub-carriers, not the bandwidth Orthogonality allows simultaneous transmission on many subcarriers in tight frequency space without interfering each others
OFDM, Interference Response, Example 1
Single carrier like water flowing from a faucet Multi-carrier like water flowing from a shower head
Page 50
OFDM, Interference Response, Example 2
Reliable delivery mechanism
Page 51
OFDM Basic-1
Converts single bit stream (wider bit rate) into multiple (smaller bit rate) parallel bit streams Efficient BW (No BPF between sub-channels) usage Orthogonal approach using FFT technique (50 yrs old, used to be expensive to implement). Signal orthogonality happens in frequency domain (peak of one signal at zeros of all others) Time & freq domain representation
Ch1 Ch2 Ch3 Ch4 Ch5
Power
FDM-Frequency
Ch2 Ch1 Ch3 Ch4 Ch5
Power
Bandwidth saving
OFDM-Frequency
Signal in Time & Frequency Domain

Subcarriers FFT
Ts TFFT Guard Intervals Tg
Frequency
Symbols
Time
OFDM Basic-2
An area under a complete sine/cosine wave is always zero
When multiplied by another integer or non-harmonics, the result is zero
It is non-zero only when multiplied by the same harmonics

Integral 0 to T of sin2(ft)*sin2(2f)t dt=0 where T is multiple of 1/f
FDM must apply a bank of RRC filters

OFDM uses optimized channel spacing and RRC does not buy you much
A proven technology, already being used in cable modem, WiFi, DSL, DVB and DAB
OFDM Basic-3
f1 Mod 1 Mod 2 Rs S/P Mod N fN Rs/N f2
Time-frequency Grid
Data
Frequency
Subcarrier f0
T=1/f0
One OFDM Symbol
Time
Time Wavefom
Carrier Freq using N point FFT
Page 55
OFDM Basic-4
Treats source Symbol as frequency domain and converts it into time domain with IFFT N carriers results in N orthogonal-sinewaves N tones are generated digitally to avoid bank of phase locked oscillators Each N determines a complex Amplitude & phase for that sub-carrier Sin(x)/x spectra for unfiltered sub-channels but their orthogonality prevents interfering to one another Output of IFFT is sum of all sub-carriers and makes up a single OFDM symbol (whose length = NT, T is IFFT input sampling period) in time domain
Data In
Mod M-QAM
IFFT
DAC
BB OFDM
Data Out
DMD M-QAM
FFT
A/D
BB OFDM
Page 56
Example of an OFDM Parameters

FFT/IFFT process must use 2^N tones (use zero filling for non-used tones)
Each sub-carrier produces its sin(x)/x spectra Relatively simple DSP algorithms 256 subcarriers (192 data + 8 pilot + 28 null at start + 27 null at end + 1 DC), center subcarrier is not used due to being easily susceptible to RF carrier feedthrough Constant number of subcarriers regardless of BW. Advantage at Narrow BW. The subcarriers are variable for S-OFDMA. IFFT Symbol (active subcarriers) vs. OFDM Symbol (IFFT plus the gap) Configurable DL and UL frame length from 2.5 to 20 ms Each UL subframe is preceded by preamble to allow BS to sync on each individual SS
1 Frame (2.5 to 20 ms)

P H B1 B2 B3 B4 TTG P B1 P B2 P B3 P B4 RTG
Downlink subframe (basestation)

Uplink subframe (subscriber)

Page 57
SC (Single-carrier) vs. MC (Multi-carrier) Modulated System

SC: each user transmits & receives data with only one carrier at the same time
Serially modulated scheme To increase throughput would require higher symbol rate which is more susceptible to channel effects Highly susceptible to crosstalk, ISI, multipath fading BW = 1+ % RRC alpha roll-off, 3 dB bandwidth of SQRT raised cosine filter Decreasing roll-off increases the PAPR and interference
MC: each user can employ number of subcarriers to transmit data simultaneously
Parallel modulation scheme Slower subcarrier rate that makes it easier to process and more rugged against interference Simpler frequency domain equalizer Manageable DSP Robust to interference
MC, Multi-carrier
Divide the shared wideband channel into N sub-channels
Data divided into Na active substreams
Substream modulated onto separate subcarriers

Substream bandwidth is B/N where B a total bandwidth
B/N < Bc implies flat fading on each subcarriers (no ISI), Bc a coherence time bandwidth
Spacing between two carriers is proportional to 1/T, where T is the IFFT symbol duration
BWOccupied = NaxTIFFT, sharper roll off due to lower sub-carrier frequency (higher IFFT rate) & DSP process. Na active subcarriers, TIFFT IFFT sampling duration
Page 59
Impairment Affects on SC vs. MC

Impairments affect differently on SC vs. MC systems
Impairment IQ gain balance IQ Quadrature skew IQ channel mismatch Uncompensated freq. error Phase noise Nonlinear distortion Linear distortion Carrier leakage Frequency error Amplifier droop Spurious
OFDM State spreading (uniform/carrier) State spreading (uniform/carrier) State spreading (nonuniform/carrier) State spreading State spreading (uniform/carrier) State spreading Usually no effect (equalize) Offset constellation for center carrier only (if used) State spreading Radial constellation distortion State spreading or shifting of affected subcarrier
Single Carrier Distortion of constellation Distortion of constellation State spreading Spinning constellation Constellation phase arcing State spreading (may be more pronounced on outer state) State spreading if not equalized Offset constellation Constellation phase arcing Radial constellation distortion State spreading, generally circular
Page 60
Multiple Carrier Modulated System

MC Advantages
Data are shared among several subcarriers and simultaneously transmitted. Pulse length ~N/B Flat fading per subcarrier N long pulses ISI is comparatively short N short EQs needed Facilitate NLOS operation with added guard interval Manage spectral efficiency with null subcarriers Easier to exploit frequency diversity Allows to deploy 2D coding techniques Dynamic signaling Exploit MIMO operation
MC Disadvantages
Higher linearity requirements due to PAPR Reduced system gain due to additional back off Higher power devices require more power dissipation, real estate space and cost Sensitive to phase-noise and clock inaccuracy Additional circuit, processing resources and cost for IFFT/FFT Reduced spectral efficiency due to added guard interval
Page 61
SC, Single Carrier

SC Advantages
Efficient and lower power consumption Complexity of transmission is much simpler than that of reception, making it suitable for asymmetrical operations High level of narrow-band noise immunity due to inherent capability by use of adaptive equalization Less susceptible to phase noise
SC Disadvantages
Data are transmitted over only one carrier. Pulse length ~1/B Selective fading
Very short pulse

ISI is comparatively long EQs are then very long Poor spectral efficiency because of band guards Sensitive to group delays
Lower peak to average ratio

Frequency domain equalization for performance improvement
MC, Multi-carrier
Delay
For each subcarrier, Rx receives a composite of sinusoids Same frequency but different phase and amplitude Fairly robust in frequency selective fading channel For single carrier transmission system, if the channel encounters interference at this frequency, the entire transmission can fail In OFDM, the problem is reduced since only a few of the N subcarriers will be affected. This means loss of a few bits instead of the entire OFDM symbol Powerful error correcting codes can be used to help restoring the erroneous bits in the corrupted subcarriers
SC
Hi-Freq Signal
Multipath Signal
Combined Signal
Delay
OFDM
Low Freq Signal
Multipath Signal
Combined Signal
Time-frequency Grid
Used Bandwidth
Data bits
Frequency
f0
Bad subcarriers
T=1/f0
One OFDM Symbol Time
Page 63
SC vs. MC, (1)

The ability to overcome delay spread, multi-path, and ISI in an efficient Symbol have wide freq. short manner that allows for symbol time higher data rate throughput
As an example, it is easier to equalize the individual OFDM subcarriers than it is to equalize the broader single carrier signal
Single Carrier
Level
OFDM Mode
Frequency S0 S1 S2 S3 S4 S5 Time
Frequency
Symbol have narrow freq. long symbol time

S0 S1 S2 S3 S4 S5
Serial symbol stream used to modulate a single wide band carrier Single Carrier
Each of the symbols is used to modulate a separate carrier
Deep Fade
OFDM Mode
Level
Frequency
Frequency
The dotted area represents the transmitted spectrum. The solid area is the receiver input
Page 64
SC vs. MC, (2)

Single Carrier systems are fairly robust to frequency offset errors and are more appropriate for mobile environment that experience large frequency offset errors The complexity of the equalizer for Single Carrier system is much greater than the multi-carrier system
Multicarrier systems are fairly robust to timing errors compared to a Single Carrier system. Their performance is similarly affected by the loss in SNR caused by frequency offset errors. Intuitively, this is easily understood from the fact that the Multicarrier symbol duration is N-times longer than its single carrier counterpart operating at the same data rate
LOS vs. NLOS

LOS, direct non-obstructed path
LFSL = 10Log(4Dm /)2, Dm distance in m, =c/f wavelength in m Optical LOS, Dkm = 3.57 SQRT (H), H antenna height
Radio LOS, Dkm = 3.57 [SQRT (kHB)+SQRT(kHM)], k effective earth kfactor
NLOS, Rx signal reaches through reflections, scatterings and diffractions

Signal have different delay spreads, attenuations, polarization & stability relative to direct path OFDM technology takes advantage of this phenomena LNLOS = Free space loss + terrain induced loss
Page 66
Fresnel Zone
Fresnel zone clearance depends on frequency & path length 1st Fz = 0.5 wavelength = 17.31 SQRT{(d1* d2) / (Dkm fGHz)}, d1 & d2 distance from obstruction to antenna, Dkm total distance Destructive affects at even orders of Fz Signal summation of same and or opposite phase
Page 67
Multipath Fading
More than one transmission path between Tx and Rx
Receive signal is the sum of many versions of the Tx signal with varied delay and attenuation
Reflection occurs when a propagating electromagnetic wave impinges upon a smooth surface with very large dimensions relative to the RF signal wavelength (=c/f, c speed of light, f operating frequency)
Buildings, ground, billboards, media
Diffraction occurs when the propagation path between Tx and Rx is obstructed by a dense body with dimension that are large relative to . Wave bends around sharp objects
Terrain, top of buildings
Scattering occurs when a radio wave impinges on either a large, rough surface or any surface whose dimensions are on the order of or less, causing the energy to be spread out or reflect in all directions. In an urban areas it is caused by lamppost, street signs and foliage. Multipath propagated signal affected by
Velocity, path, attenuation, time delay, Doppler shift, number of paths, etc.
Reflection Diffraction
PRX = PTX GTX GRX ( / 4D)2 for LOS path
Scattering
Page 68
Two-Ray Ground Propagation Model

If there are obstructions between the transmitter and receiver, wave will traverse multiple paths
Radio waves arrive at receiver from different directions and with different time delays Resultant signal at receiving antenna is vector addition of incoming signals Individual signals can add constructively (resultant signal has large power) or destructively (resultant signal has small power) depending on relative phases
EM wave impinging on surface will be reflected with some attenuation (determined by reflection coefficient) Two-ray model assumes one direct LOS path and one reflection path each reaching receiver with significant power
Typical Signal Attenuation

Environment Drywall, per cm Sheetrock wall, 2x4 Office whiteboard, per cm Clear glass, per cm Mesh glass, per cm Office wall Wooden wall Brick wall Metal wall Foliage, 3 m deep FSL, 1 km Rural open space, 1 km Suburban, 1 km Urban, Newark, 1 km Urban, Philadelphia, 1 km Urban, Tokyo, 1 km
Sig. Attn dB @ 2.5GHz 2.1 6 0.3 20 24.1 10 15 30 45 8.3 100.4 104 117 119 125 139
Page 70
Sector Antenna Pattern, example
Page 71
Frequency Selective Scheduling

OFDMA is fairly resistive to frequency selective fading since its parallel nature allows errors in sub-carriers to be corrected
Mobile WiMAX signal occupies a portion of the bandwidth. In broadband wireless channels, propagation conditions can vary over different portions of the spectrum in different ways for different users. Mobile WiMAX supports frequency selective scheduling to take full advantage of multi-user frequency diversity and improve QoS. WiMAX makes it possible to allocate a subset of sub-carriers to mobile users based on relative signal strength. By allocating a subset of sub-carriers to each MS for which the MS enjoys the strongest path gains, this multiuser diversity technique can achieve significant capacity gains.
F1
F1
F1 F1
F1
F1
F1 F1
F1
F1 F1
F1
F1
F1 F1
F1 F1
F1
F1 F1 F1 F1
F1
F1
Page 72
Frequency Reuse Approach (1)

Sectorized vs. Omni antenna
Higher directivity (higher gain). Must do up front performance tradeoffs Interference, range and cost tradeoffs
F2 F1 F3 F1 F2 F1 F3 F1 F3 F2 F3 F1 F3 F2 F2 F2 F1 F3 F1 F3 F2
Interference from overlapping section of the sectors edge

Single & dual-pole antenna Lower order modulated signal may use the same frequency adjacent sectors on single polarized antenna Higher XPD requirements on a dual poled antenna especially for higher order modulated signal Applies dynamic frequency reuse across sectors based on loading and interference conditions Allocating non-overlapping subchannels for poor SINR area at the expense of spectral efficiency
Page 73
Frequency Reuse Approach (2)

(NcxNsxNf definition), Cluster of cell x Sectors per cell x Frequencies per cell 1x3x1, Higher interference (CCI) higher spectral efficiency 1x3x3, Lowest interference (CCI & ACI) lower spectral efficiency 1x3x11-3 Subchannels, Lower interference & lower spectral efficiency at edge
Transmission from BS (all sectors) and SSs must be synchronized while using different permutation subchannels to minimize interference
PUSC typically uses 1/3 subcarriers per sector Randomly assigns subcarriers to subchannels using PUSC scheme in an unloaded network (not very affective when load increases)
FUSC & AMC would result in large coverage holes

F1S2a
F2 F1 F2 F3 F1 F3 F1
F1S1a
F1S2
F2 F3 F2 F1 F3
F1S3a
F1S1a
F1S2a
F1 F2 F3 F2 F1 F3
F1S3
F1 S 1
F1S2
F1S2a
F1S3a
F2 F1 F3
F1S2
F1S3
F1S1
F1S3
F1S1
F1S3a
F1S1a
Page 74
Frequency Reuse (3)

Mobile WiMAX also support frequency reuse one, i.e. all cells/sectors operate on one frequency channel to maximize spectrum utilization. However, due to heavy interference in (common frequency) reuse 1 deployment, users at the cell edge may suffer low connection quality
In WiMAX the sub-channel reuse pattern can be configured so that users close to the base station operate on the zone with all sub-channels available. While for the edge users, each cell/sector operates on the zone with a fraction of all sub-channels available.
Page 75
Fractional Frequency Reuse (4)

Subchannel reuse planning can improve the coverage across the cells/sectors based on network load and prevents interference
F1a, F1b and F1c represent different sets of sub-channels of the same frequency Full frequency use (maximum sub-channels) at the center while fractional frequency use at the edges The sub-channel reuse planning can be dynamically optimized Other implementation forms include time-coordinated and power-coordinated transmission
Transmission across BSs and sectors are coordinated in order to achieve maximal F=F +F +F F interference avoidance F 1x3x1 Reuse
1 2 3 2
1a
F1=F1a + F1b + F1c
F F
F1
F F2 F3 F2 F F1 F F F3 1x3x3 Reuse F3 F1: F,S1 F2: F,S2 F3: F,S3
F1
F1
F1
F F1 F F
F1b
F1c
Fractional Freq Reuse
Page 76

.
frequency
Reuse 1 Area Reuse 3 Area Reuse Partition 3 All Resources Reuse Partition 2 Reuse Partition 1 time time
DL subframe Preamble Center cell FFR = 1 Whole cell FFR = 3

G A P
UL subframe Preamble Center cell FFR = 1 Whole cell FFR = 3

Page 77

Cell 3 Cell 3 Cell 3
Cell 2
Cell 2
Cell 2
Cell 1
Cell 1
Cell 1
Frequency
Frequency
Frequency
Cell 1
Cell 2
Cell 3
Power
Power
Power
Power
Frequency
Uniform
Hard reuse 3
Fractional reuse 3
Soft reuse 3
3 2 1
3 2 1
3 2 1
3 2 1
Cell 3 Cell 2 Cell 1 Frequency
3 2 1
Reuse-3 Scheme
3 2 1
FFR-A Scheme
3 2 1
FFR-B Scheme
Page 78
DFS, Dynamic Frequency Selection

Feature used in license-exempt frequency band only
Automatically detects and avoids interference by moving to a different frequency location within the band
Prevents harmful interference into other users

Provides improved system performance A mandatory feature
Page 79
RF/ Mixed Signal Impairments

RF-LO phase noise, not correctable by adaptive equalizers Inter-modulation distortions, at inputs, at outputs or both Amplitude & group delay distortions Improper cable termination PA compression PA switching and settling time Antenna mismatch and low isolation Low signal combiner isolation Burst shaping error Recovered clock jitter IQ gain imbalance, IQ phase imbalance, (in-band spurious) Filtering distortions (normally compensated by signal processing) Distortion due to DC offset compensation DC offset (canceled in analog and signal processing) A/D and D/A converter non-linearities Thermal noise
Page 80
Digital Baseband Impairments

Improper channel estimation
IFFT, FFT Equalizer Incorrect coefficients Viterbi decoder BB-LO phase noise Timing / Frequency Sync Digital insertion loss
Latency
Processing circuit noise
Commonly Used Implementation Architectures & its Characteristics

Super-heterodyne (dual conversion)
Needs of channel IF filtering (external component) and two synthesizers. Less stringent IF filtering Good immunity from interfering signals and good selectivity performance Image is not a serious problem
Heterodyne with not fixed wide IF (2nd LO by division of 1st LO)

Removes part of DC offset issues, LO emission, pulling and flicker noise Higher complexity, more spurious, IQ imbalance & power consumption Spurious associated with the 2nd LO and IF frequencies: careful frequency plan required
Low IF
Removes part of DC offset issues and flicker noise Higher complexity, sensitive to IQ paths asymmetry
Homodyne (Zero-IF)
Reduced parts count, saves die/board size and power consumption Simple Frequency Plan Spurious and higher order mixing products associated with the 2nd LO and the IF frequencies are also eliminated from the frequency plan Isolation and dynamic rage trade-off Sensitive to DC offset , LO emission, LO pulling, flicker noise
Common Path Related Impairments

Distance dependent decay of the signal power
Blockage due to obstructions Large variation in received signal envelope
Due to constructive/destructive additions of multi-path signals
ISI due to time dispersion ICI due to local clock inaccuracy & phase noise
More critical for TDD than the FDD system If occurred, it is not correctable
Synchronization vs. clock drift Frequency dispersion due to motion Noise Interference from own & or intra-network equipment
Other Impairments
Atmospheric absorption water vapor and oxygen contribute to attenuation (not relevant for low freq WiMAX)
Multipath effects by terrain and environmental conditions
Obstacles reflect signals so that multiple copies with varying delays are received
Refraction bending of radio waves as they propagate through the atmosphere
Page 84
TX PO
Max PO, regulated by local regulatory agency
Asymmetric power level at SS, MS & BS Different device sizes may yield asymmetric performance at each end PO, determined by sum of power from all active subcarriers measured over certain number of symbols in time
Sub-carrier power varies depending on type of sub-carrier, modulation and content Total data PO =DataPWR of 1subcar +10Log(# of act data subcar) Total pilot PO =pilotPWR of 1subcar +10Log(# of act pilot subcar) Total symbol PO =10 Log(10^dataPWR/10 + 10^pilotPWR/10)
TX PO Constraints and Impairments

Requires vector power meter to measure specific symbol power (not feasible with traditional power meter due to TDD, DL/UL ratio, adaptive modulation, burst rate, training sequence etc.) MC system demands increased PA linearity for reliable high performance PAPR = 10 Log(# of subcarriers), additional requirements than SC Creates extreme peaks and valleys Normally not a serious occurrence issue due to data scrambling PA linearity may improve at the expense of Increased power back off Larger device (resulting in increased cost, power, real estate, thermal rise and lower reliability). Perform upfront trade offs Adaptive distortion control Increases cost Increases control algorithm complications Reduces processing resources Inappropriate power level affects system performance, spectral mask, spectral flatness, spurious, interference to other equipment, etc.
TX PO Impairments
PA non-linearity causes IMD that results in spectral regrowth
Select higher OIP3, OIP5, IIP3, IIP5, P1dB, SFDR to improve performance Demands more linearity at higher order modulation Non-linear distortion can not be corrected by equalizer
Spurious may also originate at other areas of the circuit such as in non-linear mixer, LO phase noise, DAC, IQs, filters, etc. Performance degradation affects at its own near-end / far-end Rx and other operators in the vicinity Regulatory agency controls the Tx signal quality (Po, Freq., BW, spectrum, spurious, noise floor, ACLR, etc.) in order to protect other operators
Page 87
TX Impairments
Tx impairments affect the performance of its own and other neighbors in the vicinity Mitigation techniques includes power back-off, distortion control, larger device, signal clipping, selective mapping, partial IFFT, etc Destructive effects resulting from IMD3 & IMD5
1dB OIP3
Fundamental Output Power SFDR

f
Output Power
Fundamental
3rd Intermod
Rx Thres IMD3 IMD5 Floor Freq

Noise NF Input Power SNR
IMD3
Fundamental
3rd 5th
f
2f1-f2
f1
f2
2f2-f1 3f2-2f1
kTB
SFDR BDR
P1dB IIP3
Page 88
Spectral Spreading Control

Commonly used mitigation techniques includes:
Operate at increased power back-off Forcing counter distortion Using larger PA devices Signal clipping Digital domain clipping also introduces spreading and minimizes the effective SNR
Passing the clipped signal through BPF prior to PA eliminates spreading

Selective mapping Partial IFFT
From OFDM modulator Simulated Transmitted signal
Clip to specified Prefilter OBO
BPF user FIR
Clip to specified Output Power Amplifier OBO
Page 89
TX spectral mask
Reference ETSI EN302 326-2 & EN320 544-1 RBW is generally set to about 1% of the BW if not specified ACLR: 44.2 dB at x1 CS, 49.2 dB at x2 CS (Channel Spacing) The spectral mask basically specifies the accuracy of the out of band signal
-5
-15
Attenuation (dBr)
4QAM 16QAM 64QAM
-25
-35
-45
-55 0
Frequency/CS
2
Page 90
TX Spectrum, (1)
Frequency domain representation of one OFDM symbol
Modulation scheme & power adjustable per sub-channel
Page 91
TX Spectrum, (2)
Higher spectrum efficiency
Place unused sub-carriers at the beginning & end of OFDM symbol Rectangular spectrum shape (almost like brick wall) For larger number of subcarriers the spectrum goes down rapidly in the beginning, which is caused by the fact that the side lobes are closer dB together Roll off relative to sub-carrier rate Small frequency guard band
BW= BOU + BOL + (FOH FOL)

BOU, BW of upper subcarrier BOL, BW of lower subcarrier
OFDM -80 Single Carrier
FOH, Upper frequency edge

FOL, Lower frequency edge
X MHz
Freq
Page 92
Tx Spectrum, (3)
Use vector spectrum analyzer to capture a non-traditional signal: TDD, DL/UL ratio, adaptive modulation, burst rate, training sequence, etc.
Sharp almost brick wall like spectrum, allows more data in the allowed BW Tx spectral flatness to be within 2 dB over all active tones for spectral lines starting from -50 to -1 and +1 to +50. +2/-4 dB over all active tones for spectral lines from -100 to -1 and +1 to +100. To be within 0.1 dB for adjacent subcarriers
Page 93
TX Spectrum, (4)
Preamble
Page 94
TX BW
Tx BW is determined by total active data and pilot sub-carriers

For example, BWAllowed =20 M, NFFT = 2048, active data = 1440, pilots = 240, subcarrier spacing = 11.160714 kHz BWOccupied = (1440+240) * 11.160714 = 18.75 MHz If an input data rate R bps, Nused of FFT, M Modulation order, 3/4 FEC then each active data subcarrier carries {(R/Nused) * (4/3) * M} load
Page 95
TX Frequency
MC system demands higher frequency stability & accuracy to deliver a consistently reliable performance
1 ppm is required for FDD & TTD operation over the life of product
This equates to 2.5 kHz for Tx only at 2.5 GHz
The subcarriers frequency is typically about 10 KHz
SS to BS synchronization tolerance to be 2 Hz Timing accuracy of 5-25 us required for TDD system Frequency inaccuracy increases the ICI
Page 96
Thermal Noise, (1)
Thermal noise due to agitation of electrons

Present in all electronic devices and transmission media It cannot be eliminated Function of temperature (increases at higher temperature, 1.68 dB from -33 C to +80 C)
Page 97
Thermal Noise, (2)
Amount of thermal noise to be found in a bandwidth of 1 Hz in any device or conductor is:
N0
kT W/Hz
N0 = noise power density in watts per 1 Hz of bandwidth N0 = -173.93 dBm/Hz into a 50 Ohms load (antenna) at room temperature k = Boltzmann's constant = 1.3803 * 10-23 J/K T = temperature, in Kelvin's (absolute temperature)
Noise is assumed to be independent of frequency
Page 98
Noise and Threshold

RX Threshold = kT +10 Log (BW) + NF + SNR
kT =10 Log(1.38e-23 * 293 K) = -204 dBW/Hz = -174 dBm/Hz into a 50 ohm antenna k Boltsmanns constant, 1.38e-23 W/Hz/K
BW is the RX signals 3 dB bandwidth

BW is computed differently for MC system Post processing SNR NF varies with Freq band, RF filter & cable losses (adds dB for dB) T room temp in Absolute term, (273+20) K NF increases due to rigid filter requirements for narrow FDD T-R spacing NF increases due to higher insertion loss in narrow band filters
NF increases with temperature rise

SNR requirements vary with Mod level, FEC power & modem design approach
Page 99
Threshold & Interference

N = ktB * NF, also known as noise floor in non-log terms
I, Interference to cause 1 dB threshold degradation at 1e-6 BER I, dBm = N+I =N+(-6)
S
Thermal Static Fade Margin
64QAM o o 4QAM- 4QAM- BPSK-
Unfaded RF RX Level
SNR, dB = Signal / Noise

SINR, dB = SNR + 1 SIR = S/(N-6)
SIR = SNR when S +6 dB
T
SIR T/I N SINR SNR
1e-6 BER with I+N 1e-6 BER Threshold with Noise

1dB
Noise Floor + Interference Noise Floor
6 dB, objective for 1 dB threshold degradation
I
Interference Level
Page 100
RX Constellation Error
Rx relative constellation error includes transmit constellation error, Rx constellation error plus the channel impairments Provides Tx-Rx and channel condition prior to error correction Tested under any of the normal operating conditions
OFDM Burst Type BPSK-1/2 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-2/3 64QAM-3/4 OFDMA Burst Type 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-1/2 64QAM-2/3 64QAM-3/4
Relative Cons Error for SS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -30.0 Relative Cons Error for SS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Relative Cons Error for BS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -31.0 Relative Cons Error for BS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Page 101
RX Threshold
Rx to remain operational at signal up to -30 dBm for all modulations No damage to equipment at signal up to -0 dBm
Requires higher IIP3 and IIP5 devices at RF front end
Hi-RSL is more sensitive to higher modulation Rx must detect Rx signal up to -90 dBm min Rx dynamic range of 50 dB min PER to be better than 0.49% Image rejection to be 60 dB min Receive threshold is determined by the number of data subcarriers & frame length while excluding the pilots & preambles
Page 102
RSSI, Receive Signal Strength Indicator

Measurement referenced at RF Rx input
Signal detection over a wide signal range (-10 to -90 dBm for 16d, 40 to -90 dBm for 16e). Measurements to continue down to -123 dBm Tolerance accuracy over environmental conditions (within 2 dB relative, 4 dB absolute) Controlling parameter for other critical functions Covers the full RF filter bandwidth Performs fast estimation to determine available BW, mod & FEC allocations
RSSI determined by excluding pilots & preambles

Fixed power DL RSSI, adaptive power UL RSSI
Typical SNR vs. Modulation

Ratio of the RX signal power to noise power
Required minimum SNR for 1e-6 BER
Modulation SNRReq BPSK-1/2 4QAM-3/4 4QAM-3/4 16QAM-1/2 3.0 6.0 8.5 11.5
Table shows typical SNR using CC & RS FEC types vs. Mod level
Further reduction with more powerful codes
A key quality metrics for RX signal
16QAM-3/4
64QAM-2/3
15.5
19.0
SNR requirements vary with Mod type, 64QAM-3/4 FEC power & modem design techniques
Minimum post processing requirement Lower requirements with higher powered FEC Lower requirements at lower modulation
21.0
Page 104
SNR
S/W performs fast background computation to determine channel conditions and fade margin
Useful to check the presence of steady interference at normal RSL
For SC System
SNR value decreases (worse) with increased path impairments

Provides current channel status conditions to optimize transmission
16QAM Error Distance 'd'
Phase Noise Effect

Accumulated phase noise from all sources (Tx-BB, Rx-BB, TX-LO, Rx-LO)
Phase noise effect appears different on OFDM system compares to a single carrier system
RX Const
1 Scatter Plot
Quadrature Amplitude
2 1 0 -1 -2 -3
0.5
-0.5
-1
-2 0 2 In-phase Amplitude
-1
-0.5 0 0.5 In-phase Amplitude
Multicarrier
Single carrier
Page 106
Interference, (1)
Spurious Caused by different combinations of signals in the Tx and Rx Harmonics are integer multiple of the primary transmitter/receiver frequency Predictable location and traceable back to primary frequency source
Typically harmonics are measured up to 5x the frequency or up to 17.5 GHz

Spurious signals are typically image frequencies caused by internal mixing of an oscillator or clock freq with the primary transmitter/receiver frequency Difficult to trace due to change in level and mixing location
Others types such as CIR, CINR, PCINR, ECINR, SINR, CCI, ACI, CW, ICI, ISI
Easy to avoid/reject (with null carriers) narrowband interference with subchannels Less interfered part of the carrier can still be used
f
Interferer Rx Filter IMD3

f
IMD5 Freq
2f1-f2
f1
f2
3f2-2f1 2f2-f1
Page 107
Interference, (2)
Tx, Rx spurious interference to be -47 dBm in 1 MHz BW
Power Rx Filter ACI
Desired Signal Thermal Noise Out of channel interference CCI Frequency
Page 108
CCI & ACI (Co & Adjacent Channel Interference)

Co-channel interference occurs when another transmission on the same carrier frequency affects the receiver 16QAM Adjacent-channel interference occurs 30 when energy from a carrier spills over into adjacent channels 64QAM 25 Standards specify a reference 1 & 3 dB degradation from interference 2x-Adjacent 20 within co-channel and from adjacent Channel channel bandwidth 15 Using same BW and type of signal Tests are performed with a same 10 order modulated signal and bandwidth Adj-Ch 5 -6 BER Degradation referred to 1e Interference tolerance to determine -25 from inversed T/I curve
Level above sensitivity (dB)

Sensitivity
Page 109
Interference
CCI and ACI requirements at 1/3 dB degradation CCI (dB) x1 ACI (dB) x2 ACI (dB) Reference 1e-6 BER 3 dB deg Modulation 1 dB deg 3 dB deg BPSK-1/2 4.0 -11.0 -30.0 4QAM-1/2 8.0 -11.0 -30.0 4QAM-3/4 9.5 -11.0 -30.0 16QAM-1/2 12.5 -11.0 -30.0 16QAM-3/4 16.0 -11.0 -30.0 64QAM-2/3 19.0 -4.0 -23.0 64QAM-3/4 22.0 -4.0 -23.0
Page 110
CW (Continuous Wave) Interference

Narrow out of band signal may affect the normal Rx AGC operation
It may exceed the device maximum overload tolerance Affects RSSI detector accuracy Mixing products may fall in- IMD3 band
Rx Filter
f
Interferer
IMD5
Tolerable limit is specified by ETSI standard

Freq
2f1-f2
f1
f2
3f2-2f1 2f2-f1
Page 111
ISI, ICI, LO Phase Noise and Clock Offset

ICI & ISI are normally caused by phase noise, poor synchronization, unstable subcarriers clock, insufficient delay spread and Doppler shift
ICI & ISI become more sensitive at higher modulation To avoid inter-carrier interference, the inter-carrier spacing is set to be equal to the inverse of the symbol duration.
Non-linear distortion and phase noise are the two largest contributing factors to a loss of orthogonality, creating an ICI. Poor frequency estimation in the receiver is another contributing factor ISI introduces an irreducible error floor which can not be removed by increasing transmit power
ISI & ICI

Signal arriving late from secondary path
Affects both ISI & ICI
(i-2) (i-1) (i+0) (i+1) (i+2)
OFDM symbol
OFDM symbol
OFDM symbol
|Ca(t)|
OFDM symbol
OFDM symbol
Magnitude of channel impulse response
OFDM symbol Fade in (ICI) OFDM symbol
Fade out (ISI)
Page 113
ISI, Inter Symbol Interference

Delay spread: defined as the RMS time difference between the arrival of the first and the last multipath signal seen by the receiver
The delay is affected by distance, frequency and environment For mobile, it is also dependent on the speed (Doppler shift)
Typical delay spread: 40 to 200 ns for indoors (50 ns in homes, 100 ns in offices, 300 ns in industrial environment), 1 to 20 us for outdoors
ISI becomes more serious as the bit rate increases (/Ts gets worse i.e., bigger). is delay spread, Ts sample time For OFDM ratio of /NTs to become smaller (better ISI) Sensitive to LO phase noise (from all sources)
For MC system, ISI is less sensitive to narrowband interfering signal and frequency selective fading
ICI, Inter Carrier Interference

OFDM systems becomes more susceptible to time-variations as symbol length increases
Increase the CP length and number of pilot tones to mitigate the ICI Lower FFT size increases the subcarrier spacing that improves the ICI and more tolerant to Doppler shift Time variations introduce ICI in frequency domain Signal arriving from multipath causes ICI. If occurred, it is not correctable LO phase noise and clock recovery error produces wider overlapping skirt at the lower part of the subcarriers in frequency domain. This phenomena is independent of clock stability
F-2
F-1
F0
F1
F2
ICI
Dynamic Range
Receive signal ratio between the maximum possible signal and the minimum signal that gives the desired signal level over noise at demodulator input
The range includes input power (signal, noise and interference) over which receiver performs adequately Performance determined at a reference 1e-6 BER Determined by aggregate AGC in the receiver chain
BER
1e-2 1e-4 1e-6 1e-10 Overload Thres
-RSSI
TDD, Time-Division Duplex

DL & UL timeshare the same RF channel
With a gap period at transition to accommodate Tx/Rx mode switching and PA settling time
BS or SS, neither transmit/receive simultaneously
On DL, SS is associated with a specific burst

On UL, SS is allotted a variable length time slot for their usage Single RF filter and single RF-LO Less stringent filter requirements Less data throughput Increased MAC control complexity (less hardware complexity) Readily available lower cost parts due to higher usage in unlicensed band
Dynamic asymmetry ratio of DL/UL

Unlicensed operation is limited to using TDD format
Page 117
FDD and TDD Frame Structure

UL Subframe (PHY PDU)
Frequency
RNG BW
UL P Burst #1 ... P
UL Burst #k
FDD
Contention UL SS #1
UL SS #k
DL Subframe (PHY PDU) P FCH DL Burst #1

...
DL Burst #m
Frame n-1
Frame n
Frame n+1
Frame n+2
DL Subframe (PHY PDU)

TTG
UL Subframe (PHY PDU)

RTG
TDD
FCH DL Burst #1 ... DL Burst #m
RNG BW
Contention
P Burst #1 ... P
UL SS #1
UL
UL Burst #k
UL SS #k
Time
FCH: Frame Control Header P: Preamble

RNG: Contention Slot for Ranging Request BW: Contention Slot for BW Request
TTG: Tx/Rx Transmission Gap RTG: Rx/Tx Transmission Gap

Page 118
TDD & FDD, HFDD
TDD
FDD
HFDD
Page 119
FDD, Frequency-Division Duplex
DL & UL on separate RF frequency channels

BS & SS transmit/receive simultaneously
Static asymmetry
Half-duplex SSs supported
SS does not transmit/receive (lower cost)
Continuous operation, no switch settling time required

Requires two frequency channels Higher performance front end RF Simpler MAC control operation
TDD & FDD

TDD
Advantages Asymmetric DL/UL ratio
FDD
Advantages Better protection against interferences (separate DL/UL ratio) Stronger synchronization of receiver
Lower cost of RF elements

More options with channels size Simple AAS in MIMO implementation Disadvantages Vulnerable to interference
Network planning is easier

Disadvantages Fixed DL/UL ratio More expensive Less options with channel sizes
Synchronization of receiver
Synchronization of network
Page 121
WiMAX
WiMAX architecture consists of two key items PHY BB & RF processor (Frequency source, Mod, IFFT/FFT, timing recovery, Sync, multiple interface access, error detection & correction etc.) MAC Standard compliant LAN and end to end interface Protocol control / process / manage and QoS toward LAN interface
Protocol control / process / manage and QoS toward end-toend system

Base Station & Subscriber Station

Base Station (BS):
Controls the entire system, frame size, scheduling, admission control, QoS, Ranging, clock synchronization, power control, handoff, privacy key and PHY management All traffic goes through BS
Subscriber Station (SS):

Finds BS, acquire PHY synchronization, obtain MAC parameters, generate bandwidth requests, make local scheduling decisions, follows transmission/reception schedule from BS, performs initial ranging, maintenance ranging and power control
Mobile Station (MS):

In addition to the SS functions, mobility management, handoff, power conversion and power management
System Design Requirements

BW, Bandwidth
Bit rate Subcarrier spacing
Tolerable delay spread

Doppler shift value Bits per OFDM sym = Bits rate * (active data subcarriers) * FEC * Log2(Mod) QoS
Modulation (M) Channel Coding Inter leaving Modulation (M) S/P Modulation (M) Modulation (M) Mapping from size N/2 to N To Channel
Source Data
IFFT
S/P
CP
Page 124
Key System Design Parameters
Channel bandwidth
Number of subcarriers CP, cyclic prefix Subcarrier spacing Modulation
FEC
PO Dynamic range Threshold
OFDM Subcarriers
An OFDMA symbol consists of three types of subcarriers:
Pilot subcarriers Data subcarriers Null subcarriers
Im Received symbol Received symbol
Subcarriers can be turned on/off dynamically based on channel conditions and to meet the required BW
Real Transmitted symbol
Page 126
Pilot Sub-carrier
Pilot subcarriers contain signal values that are known to the receiver Facilitate signal recovery and synchronization Pilot subcarriers are used in the receiver for correcting the magnitude (important in QAM) and phase shift offsets of the received symbols (see signal constellation example on previous page) Magnitude and phase of these subcarriers are known to receiver that helps to speed up channel estimation Always BPSK-1/2 modulated & its transmission repeated Higher power level (2.5 dB higher than the average power of the non-boosted data tones Transmitted with embedded Pseudo random code Inserted after the FEC stage so as not to destroy the fixed time and amplitude relationships that these signals must possess to be effective 8 pilots for OFDM (Configurable number for each transmitter in OFDMA) More pilots increases noise resiliance & processor loading while reducing the overall throughput For OFDM, pilots are common to all UL-subchannels For OFDMA, certain numbers are dedicated to specific subchannels
Data Sub-carriers
Used to transport over head control and user data
Part of the DL/UL data subcarriers contain preamble symbols for training purposes DL has two long preamble symbols of QPSK: Two training cycles at the start of each 8 us (1st containing 50 subcarriers and called short training sequence, every 4th subcarrier with a phase relationship that minimize the PAPR. This period is used for RX gain setting and course frequency correction. All have the same levels. 2nd containing 100 subcarriers and called long, 8 us, all subcarriers turned on. Allows RX to calculate frequency response of the channel and to fine tune the frequency errors). Preambles are 3 dB stronger than all other symbols in the DL frame. UL always starts with preamble (called short preamble, 100 subcarriers of QPSK. Preamble has no pilot carriers. Helps Rx to sync and perform additional channel estimation). Modulation remains the same within burst but changes from burst to burst.
Following DL Preamble is the FCH (single symbol of 88 bits, BPSK-1/2 for OFDM, QPSK-1/2 for OFDMA), occupies 1st two subcarriers in the 1st data symbol
Remainder of the data subcarriers carry the user data
Null Sub-carriers
To avoid difficulties in DAC and ADC converter offsets, and to avoid DC offset and PA saturation, the sub-carrier falling at DC is not used
Relaxes anti-aliasing and filtering requirements
DC subcarrier power must be at least 15 dB lower than the average of all other subcarriers
Provides a frequency guard band before the Nyquist frequency and allows for a realistic roll off in the analog anti-aliasing reconstruction filters Used for spectrum shaping and to fit the regulatory mask
Null subcarriers contain no power
Typical OFDMA Blockdiagram
Variable FFT & subcarriers size based on the BW

Constant subcarrier frequency spacing Configurable over-sampling factor
SCRM & FEC
Interl eaver
QAM Mod
Pilot
S/P
IFFT
P/S
CP & Win
DAC & FLTR
RFTX
FEC & DScrm
D-Int
QAM DMD
FD EQL
P/S
FFT
S/P
CP Remove
TMG & FreqSyn
ADC & FLTR
RFRX
Page 130
Scrambler
Randomization prevents long sequences of 1s or 0s in the incoming data stream Helps speed up and maintain clock recovery DL & UL data is randomized by modulo-2 addition of every data bit with output of a pseudo random binary sequence generator Randomization is performed on data bits only A pseudo random binary sequence of 1+X14+X15 Each frame starts with initialization sequence of 100101010000000 Randomization is performed on each allocation (DL or UL) independently
Page 131
FEC (1)
Probability of symbol location after passing through AWGN channel
When does error occur? Expected symbol ends up in neighbors territory Symbol vs. bit error Error multiplication at higher order modulation
Probability Density Function
Probability Density
Receive symbol position

FEC (2)
Hard decision declares error the moment it crosses the decision boundary Soft decision further adds statistical values in error computation Error distance decreases on higher modulation making it more susceptible to error BPSK = 2.0 1/4th constellation View
256QAM
15 170 13 178 194 218 250 290 338 394 226 234 250 274 306 346 394 450
QPSK = 1.414
16QAM = 0.471 64QAM = 0.283 256QAM = 0.202 Signal compression at outer most SER, BER. PER, FER
128QAM
122 11 82 9
130
146
170
202
242
290
346
90
106
130
162
202
250
306
64QAM
50 7 58 74 98 130 170 218 274
Modulation BPSK QPSK 16QAM 64QAM

Error 90 45 16.9 7.7
32QAM
5
26
34
50
74
106
146
194
250
16QAM
3
10
18
34
58
90
130
178
234
4QAM
1
10
26
50
82
122
170
226
11
13
15
Error Distance 'd'

Page 133
FEC (3)
Process computes and adds additional parity bits at the transmit which helps identify error location & possible correction by receiver
FEC implementation varies by Cap/BW/Modulation
Reed Solomon (RS) only RS + Convolution RS + Convolution + Interleaver
Detected error quality is used to control adaptive modulation, coding rate, data integrity, error performance, bandwidth allocation, subchannelization, AAS (adaptive antenna system) & MIMO calculations If the last FEC block is not filled, that block may be left shortened Shortening in both UL and DL is controlled by the BS and is implicitly communicated in the UL-MAP and DL-MAP
FEC (4)
Adds redundancy to data bits
Programmable concatenated Reed Solomon (good for low BER, 1e-8) and Convolution coding (good for 1e-4 to 1e-7 BER) Total of 7 different rate dependent combinations Support of Block Turbo Coding (BTC), Convolutional Turbo Coding (CTC) and low density parity coding (LDPC) is optional The Reed Solomon encoding shall be derived from a systematic varied length RS code (k, n, t) where n is the number of overall bytes after encoding, k is the number of data bytes before encoding, t is the number of data bytes which can be corrected Encoder supports shortened and punctured codes to accommodate variable block size Reduces overall throughput according to the selected coding rate
FEC (5)
Incoming data bytes are processed serially (byte by byte) over a fixed RS block length then adds the parity bytes at the end
Low rate coding may be punctured by deleting zeros to lower overhead (i.e., deleting 2 out of 6 bits of to create a rate)
Raw Data (lower speed)
Block 2
Block 1
Block 0
After RS (higher speed)
X
Parity Bytes
X Data Bytes
Direction of Data Flow

X Out
Data In
1 bit delay
1 bit delay
1 bit delay
1 bit delay
1 bit delay
1 bit delay
+
Y Out
Page 136
FEC (6)
Four FEC schemes defined in 802.16 802.16 defines concatenated coding schemes: inner code (random errors) and outer code (burst errors) Code type 1 (used for large data block or high coding requirements):
No inner code Outer codes: systematic Reed-Solomon (corrects errors: 16 to 0 bytes) Two modes of operation:
Fixed codeword: number of information bytes same for every RS codeword Shortened codeword: number of information bytes in the final RS block is reduced
FEC (7)
Code type 2 (useful for low to moderate coding rates that provide good performance):
Outer code: almost same as RS code as in code type 1
Inner code is a (24, 16) block convolutional code (BCC)

16 bits input block code, bi 24 bits output codeword ci (each symbol: combination of others symbol: c23 = b15+b0+b1)
Code type 3 (optional):

Outer code: almost same as RS code as in code type 1 and 2
Inner code: (9, 8) parity check code (code adds one parity bit to every eight bits)
Page 138
FEC (8)
Code type 4 (used to extend the range of a BS or increase the data rate at the same range):
No inner code
Outer code: block turbo code (BTC): The idea is to encode the data twice
Option: bit interleaving
k1 k2 n2
n1
Information bits Checks
Checks Checks on checks

Page 139
BER Curve
BER vs. Eb/No before and after the RS only FEC
Performance tradeoffs
FEC Gain
Page 140
Modulation & Coding Combination Summary
WiMAX supports 7 possible Mod / FEC rates to provide optimal data throughput
Other modulations and FEC types are optional
Uncoded Modulation Blocks (bytes) RS Code CC Code BPSK 12 (12, 12, 0) 1/2 4-QAM 24 (32, 24, 4) 2/3 4-QAM 36 (40, 36, 2) 5/6 16-QAM 48 (64, 48, 8) 2/3 16-QAM 72 (80, 72, 4) 5/6 64-QAM 96 (108, 96, 6) 3/4 64-QAM 108 (120, 108, 6) 5/6
Coded Blocks Overall (bytes) Coding 24 1/2 48 1/2 48 3/4 96 1/2 96 3/4 144 2/3 144 3/4
Page 141
Preamble (added after the FEC)

It is important that the frame control section of the DL frame be encoded with fixed set of parameters known to the SS at initialization in order to ensure that all subscribers stations can read the information The control portion of the frame is encoded with a Type 2 FEC where the outer code is a (46, 26) RS code and the inner code is a (24, 16) BCC
SCRM & FEC
Interl eaver
QAM Mod
Pilot
S/P
IFFT
P/S
CP & Win
DAC & FLTR
RFTX
FEC & DScrm
D-Int
QAM DMD
FD EQL
P/S
FFT
S/P
CP Remove
TMG & FreqSyn
ADC & FLTR
RFRX
Page 142
Interleaving
Data interleaving is very affective against burst (clustered) typed errors Process increases latency through the system All encoded data bits shall be interleaved by a block interleaver Interleaver block size corresponds to the number of coded bits per specified allocations The number of coded bits per carrier is 2, 4 or 6 for QPSK, 16QAM or 64QAM, respectively
Before Interleaving
X
No error block After Interleaving
X
No error block
Burst errored block
X
Correctable block
X
Correctable block
Correctable block
Data Input
Fill in
b1 b4 b7 b10
Data Out
b2 b5 b8 b11 b1 b4 b7 b10 b2 b5 b8 b11 b3 b6 b9 b12 b3 b6 b9 b12

Page 143
Constellation Mapping
Serial data bits are mapped into selected modulated symbol
Supports Gray-mapped BPSK, 4/ 16/ 64-QAM modulation
Support of 256-QAM is optional

Normalized to achieve unity average power regardless of modulation scheme Constellations must be normalized to achieve equal average power Supports adaptive modulation and coding
Page 144
Constellation Map and Spectral Efficiency

Modulation BPSK QPSK 16QAM 64QAM Spectral Efficiency 1 2 4 6
Imaginay b0b1b2 b3b4b5
000 100 001 100 011 100 010 100 110 100 111 100 101 100 100 100 7 000 101 001 101 011 101 010 101 110 101 111 101 101 101 100 101 5 000 111 001 111 011 111 010 111 110 111 111 111 101 111 100 111 3 000 110 001 110 011 110 010 110 110 110 111 110 101 110 100 110 1 -7 000010 -5 -3 -1 -1 000 011 001 011 011 011 010 011 110 011 111 011 101 011 100 011 -3 000 001 001 001 011 001 010 001 110 001 111 001 101 001 100 001 -5 000 000 001 000 011 000 010 000 110 000 111 000 101 000 100 000 -7 1 3 5 7 001 010 011 010 010 010 110 010 111 010 101 010 100 010
Input Bits Input bits I-Out Q-Out (b0b1b2) (b3b4b5) 000 -7 000 -7 001 -5 001 -5 011 -3 011 -3 010 -1 010 -1 110 1 110 1 111 3 111 3 101 5 101 5 100 7 100 7
Q = b 3 b4 b 5 . . b5 b4 b3 6 bits b2 b1 b0 I = b0 b1 b2 1 Symbol
Real
Page 145
Constellation Mapping (2)
Digital modulation: how bits are mapped to symbols. Constellation can be selected per subscriber (quality of the RF channel)
In the DL: QPSK, 16-QAM and 64-QAM Distance to origin: power that sends the signal, follows 2 adjustment rules
Constant constellation peak power and constant constellation mean power
Before Mod I & Q signals filtered by square root raised cosine pulse shaping filter:
S(t) = I(t)*cos (2fct) * Q(t)sin(2 fct)
Constellation Mapping (3)

After bit interleaving, the data bits are entered serially to the constellation mapper
Mapped to form symbol for selected modulation
Modulation with Gray coding

Normalized to achieve unity average power regardless of modulation scheme in order to facilitate timing recovery
SCRM & FEC
Interl eaver
QAM Mod
Pilot
S/P
IFFT
P/S
CP & Win
DAC & FLTR
RFTX
FEC & DScrm
D-Int
QAM DMD
FD EQL
P/S
FFT
S/P
CP Remove
TMG & FreqSyn
ADC & FLTR
RFRX
Page 147
Constellation Display
Multiple modulations captured in a single frame
Constant average power to stabilize Rx AGC loop gain
Page 148
Constellation Map
The Frequency Domain description includes the basic structure of an OFDM symbol
An OFDM symbol is made up from subcarriers, the number of which determines the FFT size used. There are several carrier types
Data subcarriers: for data transmission & down stream synchronization Pilot subcarriers: for various estimation purposes
Null subcarriers: no transmission at all, for guard bands and DC carrier
Page 149
IFFT/FFT
A DSP process that uses N points IFFT of a signal X(k)
Parallel data streams are used as inputs to an IFFT IFFT output contains N times data buckets Each bucket contains sum of many samples of many sinusoids Same frequency, different amplitude and phase At center of the subcarrier there is no cross talks from other subcarriers and hence makes receiver to correctly recover data IFFT does modulation and multiplexing in one step Normal DFT would require (N-1)^2 operation whereas the FFT would require only N/2*Log2(N) operations (i.e., 65025 vs. 1024 multiplications for 256 point FFT)
Page 150
IFFT/FFT
The IFFT operation in OFDM partitions a wide band channel into multiple narrowband subchannels The IFFT & FFT operations are almost identical. The IFFT can be made using an FFT by conjugating input and output of the FFT and dividing the output by the FFT size. May use the same hardware for Tx & Rx in TDD mode
IFFT modulates and multiplexes the signal in one step

DSP algorithms replace a required bank of IQ Mod-DMD that would otherwise be Time Domain Frequency Domain required
x x x
199
Parallel to Serial
Subcarrier Modulation Data IQ Vector x x x Zeros
IFFT
Q Output Base Band OFDM Signal
Input Data
1 0
Guard Period
Symbol Start
Page 151
Guard Time
Signal from multiple reflected paths arrive at various delays Delayed signal may corrupts the front part of the next symbol The guard time acts as a buffer to allow time for multipath signals from previous symbol to die away before the information from the current symbol to get collected by receiver It is like water splash when driving too close to a car in front A simple gap is not acceptable for optimal signal recovery at Rx Adding guard time lowers the symbol rate but does not affects the subcarrier spacing Environment Sig. Delay, ns Office/home NLoS 50 Open space office NLoS 100 Large open space office NLoS 150 Manufacturing area 200-300 Microcell 500 Large open space LoS 140 Large open space NLoS 250 Mobile city 2500 Mobile rural area 25000 Introduction to WiMAX Technology
Page 152
CP, Cyclic Prefix Plot

Data Carrier (192)
CP x(0), x(1), ..., x(N-2-v), x(N-1-v) Tb = FFT symbol duration Ts = OFDM symbol duration
,...,x(N-1)
DC Carrier
Pilot Carrier (8)
Tg
Guard Band Low (28)
Guard Band High (27)
Frequency Domain
CP vs. Guard Band

max Sampling start
Time Domain
MC Sample
Level 1/Fs
~ ~
NFFT Time
Tg
NFFT*1/FS=Tb(=1/f) Ts=Tb+Tg
Tg=G-Tb -1
NFFT
Equaliazer Length
SC Sym period
Page 153
Delayed Signal
Signal travels through various paths and ultimately arrives at different time
FFT symbol portion must contain integer number of cycles
Append tail part of the FFT symbol to its front part in order to make it a continuous signal
ICI & timing recovery issues if not appended
Guard interval length may or may not contain integer cycles
Page 154
Guard Time Consideration

Guard interval reduces the signal energy available at receiver Guard interval reduces the data rate throughput while increasing the noise bandwidth (spectrally inefficient) Adding a guard interval lowers the symbol rate, however it does not affects the subcarrier spacing see by the receiver
Subcarrier spacing f = Fs / FFFT
Guard interval simplifies equalization at the Rx if guard interval time is greater than the maximum delay spread Guard interval should be short (performance trade offs) Guard interval should be chosen longer than the actual RMS delay spread, 3x to 4x longer ( 0.1 of symbol length, SNR 1 dB = 10log(1- Tg/TOFDM Sym)) Guard interval is discarded by the receiver
SNR Loss, in dB = -10Log (1 - Tguard interval length / TOFDM symbol duration)
Page 155
Guard Time Effects

Insufficient guard time (CP) causes ISI & ICI
Higher order modulations & timing recovery circuits are more sensitive to ICI & ISI
16QAM 256 points FFT receive constellation plots (a, delay guard time. b delay exceeds guard time by 3% of FFT internal. c, delay exceeds guard time by 10% of the FFT interval)
a
-10
RX Const 3
b
2 1 0 -1 -2 -3
RX Const 3
c
2 1 0 -1 -2 -3
RX Const
Magnitude-squared, dB
-15 -20 -25 -30 -35 -40 0 5 10 Frequency (MHz) 15
2 1 0 -1 -2 -3 -2 0 2 In-phase Amplitude
Page 156
CP, Cyclic Prefix

Usage of CP is necessary to combat MP distortions CP reduces the BW efficiency (a tradeoff between throughput performance vs. BW) CP should be longer than the maximum expected RMS delay spread Programmable (1/4, 1/8, 1/16, 1/32), is the most robust in the multipath Delayed replicas of the OFDM symbol always have an integer number of cycles within FFT interval
A copy of the last OFDM symbol is appended to the front of transmitted OFDM symbol Actually the Tg can be realized by adding zeros, but using the CP as guard interval can transform the linear convolution with the channel into circular convolution
CP is added after the IFFT on a combined signal rather than for each sub-carrier Accommodates the decaying transient of the previous symbol Smooth initial transient to reach the current symbol Impact of CP is similar to the roll-off factor in raised cosine filtered SC systems
Page 157
Doppler Shift
A measure of spectral broadening caused by the channel time variation Motion of the mobile causes periodic phase shifts which change with time. The rate of change of phase gives rise to Doppler frequency, which varies with mobile speed and arrival angle of rays fd, Hz = v/, where v velocity, wavelength Inter-carrier spacing must be at least 10 times higher than the maximum fd
Value increases if moving toward source and lower when moving away from each other Symbol rate must be much higher than the Doppler shift. Inter-carrier spacing of the system must be chosen large, compared to the maximal Doppler frequency of the fading channel. Coherence time, Tc = 0.423/fd , fd Doppler shift TC >> T, slow fading TC T, fast fading Coherence time is a time period to correlate the channels value Coherence distance, DC = 0.179 , to determine antenna spacing
Page 158
Group Delay
It is defined as derivative of the phase response versus frequency, that is the slope of the phase response
Prime contribution of the group delay comes from tighter band pass filter response in the baseband, IF & RF sections. It is also contributed from improper cable termination, noncompensated sin(x)/x, antenna mismatch, signal combiners and multipath effects MC systems are more tolerant to group delays Equalizer is an effective tool to remove linear distortion
Page 159
OFDM Symbol
Chosen NFFT = 2^n

Carrier filler for unused carriers
Total OFDM Symbol time, Ts = (1/subcarrier frequency spacing) + Tg Tb: Useful symbol time, N/Fs
Tg or CP (to improve the impulse response) guard time

Generally kept under 1 dB (pick about 4x the delay spread) Increased roll off time reduces the spread tolerance
G = Tg/Ts, four programmable intervals (1/4, 1/8, 1/16, 1/32)

Windowing
Makes amplitude go down smoothly to zero at symbol boundaries (minimizes interference to others)
Tx signal without windowing will have wide bandwidth due to the side lobes of the IFFT being a Sinc function
Tx signal is band limited in time domain by using windowing technique (raised cosine function). There is no band limits in frequency domain
Applied window must not influence the signal in its effective period. In other words pulse-shaping affects on the CP T
Tguard Sym-1 Prefix Twin TFFT
TFFT Current Symbol Effective TX-time Postfix Twin Sym+1 Time
TX PO
BS to provide 10 dB min settable attenuation in 1 dB min step size with better than 1.5 dB (within 0.5 dB for relative steps) measurement accuracy SS to provide 30 dB (40 dB for 16e) min settable transmit attenuation in 1 dB min step size
50 dB min settable attenuation for devices with sub-channelization The measurement accuracy for 1 dB step size must be within 1.5 dB (within 0.5 dB for relative step) for the first 30 dB range, 3 dB for larger step size
Preamble level between adjacent sub-carriers must be within 0.1 dB Preamble bursts are 3 dB (4.6 dB for 16e) higher than the FCH & DL data Training symbol (for sync., estimation and tracking time-varying channels) are transmitted via preamble or pilot carriers UL pilots are not boosted
BS, TX Performance Requirements

PO +38.5 dBm/MHz max or per local regulatory requirements
Tx output noise spectral density of -80 dBm/MHz max when Tx is not transmitting Ramp up & down time of 8 sym Mod accuracy
12% for QPSK, 6% for 16QAM without equalizer

10% for QPSK, 3% for 16QAM, 1.5% for 64QAM with equalizer, linear distortion removed Symbol timing accuracy
0.02 pk-pk of nominal sym relative to previous sym over 2s duration

Tx Sym clock accuracy to be within Tx burst timing step size 0.25 of a sym 0.125 of a sym Tx burst timing step accuracy 1e-6
SNDR -31 dBc

Tx spectral mask regulated per local regulatory agency
Page 163
TX Waveform Accuracy
The accuracy of the modulated waveform is affected internally by
Root raised cosine filter length and coefficients accuracy D/A converter accuracy Modulator imbalances Synthesizer phase noise PA nonlinearities
Q Error Vector Magnitude Actual
Error
Externally affected by
Cable mismatch
Antenna mismatch Interference Terrain Environmental
Ideal Phase Error Carrier Leakage
Magnitude error I
Page 164
BS, min TX Performance

Provides TX Mod accuracy and PA linearity conditions
Tx relative constellation error is to measure with ideal receiver with carrier recovery loop BW of 1% of the symbol rate Measurement method determines the magnitude error of each constellation point at the sampling instances and RMS averages them together across multiple symbols, frame, and packets Provides signal quality prior to channel impairments at the sampling instances
OFDM Burst Type BPSK-1/2 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-2/3 64QAM-3/4 OFDMA Burst Type 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-1/2 64QAM-2/3 64QAM-3/4
Relative Cons Error for SS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -30.0 Relative Cons Error for SS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Relative Cons Error for BS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -31.0 Relative Cons Error for BS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Page 165
SS/MS, min TX Perf

+39.5 dBm/MHz or per regulatory requirements
BER at 1e-6 & carrier symbol rate of R in Mbps
-90dBm+10Log(R) for QPSK -83dBm+10Log(R) for 16QAM -74dBm+10Log(R) for 64QPSK
Transmission time from Tx to Rx, 2 us for TDD, 20 us for FDD & HD-FD ACI at 1e-6 BER & 1 dB degradation
-1 dB for QPSK, +6 dB for 16QAM, +13 dB for 64QAM
2nd ACI at 1e-6 BER & 1 dB degradation

-30 dB for QPSK, -30 dB for 16QAM, -23 dB for 64QAM
Page 166
SS/MS, min TX Perf-2

40 dB min dynamic range Tx PO of +15 dBm min for QPSK Tx PO adjustment in 0.5 dB step Tx pk-pk jitter, 0.02 of the symbol duration in over 2 s period Symbol clock to be locked on BS TX burst timing accuracy, self correction for burst step up to 0.5 of a symbol with step accuracy of 0.25 of symbol TX RF frequency accuracy, 1 ppm Spectral mask per local regulatory requirements Ramp up and down time 8 symbols Noise density, -80 dBm/MHz when not transmitting Modulation accuracy: 10% (QPSK), 3% (16QAM) 1.5% (64QAM) with equalizer distortion removed
Page 167
SS/MS, min TX Perf-3 (Flatness)
DC subcarrier to be suppressed by 15 dB min relative to the total average power from all data and pilot subcarriers
The outer subcarriers need to be within +2/-4 dB from average power transmitted from all active subcarriers
The inner subcarriers must be within 2 dB 0.4 dB The adjacent subcarriers must be within
Page 168
Po, Attenuation & Accuracy

Preamble power level may change from burst to burst Preamble aids in synchronizing the RX, perform channel estimation & Equalization processes Preamble spectral flatness is specified across all sub-carriers Po within 0.1 dB of adjacent subcarriers for both DL/UL Preamble applies to every second or 4th channel (computational adjacent) 2 dB ave over all active tones from -50 to -1 & +1 to +50, +2/-4 dB ave over all active tones from -100 to -50 and +50 to 100 for both DL/UL Preamble symbol contains no pilot 3 dB higher power than all other data subcarriers in the DL subframe Requires extremely sharp notch filters for reliable measurement. Requires 10 dB min range for BS. 30 dB min range for OFDM SS. 50 dB min range for OFDMA SS. 1 dB step with 1.5 dB min relative accuracy for 30 dB, Larger steps with 3 dB min relative accuracy for over 30-50 dB for MS/SS Power control to support 30 dB/s signal fluctuations
BS performance
TX center freq tolerance

BS 1 ppm, SS must be locked to BS, SS to be within ppm of BS 1
Tx symbol clock frequency tolerance

Rx freq & timing requirement
Time accuracy
5 to 25 us for TDD N/A for FDD GPS option (more expensive and difficult to access open sky if in the basement)
MS Performance Parameters
RSL power determined from per subcarrier level
RSL per unboosted subcarrier = RSSI-10Log(8)10Log(number of preamble subcarriers)
Fast feedback channel from MS to up date the time critical information such as CINR, MIMO, AAS, spatial multiplexing, etc.
Page 171
RX Constellation Error
Relative constellation error
Intended to ensure that RX SNR does not degrade more than 0.5 dB due to TX SNR. Measured by an ideal receiver with carrier recovery loop bandwidth of 1% of the symbol rate
EVM value includes PA nonlinearities, untracked phase noise, inband amplitude OFDM Relative Cons Relative Cons ripple and DAC inaccuracies
Results are independent of the FEC
Q Error Vector Magnitude Actual
Error
Ideal Phase Error Carrier Leakage
Magnitude error I
Burst Type BPSK-1/2 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-2/3 64QAM-3/4 OFDMA Burst Type 4QAM-1/2 4QAM-3/4 16QAM-1/2 16QAM-3/4 64QAM-1/2 64QAM-2/3 64QAM-3/4
Error for SS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -30.0 Relative Cons Error for SS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Error for BS (dB) -13.0 -16.0 -18.5 -21.5 -25.0 -29.0 -31.0 Relative Cons Error for BS (dB) -15.0 -18.5 -20.5 -24.0 -26.0 -28.0 -30.0
Page 172
Adaptive Frequency Domain Equalizer

A perfectly normal signal becomes distorted after going through multipath channel
Same frequency but different amplitude & phased signals are received Equalizer estimates the error amplitude Multiply the subcarrier by inverse phase and magnitude of the estimated channel Time domain equalization in a time dispersive channel becomes prohibitively expensive for a SC system as data rate increases. For MC, each subchannel can be modeled as flat fading channel requiring simple (N-short) Frequency Domain EQ. FDE is an attractive alternative to mitigate complexity.
OFDM moves the IFFT operation to Tx to load balance complexity between Tx 1/|a| e-j and Rx Single Single Multipath
Distorted Phase & Magnitude
An adaptive process for varied conditions Removes only the linear distortions
Transmitted Subcarrier
Channel
Received Subcarrier
Scatter Plot
TX Const 1
1.5 1
|a|ej
1
Scatter Plot
0.5
0.5
0.5 0 -0.5 -1
-0.5
-0.5
P/S
QAM DMD
FD EQL
FFT
S/P
CP Remove
FLTR ADC
-1
-1 -1 -0.5 0 0.5 In-phase Amplitude 1

-1 0 In-phase Amplitude 1
-1
-0.5 0 0.5 In-phase Amplitude
Page 173
OFDM, Rx Threshold
OFDM receive signal is made up of a sum of attenuated, phase shifted and time delayed versions of the transmitted signal
Rx Thresh=-114 -10Log(R) +10Log(FS * NUsed/NFFT) +NF + SNR + LImp
Add 10Log(Nsubchannel used/32) for OFDMA, when using less subchannels in the BS Rx NF = SNRIn / SNRO, for front end cascaded Rx chain
R, number of repetitions for the modulation/FEC rate

FS, sampling frequency in MHz
Packet vs. Frame

Packet address stays with user data up till the final destination
Link address of the frame changes at each physical device
Packet
Network (destination) Address User Data
Link Address (Destination-source addresses change along the path)
Control Info Frame
Data Payload
Pad
CRC
Page 175
FER, Frame Error Rate

Error rate, a quality metrics for data applications Performance dependent on the frame length & BER Receiving equipment discards the entire packet upon receiving error(s) and requests retransmission Lower data throughput vs. BER Further reduction in data throughput due to retransmission FER1 = BER * (1-BER)^FR*8-1 * (FR*8), for 1 bit error per frame FER2 = BER2 *(1-BER)^FR*8-1 *(FR*8)*(FR*8 -1)/2, for 2 bits error per frame For example, FER1&2 for frame rate of 64 bytes FERActual = 1 - # of non-errored frames Frame received / # of frames transmitted BER 1e-6 1e-7 1e-8 1e-10 1e-12
FER1 5.12e-4 5.12e-5 5.12e-6 5.12e-8 5.12e-10
FER2 1.31e-7 1.31e-9 1.31e-11 1.31e-15 1.31e-19

Page 176
RX Requirements
Minimum threshold requirements based on 7 dB NF & 1e-6 BER
Residual bit error rate to be 1e-10
Bandwidth (MHz) 1.5 1.75 3 3.5 5 6 7 10 12 14 20 Rx SNR, dB

Rx Threshold vs. Modulation & Coding rate, dBm BPSK QPSK QPSK 16QAM 16QAM 64QAM 64QAM 1/2 1/2 3/4 1/2 3/4 2/3 3/4 -94 -91 -89 -84 -82 -78 -76 -93 -90 -87 -83 -81 -77 -75 -91 -88 -86 -81 -79 -75 -73 -90 -87 -85 -80 -78 -74 -72 -89 -86 -84 -79 -77 -72 -71 -88 -85 -83 -78 -76 -72 -70 -87 -84 -82 -77 -75 -71 -69 -86 -83 -81 -76 -74 -69 -68 -85 -82 -80 -75 -73 -69 -67 -84 -81 -79 -74 -72 -68 -66 -83 -80 -78 -73 -71 -66 -65 3.0 5.0 8.0 10.5 14.0 18.0 20.0
Page 177
SS/MS Power Restraint
BS configures each SS/MS PO such that the Rx power arriving at the BS to remain constant and consistent with all others regardless of the distance
Receiver front end must be able to tolerate high incoming signal level demanding linearity with higher IIP3 devices
For direct conversion system, IIP3 demand further increases due to lack of sharp IF filtering and limited AGC range
Page 178
Fade Mitigation
Narrow band system

Time diversity Freq diversity Diversity type interactions
Wide band system

Equalization
Page 179
OFDM, Benefit Summary, (1)

High spectral efficiency
Simple implementation by FFT, modulate by switching between time and frequency domain
Lower Rx complexity as Tx combat the channel effect to some extends

Resilient to ICI, ISI (by increasing symbol time)
Immunity to delay spread and resilient to MPF

Equalization is simplified or eliminated altogether
Suitable for high data rate transmission Highly flexible in term of link adaptation Low complexity multiple access (OFDMA) Mod/code change on frame to frame and SS to SS depending on robustness (trade-off cap vs. robustness in real time) QoS based on latency, jitter & reliable throughput Channel impairments and timing problems are both solved with simple phase and channel estimators
Page 180
OFDM, Benefit Summary, (2)

Interoperability
Higher subchannel count smaller guard band
Robust against narrow band interference

Suitable for coherent demodulation TDD, FDD or half FDD NLOS
Simple equalization. EQ complexity B*Log(BTg) vs. SC is B^2 * Tg Frequency diversity capable

Graceful degradation due to delay spread (ideal for AMC)
Easier time-frequency synchronization

No inter-carrier guard band
Multi-access using OFDMA
Resistance to frequencyselective fading

Page 181
OFDM, Disadvantages
Synchronization
Requires more complex algorithms for time / frequency sync Additional circuit for FFT and IFFT is needed Greater complexity More expensive Tx & Rx Reduced efficiency due to guard interval Sensitive to phase noise, timing & frequency offsets Tight specifications for local oscillators Doppler limitation High peak to average ratio (PAPR) Approximately 10 Log (N), in dB
Large signal peaks require higher power amplifiers

Amplifier cost grows nonlinearly with required power Need very linear amplifiers with large dynamic range
OFDM Snapshot
8 BPSK pilots at fixed location 192 data subcarriers, 55 null subcarriers and 1 DC subcarrier OFDM Symbol = (1+Cyclic Prefix)/f f (Sub-carrier spacing) is proportional to Channel BW/FFT size Sub-channel spacing varies according to the BW For narrow BW, sub-channel spacing becomes closer that makes the symbol time longer yielding better performance in NLOS channel
Active Subcarriers: 200 Subcarrier Spacing: 90 kHz 8 BPSK pilots Fixed location BPSK, QPSK, 16QAM, 64QAM
Guardband
DC
20 MHz BW 14 MHz o o o 2.5 MHz 1.25 MHz BW

Page 183
Active Subcarriers: 200 Subcarrier Spacing: 5.6 kHz
8 BPSK pilots Fixed location
OFDMA
Combination of FDMA and OFDM. No guard band between subcarriers.
FFT size is scalable from 128 to 2048 Fill unused channels with null subcarriers to bring up to next 2N Increase the FFT as the BW increases such that subcarriers spacing remains 10.94 kHz (depends on configurable over-sampling rate) Keeps constant symbol duration and have minimal impact on higher layers Sub-carrier spacing can support delay spread up to 20 us, 125 kmph at frequency 3.5 GHz 4QAM,16QAM & 64QAM are used for data. BPSK is used during preamble, pilot & when modulating subcarriers in the ranging channel
OFDMA, Example
Sub-carrier separation remains constant regardless of the BW FFT and Subcarriers (Data, Pilot & Null) vary with the BW Number of OFDM symbols remain constant regardless of BW in a specific frame FFT: 2048 rate 166-240 QPSK Pilots Active Carriers: 1680 QPSK, 16QAM, 64QAM
Subcarrier Spacing: 11.161 kHz Fixed & variable location
20 MHz FFT: 1024 82-120 pilots Active Carriers: 840 Subcar. Spacing: 11.161 kHz
10 MHz FFT: 512 42-60 pilots Active Carriers: 420 Subcarrier Spacing: 11.161 kHz
5 MHz DC FFT: 128 10-16 pilots Active Carriers: 84 Subcarrier Spacing: 11.161 kHz
1.25 MHz
OFDMA
In OFDM, only one MS is transmitted in one time slot
In OFDMA, several MSs can be transmitted in the same time slot over several sub-channels
Time-frequency allocations are done dynamically to improve performance at the expense of complexity
OFDM
OFDMA
Subcarriers, frequency
Subchannels, frequency
User 1 User 2 User 3 User 4 Time

Page 186
FFT symbol
Time
OFDMA Symbol Parameters

Adds additional multiple access features in the frequency domain BW is divided into slots for the user in the time and the frequency domain OFDMA carriers for different users are very close together (10kHz) & that the order of the physical carriers may change from Symbol to Symbol Difficult to design variable subcarriers spacing. This is mitigated by using FFT size vs. BW
Parameter System bandwidth, MHz Sampling frequency, MHz Sample time, us FFT size # of used data subcarriers # of pilot subcarriers # of null/guard subcarriers Subcarrier spacing, kHz Useful symbol time, us Useful symbol BW, MHz Fixed Mode Mobile Mode BW 3.5-28 1.25 2.5 5 10 20 Fs=8/7*BW 4.000 1.429 2.857 5.714 11.429 22.857 Ts=1/Fs 0.250 0.700 0.350 0.175 0.088 0.044 N 256 128 256 512 1024 2048 NData 192 72 180 360 720 1440 NPilot 8 12 30 60 120 240 NNull/Guard 56 44 46 92 184 368 f=Fs/N 15.625 11.160714 Tb = 1/f 64 89.6 (exact) f*(Ndata+Npilot) 3.125 0.9 2.3 4.687 9.4 18.7
Available guard time settings Tg = Guard time, us Tg OFDMA symbol time, us Ts=Tb+Tg TTG+TRG, us PS=4/Fs 10/TS Symbol per 10 ms frame
12.50% 8 72 1.000 137
Tb/4 22.4 112.0 2.800 87
Tb/8 11.2 100.8 1.400 98
Tb/16 5.6 95.2 0.700 104
Tb/32 2.8 92.4 0.175 108

Page 187
IEEE 802.16-2005 OFDMA Physical Layer Parameters
Modulation
Error correction code Overall coding rate Cyclic Prefix Subchannels
QPSK, 16-QAM, 64-QAM

CC, BTC, CTC , , 2/3 1/32, 1/16, 1/8, 1, 2, 4, 8,16, 32
Bandwidth
FFT
1.25, 2.5, 5, 10, 20

128, 256, 512, 1024, 2048
Introduction to WiMAX Technology 188
Page 188
OFDMA Frame Structure for TDD, Example

No UL preamble at start of UL subframe but an increased number of pilots. Pilots in the UL are never transmitted without data subcarriers SSs use PN CDMA technique to access the BS in the contention region Pilot and null subcarriers are not shown
Page 189
OFDMA Frame
A frame is one complete set of DL & UL transmission, meaning the time between two preambles of the DL signal
Frame consists of DL & UL subframe with flexible boundaries
PS (physical slot) is a unit of time defined as 4 modulationSymbol length

FS, sampling freq. = FFT-size * channel-spacing
UL has no preamble except for system using AAS, but there are increased number of pilots. Data is transmitted in bursts that are as long as the UL sub-frame zone allows and wrapped to further sub-channel as required.
Page 190
OFDM Frame Structure, Example

OFDM symbol number (time) Time
... ... k+31 k+33
Adaptive & variable length duration for DLMAP, UL-MAP, fast feedback, ranging and data burst
Subcarrier (frequency)
k,k+1 k+2 k+3 k+4 k+7 k+9 k+11 k+13
...
k+17 k+20 k+23
UL-MAP (conti'd)
UL burst #2
Preamble
DL-MAP
UL burst #4 DL burst #4
UL-MAP UL-MAP
UL burst #5 DL burst #5 Fast Feedback Ranging

DL subframe TTG UL subframe RTG
S+L
Preamble
Page 191
Pilot and null subcarriers are not shown
DL burst #1
DL burst #3
UL burst #3
DL-MAP
S S+1 S+2
FCH
FCH DL burst #2 UL burst #1
OFDMA Subchannels, (1)

A subchannel describes the smallest logical allocation unit in the frequency domain. It contains one or more physical carriers, which are adjacent or non-adjacent and whose order may change within a burst from symbol to symbol. Subchannelization is a sophisticated form of frequency division multiple access where multiple subcarriers are grouped into subchannels to enhance system performance. (The number of subchannels varies from 32 to 96, depending on the zone type) Subchannel is the basis of OFDMA Multiple Access Method
Frequency space is divided into subchannels, i.e. Group of subcarriers forms subchannel Dynamically allocating time-frequency resources to DL/UL subframe At certain moment subchannel is utilized by one transmitter only Basic OFDMA time-frequency unit utilized for communication is determined through subchannel-OFDMA symbol combination WiMAX uses the term slot for minimum data allocation unit and a slot contains 48 data subcarriers
Zone and Burst

Zone (contains bursts) A zone is one complete logical part of a frame. There are DL and UL zones, and there are different zone types that may use all subchannels of the OFDMA frequency range (full usage of subchannels = FUSC) or only parts of them (partial usage of subchannels = PUSC).
Grouping of contiguous symbols that use a specific type of subchannel assignment

All zones except for AMC use the distributed allocation of subcarriers for subchannelization OFDMA PHY specifies 7 different zones: FUSC, OFUSC, PUSC, OPUSC, AMC, TUSC1 and TUSC2 Burst (contains slots) A burst is an area within a zone which is assigned to one dedicated user. It uses a certain number of subchannels (frequency) and a certain number of symbols (time).
Page 193
Slot
Slot
A slot is the minimum possible data allocation unit within OFDMA, defined in time and frequency (number of contiguous symbols times number of subcarriers). It always contains one subchannel and can contain one to three symbols (depending on the zone type). A DL-PUSC slot is two symbols wide, a UL-PUSC slot three symbols wide. Minislot
A unit of UL BW allocation equivalent to n physical slots, where n=2^m, m is an integer ranging from 0 through 7
Page 194
OFDMA Subchannels Terms
OFDMA Symbol Number

k+0 k+1 k+2 k+3 k+4 k+5 k+6 k+7 k+8 k+9 k+10 k+11 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 k+12 k+13
Time
k+14
Slot
Subchannel offset
Subchannel Logical Number
Sym offset Subcarriers
# of OFDMA symbols
Slot
Data Region
Segment
Permutation zone
Page 195
Mandatory and Optional Zones

In OFDMA PHY, the mapping from data bit to physical subcarriers is achieved in two steps: The 1st step is to map the data to one or more time slots and one or more logical subchannels The 2nd step is called permutation, in which the logical subchannels are mapped to physical subcarriers Multiple permutation zones marked by Zone Switch IEs (AAS_DL_IE, AAS_UL_IE, STC_DL_Zone_IE) Switching from Non-STC to STC, and Non-AAS zones is defined by the IEs
DL Subframe
PUSC (DL_PermBase X) FUSC (DL_PermBase Y) FUSC (DL_PermBase Z) PUSC (1 zone contains FCH & DL-MAP) Optional FUSC
UL Subframe
Optional PUSC
Preamble
TUSC1
TUSC2
PUSC
AMC
st
Must appear in every frame Zone switch IEs in DL-MAP

May appear in frame

Page 196
AMC

Like an OFDM, OFDMA symbol contains subcarriers Subchannel means splitting a normal channel BW into more than one Subchannel contains a group of subcarriers User is assigned one or more subchannel 1, 2, 4, 8, 16 or 32 subchannels for UL. Two of the UL subchannels are used for ranging and BW request Link budget improvement in UL for fixed WiMAX, 12 dB when used 1/16th of BW Power concentration on few subcarriers Reduces over head control algorithm complexity Subcarriers of a subchannel can be contiguous or distributed over the BW Randomly distributed type provides frequency diversity in frequency-selective fading channels and inter-cell interference averaging. Well suited for mobile applications Adjacent type is useful for frequency non-selective and slowly fading channels, and for implementing ACM. Typically used for fixed or low mobility applications. Channel estimation is easier as the subcarriers are adjacent Well suited for fixed and low mobility applications

Basic principle is to trade off mobility for throughput
Subchannels are dynamically allocated to users for UL & DL data based on CINR
Subcarriers are randomly assigned to subchannel and changed every symbol time
Different subchannel allocation methods to
Optimize the frequency band for stationary or mobile usage
High or low interference from neighboring sectors, cells Optimize diversity and beamforming techniques performance
16e system uses subchannels in both directions whereas 16d applies in the UL only
Subchannels, Power Concentration

Increases the Tx power & improve asymmetric link budget on cost prohibitive CPE OFDMA example: 5 MHz BW, 480 active subcarriers, 200 mW CPE PO, subchannelization assigns 120 subcarriers to CPE PSD = PWR/BW = 200mW/5M =40 uW/Hz for all subcarriers PSD = PWR/BW = 200mW/1.25M =160 uW/Hz for 120 subcarriers, yielding 6 dB gain, 10Log(480/120)
1/2 of tones yield 3 dB power gain 1/4 of tones yield 6 dB power gain 1/8 of tones yield 9 dB power gain 1/16 of tones yield 12 dB power gain 1/32 of tones yield 15 dB power gain
DL Bandwidth
All-ch Po
UL All Sub-channels
4x sub-ch Po
UL 1/4 Active Subchannels

OFDMA Subcarriers Mapping into Subchannel

Distributed allocation
Subchannel1 DC subcarrier Subchannel 3 Subchannel N
Subchannel2
guard band
guard band DC subcarrier Subchannel N-1 Subchannel N
Adjacent Allocation
Subchannel1
Subchannel2
guard band
Introduction to WiMAX Technology 200
guard band
Page 200
Subchannelization
Very effective under stressful channel conditions
Page 201
Subchannelization
Narrowband interference rejection Easy to avoid/reject narrowband dominant interference
Less interfered part of the carrier can still be used

User Subcarriers Allocation Interference Subcarriers
Before
User Subcarriers Allocation
After
Null Subcarriers
Interference Subcarriers
Total Frequency Band

FUSC, Full Usage of Subcarriers

DL FUSC
Fixed & Variable pilot tones are added for each OFDMA symbol independently One set of common pilot subcarriers and always at the same location Remaining subcarriers are divided into subchannels that are used exclusively for data User are allocated slots for DL data transfer One slot is a single subchannel with 48 subcarriers by one OFDMA symbol (48 tone-symbols)
Provides full frequency diversity and inter-cell interference averaging by spreading the subcarriers over the entire band
BW (MHz) 1.25 5 10 20 FFT Size 128 512 1024 2048 AMC Subchannels (downlin/Uplink) (downlink/Uplink) 2/2 8/8 16/16 32/32 PUSC Subchannels (downlin/Uplink) (downlink/Uplink) 3/4 15/17 30/35 60/92 FUSC Subchannels (downlink only) 2 8 16 32
Page 203
PUSC, Partial Usage of Subcarriers

DL & UL PUSC
The set of used subcarriers is partitioned into subchannels Groups the sub-carriers into tiles to enable fractional frequency reuse scheme (FFRS) Pilot subcarrier are allocated from within each subchannel Each subchannel contains its own set of pilot subcarriers User are allocated slots for DL/UL data transfer For DL, one slot is a single subchannel by two OFDMA symbols DL-PUSC slot uses a cluster structure. One subchannel contains two cluster. One cluster contains 12 data subcarriers and two plot subcarriers For UL, one slot is a single subchannel by three OFDMA symbols One subchannel contains 6 tiles. One tile contains 4 subcarriers. One tile with three symbols contains 4 pilots subcarriers and 8 data subcarriers One subchannel symbol consists of 48 tones minimum Provides frequency diversity function (minimizes interference between Inter-cell as well as between adjacent sectors)
Page 204
OFDMA Permutation
1.25 MHz BW 105 96 85 72 108 96 108 96 97 5 MHz BW 426 384 421 360 432 384 432 384 408 272 433 432 384 10 MHz BW 851 768 841 720 864 768 864 768 840 840 865 864 768 20 MHz BW 1702 1536 1681 1440 1728 1536 1728 1536 1681
OFDMA Permutation DL FUSC No. of subcarrier No. of data subcarrier DL PUSC No. of subcarrier No. of data subcarrier DL O-FUSC No. of subcarrier No. of data subcarrier DL O-AMC No. of subcarrier No. of data subcarrier UL PUSC No. of subcarrier No. of data subcarrier UL O-PUSC No. of subcarrier No. of data subcarrier Ul O-AMC No. of subcarrier No. of data subcarrier
109 108 96
1729 1728 1536
Page 205
PUSC, Example
PUSC with STC
DL-PUSC Subframe with STC Zone (2 Antennas UL-PUSC Subframe with STC Zone
User 1 Matrix B
User 2 Matrix B
User 1, Matrix A, 2 Tx Antennas
User 1, Matrix A
User 1 None, 1 Tx Antenna
User 3 Matrix A User 2 Matrix A
User 3, Matrix B Unallocated
User 2 Matrix B, 2 Antennas Unallocated OFDMA Symbol Number
OFDMA Symbol Number
Page 206
OFDMA Highlights
16e intended for mobile broadband connection for pedestrians and automobiles in 1-3 mile radius range
Maximum subscriber throughput 3 Mbps per DL, 1 Mbps per UL
Maximum sector throughput (10MHz band) Frequency reuse 18 Mbps per DL, 6 Mbps per UL 1
Mobility
Handoff Service coverage Roaming QoS offering Uplink/Downlink ratio
Up to 120 km/h
Under 150 ms Macro (1km), Micro (400m), Pico (100m) Seamless roaming with cellular and WLAN Unsolicited grant service, extended realtime, real time, non real-time, best-effort Software adjustable
Page 207
OFDMA Data Rates

Mod Code rate 5MHz 5MHz 10MHz 10MHz
DL rate Mbps
4QAM CTC, 6x CTC, 4x CTC, 2x CTC, 1x 3/4 CTC 16QAM CTC CTC 64QAM CTC 2/3 CTC 3/4 CTC 5/6 CTC
DL rate Mbps
0.53 0.79 1.58 3.17 4.75 6.34 9.50 9.50 12.67 14.26 15.84
UL rate Mbps
0.38 0.57 1.14 2.28 3.43 4.57 6.85 6.85 9.14 10.28 11.42
DL rate Mbps
1.06 1.58 3.17 6.34 9.50 12.07 19.01 19.01 26.34 28.51 31.68
UL rate Mbps
0.78 1.18 2.35 4.70 7.06 9.41 14.11 14.11 18.82 21.17 23.52
Page 208
OFDMA Advantages-Summary, (1)

Enables adaptive modulation for every user using QPSK, 16QAM & 64QAM
Performs adaptive FEC based on CINR, RSSI & error rate
Enables dynamic subcarrier allocation

Efficient use of air resources for mobile applications
Enables spatial diversity by using antenna diversity at the Base station and possible at the Subscriber Unit Gives frequency diversity by spreading the carriers all over the used spectrum
Gives time diversity by optional interleaving of carrier groups in time


Using the cell capacity to the outmost by adaptively using the highest modulation a user can use, this is allowed by the gain added when less subcarriers are allocated, therefore gaining in overall cell capacity
The power gain can be translated to distance
Doubles the distance for each 6 dB gain in LOS conditions
Enabling the usage of indoor Omni Directional antennas for the users
MAC complexity is the same as for TDMA systems
Page 210

Allocating carrier by OFDMA/TDMA strategy
Minimal delay per OFDMA symbol of 300us Using small burst per user of about 100 symbols for better statistical multiplexing and smaller jitter User symbol is several times longer than for TDMA systems Using the FEC to overcome error in disturbed frequencies Adaptive modulation and coding
OFDM is a proven technology for transporting high data rates for NLOS, long ranges with multipath conditions like for DVB-T, DAB etc.
Drawbacks of OFDMA
OFDMA signal due to its extreme amplitude variation does not behave well in non-linear channel. PAPR, dB = 10Log(N), OFDM composite signal exhibits significant peaks and valleys (when all carriers add in phase) with depths of more than 50 dB but the probability of its occurring is low due to scrambling and higher number of N. Some uses advanced coding techniques to minimize its effect. Higher N means higher peak, requiring linear & expensive PA. Increases ADC/DAC complexity. Highly sensitive to timing jitter and frequency offsets Doppler limitation Susceptible to phase noise as each subcarrier is Mod by phase noise of the LO Loss due to guard band, typically set 1 dB Reduced channel spacing at higher N increases chance of ICI Sync: requires more complex algorithms for time/frequency sync
AMC, Adaptive Modulation & Coding, (1)

Uses adjacent sub-carriers for each subchannel for use with beam forming Modulation, power and or Coding change on a burst by burst basis per link Channel response remains flat over narrow subcarrier A closed loop control process TX controls the Capacity, QoS, ECC, symbol mapping and power Rx feeds back the current SINR, BER & RSL information Basic idea is to transmit as mush data as possible and throttle down when channel is not good AMC for OFDMA: each user is allocated a group of subcarriers, each having different SINR. Care needs to be paid in selecting Mod-coding based on varying SINR across subcarriers DL burst goes to one or more SS using the same mod & coding. The UL burst comes from individual users with individual devices. The SS are told when to transmit. Demands fast settling response from PA
256QAM
64QAM SNR = 22 dB 16QAM SNR = 16 dB
Average SNR
64QAM 16QAM 4QAM BPSK
Not used due to low SNR
4QAM SNR = 9 dB BPSK SNR = 6 dB

Page 213
Subcarriers

Down Link
Modulation BPSK, QPSK, 16QAM, 64QAM BPSK optional for OFDMA-PHY
Up Link
Modulation BPSK, QPSK, 16QAM 64QAM optional
256QAM optional
Adaptive power back-off per Modulation Coding
Adaptive power back-off per Modulation

Coding Mandatory: concatenated convolutional codes at rate , 2/3, , 5/6 Optional: convolutional turbo codes at rate , 2/3, , 5/6. Repetition codes at rate , 1/3, 1/6 & LDPC Capacity, QoS and power back-off management
Page 214
Mandatory: concatenated convolutional codes at rate , 2/3, , 5/6

Optional: convolutional turbo codes at rate , 2/3, , 5/6 & LDPC Capacity, QoS and power back-off management

A closed loop controlled process
Affects on incoming data queuing, FEC, Mod and PO in the transmit direction
Dynamically adapts the transmitting signal based on channel status (RSL, CINR & PER) from the far end receiver
Assigns adjacent subcarriers to specific SS/MS
Improved performance in OFDMA subchannelization

Transmitter
Bits In
Receiver
4
ECC Encoder
Select Code
Symbol Mapper
Select Constellation
PO
Power Control
Channel SINR
Bits Out
Demod
Decoder
Queue
Adaptive Modulation and Coding Controller

Feedback Channel RSL, SINR, PER
Channel Estimation
Page 215
PHY Parameters
Table assumes 5 ms frame rate and a Tg 12.5% (1/8) of Tb
Fixed Mobile WiMAX Scalable Parameter WiMAX OFDMA-PHY OFDM-PHY FFT size 256 128 512 1024 2048 Number of used data subcarriers 192 72 360 720 1440 Number of pilot subcarriers 8 12 60 120 240 Number of null/guardband subcarriers 56 44 92 184 368 Cyclic prefix or guard time (Tg/Tb) 12.5% 1/32, 1/16, 1/8, 1/4 Oversample rate Fs/BW Depends on bandwidth: 7/6 for 256 OFDM, 28/25 8/7 for OFDM, 28/25 for OFDMA 8/7 for multiples of 1.75 MHz, 2 MHz or 2.75 MHz Channel bandwidth (MHz) 3.5-28 1.25 5 10 20 Subcarrier frequency spacing (kHz),F 15.625 10.94 Useful FFT symbol time (us), Tfft=1/F 64 91.4 Guard time assuing 12.5% (us), Tg 8 11.4 OFDM symbol duration (us), T=Tfft+Tg 72 102.9 Number of OFDM symbol in 5 ms frame 69 48
PHY Estimated Data Rate vs. BW
Parameter Channel Bandwidth (MHz) FFT size Oversampling Rate Mod & Code Rate BPSK, 1/2 QPSK, 1/2 QPSK, 3/4 16QAM, 1/2 16QAM, 3/4 64QAM, 1/4 64QAM, 2/3 64QAM, 3/4 64QAM, 5/6
Fixed WiMAX OFDM 3.5 256 8/7 DL 946 1882 2822 3763 5645 5645 7526 8467 9408 UL 326 653 979 1306 1958 1958 2611 2938 3264
Mobile WiMAX Scalable OFDMA 1.25 5 10 128 512 1024 28/25 28/25 28/25 PHY-Layer Data Rate (kbps) DL UL DL UL DL UL Not applicable 504 154 2520 653 5040 1344 756 230 3780 979 7560 2016 1008 307 5040 1306 10080 2688 1512 461 7560 1958 15120 4032 1512 461 7560 1958 15120 4032 2016 614 10080 2611 20160 5376 2268 691 11340 2938 22680 6048 2520 768 12600 3264 25200 6720 8.75 1024 28/25 DL 4464 6696 8928 13392 13392 17856 20088 22320 UL 1120 1680 2240 3360 3360 4480 5040 5600
Page 217
Interference, (1)
Presence of one or more undesirable signals that degrades a normal performance Potential source could be from its own internal, external equipment or both Non-optimized path, network and frequency planning Sharing dual pole antenna & crossed pole interference
Low antenna XPD

Co-located antennas, low discrimination antenna & sector spill over Most interference detection tests are traffic affecting Maintenance activity may introduce interference EMI and interference from co-located equipment ATPC helps minimizing the interference affect OFDMA spreads the energy of an impulse noise over an OFDMA burst that results in smaller increased noise rather than losing symbols Triple transient signal interference due to poor termination, return loss & simple Frequency Domain Equalizer
Page 218
Interference, (2)
Ways to check interference

Easier to test prior to commissioning Scan RX frequency & power facing toward remote with spectrum analyzer Difficult to detect intermittent types of interference TX fade test with built-in attenuator Capturing and plotting RSSI & CINR over time Mute remote adjacent/opposite polarity TX and check RSSI & PER Mute remote opposite polarity TX and check RSSI, CINR & PER Mute remote TX and check RSSI, CINR & PER Change polarization and check RSSI & BER Change operating frequency within the band and check RSSI, CINR & PER
Interference, (3)
Acceptable interference level = 1e-6 Threshold - T/I - MEA
T/I, ratio of the RX threshold and interference vs. Freq offset Interference level not to cause more than 1 dB degradation at 1e-6 BER Using same bandwidth and type of signal MEA maximum number of exposure allowed
The terms 1e-3 BER threshold, 3 dB degradation, C/I & MEA are more relevant to analog radios Determining threshold degradation at specific interference level
10Log[1+10(I-N)/10]
I interference level in dB
N noise level in dB
Noise & Interference Relationship

S
Thermal Static Fade Margin
Unfaded RF RX Level
T
1dB
1e-6 BER with I+N 1e-6 BER Threshold with Noise
SIR T/I N
SNR
SINRReq
Noise Floor + Interference Noise Floor
6 dB, objective for 1 dB threshold degradation
I
Interference Level
Page 221
PCINR, Physical CINR

A quick and accurate CINR estimation information is required from the SS in order for BS to select an appropriate modulation coding scheme for that SS I+N conditions over a symbol or channel vary rapidly therefore it is important to estimate both average and instantaneous CINR SS/MS measures the CINR from DL preambles and reports back to BS in REP-RSP message C/N, Carrier to thermal noise ratio C/(N+I), Carrier to thermal noise plus interference ratio
Interference level that causes 1 dB degradation During non-boosted data subcarriers
PCINR:
PCINR = (3/8 *Cpmbl) / (3/8 * Ipmbl + N) = Cpmbl / (Ipmbl + 8/3 * N)
Cpmbl & Ipmbl power measured during preamble, 3/8 to scale down due to preamble
ECINR, CINR estimation based on the pilot sub-carriers

MAC, Media Access Control Overview (1)

Designed for Point-to-Multipoint based on collision sense multiple access with collision avoidance (CSMA/CA listen before transmit) Targeted for Metropolitan Area Network applications Connection-oriented MAC
Connection ID (CID), Service Flows(SF)
Supports difficult user environments

High bandwidth, hundreds of users per channel Continuous and burst traffic
Very efficient use of spectrum Protocol-Independent core (ATM, IP, Ethernet,...) Balances between stability of contentionless and efficiency of contention-based operation Data control plane (traffic scheduling to provide QoS) with speed up to 268 Mbps each way Supports multiple 802.16 PHYs
Page 223
MAC, Media Access Control Overview, (2)

Interfaces between higher transport layers and PHY. MAC layer takes packets from upper layer. These packets are called MAC service data units (MSDUs) and organizes into MAC protocol data units (MPDUs) for transmission over the air. For Rx it performs reverse process. MAC includes convergence sub-layer that can interface with a variety of higher layer protocols such as ATM, TDM, voice, Ethernet, IP and other unknown future protocols. Beside providing a mapping to and from the higher layer, the convergence sub-layer supports MSDU header suppression to reduce the higher layer overhead on all packets.
Supports very high peak bit rates while delivering QoS similar to ATM and DOCSIS (data over cable service interface specifications).
Uses variable length MPDU and offers a lot of flexibility to a lower layer for their efficient transmission (multiple MPDUs of the same or different length may be aggregated into a single burst to save PHY overhead. Conversely, large MSDUs may be fragmented into smaller MPDUs and sent across multiple frame)
MAC, (1)
Management messages: FCH, DL-MAP, UL-MAP, DCD, UCD, Data.
FCH frame control header, DCD downlink channel description (PHY characteristics. DL-frame prefix (24 bits)
OFDM MAC is designed for efficient use of spectrum
Very high bit rate, DL & UL broadband services

Supports multi-services simultaneously with full QoS
Efficient transport IPv4, IPv6, ATM, Ethernet, VLAN, etc.

Flexible QoS offerings
CBR (unsolicited grant service, highest), rt-VBR, ert-VBR, nrt-VBR, BE (lowest), with granularity within classes
QoS per user and per connection basis
Protocol-independent engine
Convergence layers to ATM, IP, Ethernet, ...

Extensive & strong security types encryption/decryption (DES, 3DES, AES, RC4, data encryption standards)
MAC, (2)
ARQ, Automatic Repeat Query Done at MAC layer rather than at TCP layer 4 (yields less outage) Adds error detection ability in data stream Bad packets are retransmitted Detecting errors using CRC-32 codes Not efficient in broadcast systems Not used in voice services OFDM/OFDMA support Dynamic Frequency Selection For license-exempt applications Adaptive antenna system support Mesh mode Optional topology for license-exempt operation Subscriber to subscriber communications Complex topology and messaging, but: Addresses license-exempt interference Scales well Alternative approach to non-line-of-sight
MAC, (3)
Packet classification IP and Ethernet
DSCP / TOS (any bit) Source / destination MAC and / or IP Source / destination Port ranges
Packet convergence sublayer support for:

IPv4 and IPv6 Packet IPv4 & IPv6 over 802.3 (Ethernet)
Dynamic service flow creation BS/MS initiated

PHS (packet header suppression) & ROHC (robust header compression Rel 4.x) PKMv2 (privacy key management) support for
EAP based authorization Cryptographic suites (for data encryption, CCM-mode 128 bit AES key)
MAC Operation
Page 228
MAC Requirements
Provide Network Access Address the Wireless environment
e.g., very efficient use of spectrum
Broadband services
Very high bit rates, downlink and uplink
A range of QoS requirements Ethernet, IPv4, IPv6, ATM, ...
Likelihood of terminal being shared

Base station may be heavily loaded
Security Protocol-Independent Engine

Convergence layers to ATM, IP, Ethernet,...
Supports for both TDD and FDD in PHY

Physical Layer Convergence Procedure (PLCP Packet Structure 802.11)

PLCP Preamble
Short training symbol t1 to t10 Long training symbol (T1 & T2)
PLCP Signal field

PLCP Data The training length period varies for 802.16 due to varied IFFT size
8+8 = 16 us, Preamble
10x0.8 = 8 us t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 GI2 2x0.8 + 2x3.2 = 8 us T1 T2
0.8+3.2 = 4 us
0.8+3.2 = 4 us 0.8+3.2 = 4 us
GI Signal GI Data 1 GI Data 2
Coarse Freq Signal Detect, AGC, Diversity Selection Offset Estimation Timing Sync
Channel and Fine Frequency Offset Estimation
Rate Length
OFDM Service+Data
Page 230
MAC Management Messages, (1)

Connection orienteded For each direction, connection identified with a 16 bit CID Each CID is associated with a Service Flow ID (SFID) that determines the QoS parameter for that CID Admission control plane (ensures that new flows do not degrade the quality of established flows) Channel access: UL-MAP Defines uplink channel access Defines uplink data burst profiles DL-MAP Defines downlink data burst profiles QPSK-1/2, Defines allocated data regions for UL-MAP Defines UL BW-request, Ranging, CQICH ... regions Defines allocated data regions for SS/MS DL/UL reception/transmission Defines multiple permutation zones (if present) UL-MAP and DL-MAP are both transmitted in the beginning of each downlink subframe (FDD and TDD)
MAC Management Messages, (2)

From BS: Preamble>FCH>DL-MAP>UL-MAP>DCD>UCD>Data> Data...
Preamble: Provides fixed known pattern to aid in Rx timing recover
FCH: The FCH specifies the burst profile and length of one or more bursts that follows the FCH. DL_Frame Prefix (24 bits) FCH: In fixed WiMAX, FCH is transmitted at the lowest mod and highest coding rate. BPSK rate and occupies 1st two subcarriers FCH: In mobile WiMAX, FCH transmission is repeated for robustness. QPSK rate, repetitions coding of 4 and occupies 1st two subchannels
DL-MAP: describes the DL allocations, PHY sync info, BS identifier, DCD identifier that is used in the allocation. UL-MAP: describes the UL allocations. Consists of UL Ch ID, UCD (UL Ch descriptor) identifier, start time, PHY specific UL-MAP elements that define allocations. Configurable fixed duration frames
QoS
Advanced QoS features:
Weighted fair queuing Traffic shaping
Congestion management
Random early detection Hierarchical QoS
Page 233
DL BW Calculation, Example
input chan size (MHz) Calculates frame O/H Calculates DL map O/H Calculates MAC hdr O/H
input cyclic prefix ( to 1/32)
Calculates frame BW
Calculates UL map O/H
Calculates MAC subhdr O/H
input mod and coding distribution
Calculates preamble O/H
Calculates useful frame BW
Calculates MAC CRC O/H
Calculates useful Channel BW
Calculates FCH O/H
input avg user pk (B)
Calculates useful MAC BW
input frame length (2 to 20 ms)
Page 234
UL BW Calculation, Example
input chan size (MHz) Calculates frame O/H Calculates contention O/H input avg user pk (B)
input cyclic prefix ( to 1/32)
Calculates frame BW
input burst sizee
Calculates MAC hdr O/H
input mod and coding distribution
Calculates ranging O/H
Calculates Subchannel O/H
Calculates MAC subhdr O/H
Calculates useful Channel BW
input subchannel size
Calculates Preamble O/H
Calculates MAC CRC O/H
input frame length (2 to 20 ms)
Calculates MAU
Calculates useful frame BW
Calculates useful MAC BW
Page 235
Allocating Network BW, Example

Calculate user Channel BW Calc remaining Channel BW Calc BE MS% remaining BW
Allocate CBR BW
Calculate VBR MS BW
Allocate VBR MR BW
Calc remaining Channel BW
Calculate BE MS BW
Allocate BE MS BW
Allocate VBR MR BW
Calc VBR MS % remaining BW
No remaining BW done
Page 236
Power Saving Modes

Three types of subscriber power management support
Normal operation Sleep mode Idle with paging support
Page 237
PHY Operating Modes, Optional

Sleep
MS with active connection temporarily disrupt connection for a predetermined time followed by a listen window Sleep and listen windows are negotiated between the BS and the MS Duration depends on the saving class Class I, sleep window increases exponentially from minimum to maximum. Used for BE or non-real time traffic Class II, fixed length sleep window and used for UGS service Class III, One time sleep window typically used for multicast or management
Idle
Increased power saving than sleep mode MS to receive broadcast DL transmission from BS without registering itself with the network No handover action BS conserves PHY and MAC resources
Security
Private Key Management (PKM) for MAC layer security
56 bit DES-CBC, 128 bit AES-CCM encryption X.509 certification RSA authorization HMAC message integrity protection
PKM version 2 supports Extensible Authentication Protocol (EAP)
Page 239
MBS, Multicast and Broadcast Services

Multicast polling is done when there is insufficient bandwidth to poll each MS/SS individually Signaling mechanisms for MS to request and establish MBS SS access to MBS over a single or multi BS, depending on its capability & desire MBS associated QoS & encryption using a globally defined traffic encryption key A separate zone within the MAC frame with its own MAP information for MBS traffic Method for delivering MBS traffic to idle mode SS Support for macro diversity to enhance the delivery performance of MBS traffic Certain CIDs are reserved for multicast groups and for broadcast messages
Page 240
Advanced Features
H-ARQ, Packing, PHS, PKMv1 Hitless Handoff Antennas: SIMO, MISO, MIMO MIMO Matrix-C (4x4)
Four data streams are transmitted in parallel from four antennas per symbol yielding four times the baseline data rate Multiple separately encoding (horizontal) streams are transmitted over multiple antennas
Adaptive MIMO mode switching
Page 241
MIMO
Multiple Inputs to the TX antenna(s) and Multiple Outputs of RX antenna(s)
TX RF MAC PHY RF RF RF PHY MAC Rx
RF
RF PHY
BS
RF RF
TX PHY
SS
RF Rx
Page 242
MIMO
MIMO techniques improves system performance, robustness, throughput and coverage It takes advantage of multi-path and reflected signals that occur in NLOS environments BW of each subcarrier is small that enables the low cost DSP (PHY layer technology) to practically calculate the MIMO coefficient MIMO needs a better SNR than SISO Reduces interference and improves fade margin by using multiple adaptive antennas in TX/RX diversity. Approximate gain increase of 10Log(# of antenna array elements) Space time coding (transmit diversity technique by taking pair of symbols, time reverse each pair for transmission on a second antenna) Space-time diversity coding (up to NxCap but no increase in peak data rate) MIMO (spatial division multiplexing), beam forming with multiple antennas. Spatial multiplexing increases peak data rate by up to Nx with Nx antennas
Switched SISO
Receiver performs a quick AGC and level check on arriving packet and switches to a stronger signal based on the signal level
Transmitter keeps knowledge of the channel performance and switches to the better TX.
Transmitter
Signal Strength
Transmitter
Select Antenna
Knowledge of the channel
Receiver
Page 244
MIMO Matrix A (STC)

MIMO Matrix-A uses two or more antennas at transmitter (2x1) and one or more at far end receiver (1x2) Space time coding (transmit diversity) is a method which yields diversity gain without channel knowledge in the transmitter by coding across antennas (space) and across time
Applies well known Alamouti code in the downlink direction Provide reliability improvement via diversity with transmitting two redundantly encoded data streams (time reversing each pair of symbols for transmission on second antenna) during the same symbol and enables to utilize spatial (or polarization) diversity gain Overall data transfer rate remains the same as the baseline data but holds the throughput under difficult conditions Increases signal strength (3 dB higher SNR at stable conditions and about 10 dB at faded conditions compared to non-STC) by coherently combining multiple signals The Tx streams must originate at the same frequency and phase
High order modulations are more sensitive to multi-path and other impairments
One remedy is an aggressive use spatial / frequency / time diversity Space time coding (STC) is a well proven way to improve system performance Performance equivalent to maximum ratio combining with two RX antennas
Page 245
STC, (1)
There are two transmit antennas at the BS side and one reception antenna at the SS side (MISO system) Each TX antenna has its own OFDM chain Distinct pilot subcarrier location for each antenna Common location for data subcarrier but its content in a different order This technique requires Multiple Inputs Single Output channel estimation Decoding is very similar to maximum ratio combining STC achieves near optimal diversity gain in slow fading (coherence time is 10 times the channel estimate update period) environment
Antenna 1
Frequency Pilot Subcarrier Data Subcarrier DC Subcarrier
Antenna 2
Time
Page 246
STC, (2)
Cheaper to implement in BS than the SS
Applies cyclic shift into one Tx path (typical delay of 50 to 200 ns)
Two forms: with coding and without coding
TX A Data
C B A
TX B Modified Signal
Rx
C A B
C B A
Page 247
STC, (3)
First channel uses: antenna 1 transmits S0 and antenna 2 transmits S1
Second channel uses: antenna 1 transmits -S1* and antenna 2 transmits S0*
Tx-1 IFFT Filter DAC/RF Tx-2 IFFT Filter DAC/RF
Subch. Mod
IFFT Input Packing
[S0, -S1*] TX Diversity Coding [S1, S0*]
Rx Decoder Log Likelyhood Ratio Subchannel detection Diversity Combining FFT Filter RF/ADC
Page 248
STC, Space Time Coding

DL Tx diversity that can add up 3 to 10 dB link margin in faded NLOS environments
Up to N times the capacity/frequency but no increase in peak data rate Provides large coverage regardless of channel condition Adds robustness to time fluctuations and decreases frequency selectivity Applies new coefficients to the computation equations upon receiving signal offset by half wavelength
BS continuously optimizes algorithm by obtaining performance data results from the MS

Page 249
MIMO Matrix B (SDM)

Two independent data streams are transmitted over two antennas in the same time-frequency slots
An independent data stream is mapped to each transmit antenna and sends only once (unlike STC which sends the data twice)
Requires two receive antennas at the MS with proper signal equalization and decoding
Enables spatial multiplexing in down link, doubling the capacity and providing unmatched spectral efficiency when channel conditions are poor
Up to N-times the peak data rate increase with N-times antennas

Good signal quality is required and the correlation has to be low enough Huge capacity increase is expected for pico and nano cell Improves robustness to multipath fading using space diversity BS continuously optimizes algorithm by obtaining performance data from MS
Page 250
MIMO Matrix B (SDM)

SDM: Space Division Multiplexing (2x2) increases the capacity by transmitting multiple data streams in parallel on different antennas while reducing the signal quality
No increase in cell range because users near the cell edge typically have low SNR
Page 251
MRC, Maximal Ratio Combining

Both BS and MS receivers are equipped with two receive antennas performing Maximal Ratio Combining (MRC) technique for both DL & UL
Increases SNR and robustness, especially in dynamic and frequency selective channel by averaging symbol error probability in an additive white noise channel Adds spatial diversity gain at the receiver to further increase the link budget
Transmitter
Page 252
Adaptive MIMO Mode Switching

Matrix A adds robustness and coverage, while Matrix B increases capacity, but decreases the robustness
Smart adaptation algorithms are required for making decisions when to use Matrix A and when to use Matrix B Switching and algorithms decisions are made based on CINR and antenna correlation, but also on the capacity and QoS requirements Certain MIMO techniques apply pre-coded transmission and use fast feedback slot (6-bit payload) for coefficient update
1 subchannel x 6 Tiles x 8 Data-subcarriers = 48 QPSK modulated subcarriers Mapping of the MIMO coefficients to the 6-bits payload done by a codebook and a sequence of signal phase
Page 253
Adaptive MIMO Switching
Page 254
Collaborative Spatial Multiplexing

Two MS transmitters, each with one Tx antenna, may transmit at the same time and on the subchannels
In the UL, BS can receive signals simultaneously from two MSs in the same time-frequency slot Achieves multiplexing gain (capacity increase) through collaboration of MSs
Does not increase the peak UL data rate of the modem, but can double the cumulative uplink data rate in a sector
This technique is also called space division multiple access (SDMA) and requires multiple antennas at the base station
Page 255
Beam Forming (SDM)

Antenna array focuses energy in selective area or null steering in interferer
The narrowness of the beam is directly proportional to number of antennas and their gain Achieve additional robustness and capacity Higher peak rate at cells edge Robustness against inter-cell interference
Multiple antennas are required at the BS side

Emission patterns are controlled with phase and amplitude
Especially beneficial in larger cell with higher antennas Increases link budget and decreases interferences
Page 256
Two Types of Beamforming

Beamforming with phase array antenna
applies to LOS & SC applies to TX or RX Requires regular scanning mechanism like omnidirection beacon Interference rejection at RX is equally important as increased wanted signal RX signal strength depends on phase alignment of the incoming signal Beamforming with MIMO SDM
Desired
Transmitter
Gain=A+jB Delay Multipath Gain=C+jD

Algorithm
Co-Channel
Interference Mitigation with Beam Forming
Page 258
AAS, Adaptive Antenna System

AAS optional feature improves system capacity by spatially overlay coverage area by adding additional independent antennas systems Increases SNR gains toward SS while placing nulls on interfering transmitter
Increases or decreases antenna gains toward affected direction Enables transmission of DL and UL burst using directed beams to intended one or more SSs Increases expense Implemented by using multi-element phase array BS antennas
ARQ (16d), Automatic Repeat Query

Provides a rapid retransmission Implemented at below the MAC layer
Process hides the errors from TCP stack and simulates TCP error correction at lower layer
Allows selective repeat (stop, wait, go back to n) ARQ block size negotiated at connection setup (depends upon the type of service, expected delay, etc.) ARQ block cannot be fragmented Monitors Rx packets and requests retransmission if found in error(s) Protocol overhead and processing resource burden Not used in VoIP applications Ineffective in broadcast system Configurable enable/disable function Latency impact
HARQ, Hybrid ARQ (16e)

ARQ discards the previously transmitted data while HARQ combines the previous and retransmitted data to gain time diversity Uses dedicated ACK channel and PHY functions to implement a stop & wait protocol Makes use of the faster responding physical layer Complement to FEC HARQ combines ARQ with FEC such as convolutional or turbo codes. Transmitter sends a coded block. If transmission cannot be recovered by the decoder then: 1. coded data blocks are stored at the receiver. 2. retransmission is initiated. When additional coded data block is received: 1. both coded data blocks are combined and fed to decoder, 2. adds incremental redundancy and hence improves probability of recovering the data Greatly increases the data rate when SNR is very low, hence increases the coverage Typically increases the ideal BLER (block error rate) operating point by about a factor of 10 Latency impact
Page 261
Adaptive Burst Profiles

Burst profiles are transmitted in decreasing robustness
Modulation and or FEC
Dynamically throttle up or down according to the link conditions

Burst by burst, per subscriber station Trade-off capacity vs. robustness in real time
Roughly doubles the capacity for the same cell area

Burst profile for DL broadcast channel is well known and robust
Other burst profiles can be configured on the fly
SS looks for its MAC header to receive rest of the data SS capabilities are recognized at registration
Frame Diagram in Time Domain

Frame n-1
DL Subframe DL-PHY PDU
G A P
Frame n
UL Subframe CR CBR UL-PHY PDU from SS#1
Frame n+1
UL-PHY PDU from SS#n

one UL burst per UL PHY PDU
G A P
Preamble FCH DL Burst #1

Broadcast Message
DL Burst #n
Preamble UL Burst
DLFP
1 ODFM Sym with BPSK-1/2
DL-MAP, UL-MAP DCD, UCD
MAC PDUs
MAC PDU
MAC PDU MAC MSDU Header
PAD
MAC MSDU Header

CRC
CRC
Page 263
Frame Partitioning
Normal region: frequency-diverse sub-channels
Time scheduling possible but no frequency-specific scheduling i.e., used for voice services without scheduling or for flat channels Band AMC region: adjacent sub-carriers Time and frequency scheduling possible Broadcast region: frequency-diverse sub-channels in simulcast mode Borrows concept of single frequency network (SFN) from DVB/DAB etc.
Frame DL Subframe UL Subframe PUSC (Cell ID Y)
Preamble
FCH & PUSC PUSC FUSC DL-MAP (Cell ID Y) (Cell ID Z) (Cell ID Z) AMC (signaling)
AMC
Frame DL Subframe UL Subframe
Preamble
CQI/ACK
Normal Region
Band AMC Region
Broadcast Region
Normal Region
Band AMC Region NAMG

Page 264
NNG
NAMG
NBR
Guard
NNG
Frame structure (1)

Frame consists of DL & UL sub-frames
Asymmetric traffic distribution between DL and UL is always expected DL subframe consists of only one DL PHY PDU and followed by one or more UL sub-frames
UL subframe consists of:

Contention slot for initial ranging Contention slot for BW requesting UL PHY PDUs from different SS Each UL PHY PDU consists of UL preamble and UL burst
Frame n-1 Frame n Frame n+1 Frame n+1
DL PHY PDU
Contention slot A
Contention slot B For BW requests
UL PHY burst 1
UL PHY burst n
TDM signal in For initial DL ranging Adaptive

TDMA burst from different SSs (each with its own preamble)
Page 265
Frame Structure (2)

DL: One transmitter and multiple receivers (multiplexed TDM)
UL: Several transmitters and one receiver (TDMA) In UL, all transmitters have unique time and frequency offset, thus, UL system design is more difficult than the DL
The SSs are accurately synchronized such that their transmission do not overlap each other as they arrive at the BS
7 different frame durations (2.5 to 20 ms, 5 ms typical) TTG transmit/receive transition gap between DL & UL RTG receive/transmit transition gap after UL before DL Transition gap duration is a function of channel BW and OFDM symbol time This is also used for Tx/Rx mode selection and PA to settle gracefully at both ends Header suppression, packing and fragmentation techniques are applied in the frame structure for efficient use of spectrum
Page 266
DL/UL Sub-frame Sample Trace
Page 267
TDD frame structure, Sample

OFDM symbol number (time)
k S S+1 S+2 k+1 k+3 k+5 k+7 k+9 k+11 k+13 k+15 k+17 k+20 k+23 k+26 FCH
Time
k+30 k+31 k+33
UL-MAP (conti'd)
FCH
DL-MAP Preamble
RTG
Page 268
DL burst #2
UL burst #1
UL burst #2
DL burst #1
Subcarrier (frequency) Preamble
DL burst #3
UL burst #3
DL-MAP
UL-MAP
UL-MAP
UL burst #5 DL burst #5 Fast Feedback Ranging

DL subframe TTG UL subframe
S+L
OFDMA Frame Structure

DL-MAP and UL-MAP indicate the current frame structure
BS periodically broadcasts Downlink Channel Descriptor (DCD) and Uplink Channel Descriptor (UCD) messages to indicate burst profiles (modulation and FEC schemes)
Page 269
MAC Data Frame Format, Basic 802.3

Flexible frame structure allows terminals to be dynamically assigned UL & DL burst profiles according to their link conditions
Transmission order: left to right, bit serial
FCS error detection coverage FCS generation span PRE 7 SFD 1 DA 6 SA 6 Length/Type 4 Data 46 to 1500 Pad FCS 4
Field length in bytes

PRE = Preamble SFD = Start of frame delimiter DA = Destination address SA = Source address FCS = Frame check sequence
MAC
WiMAX system can be deployed as TDD, FDD or half-duplex FDD
A short gap between each DL & UL SS to remain synchronized to BS Each UL preceded by preamble (called short) that allows BS to sync with each individual SS DL starts with preamble followed by FC header then one or more DL bursts of data All symbols in the FCH and DL data bursts are transmitted with equal power to simplify the Tx & Rx design Mod-Coding remains the same within a burst but may change from burst to burst Preamble bursts are 3 dB higher than the FCH & DL data Burst generally starts with BPSK or QPSK then moves up depending on the performance
1 Frame (2.5 to 20 ms) B4 TTG P B1 P H B1 B2 B3 P B2 P B3 P B4 RTG
Downlink subframe (basestation)

Uplink subframe (subscriber)

Page 271
TDD Frame (1)

DL subframe starts with preamble that helps SS to do time and frequency synchronization and initial channel estimation
FCH provides frame configuration information such as MAP message length, modulation, coding and usable carriers. Multiple users are allocated data regions within frame and it is relayed by DLMAP & UL-MAP. MAP contains burst profile for each user such as modulation, coding. It is usually sent in BPSK- coding and repeated. Potential of increasing overhead when too many users with small packet like VoIP. Possible mitigation by use of multiple sub-MAP messages at higher rate (if there is good SNR), compress or use broadcast MAP. BPSK is used for preamble, pilot & when modulating sub-carriers in the ranging channels
TDD frame (2)

Support multiple users on a same frame
Varied size, type of data for several users Variable frame size (2-20 ms but typically 5 ms), variable packets or fragmented packets from higher layers UL sub-frame has a channel quality information that is used by scheduler (change the modulation & coding). Repeat pilots in lower modulation to improve recovery Supports Convolution, RS and optionally turbo LDPC coding
Page 273
Frame Format (3)

TDM portion Tx/Rx Transition Gap (TDD only)
Schedule
G A P
Broadcast TDM TDM TDM G control DIUC a DIUC b DIUC c A DIUC = 0 4QAM 16QAM 64QAM P 4QAM
TDMA portion
G A P
TDD For FDD

Preamble Preamble Preamble
TDMA DIUC d TDMA DIUC e TDMA DIUC f
Preamble
TDMA DIUC g
Preamble
Burst start points

DL-MAP UL-MAP PHY-Cntl MAC-Cntl
Page 274
Various MAC PDU, Example

GMH Other Packet Fixed Packet Fixed Packet Fixed ... SH Sized MSDU Sized MSDU Sized MSDU MAC PDU frame carrying several fixed length MSDUs packed together Other FSH MSDU Fragment SH MAC PDU frame carrying a single fragment MSDU CRC
GMH
CRC
GMH
Variable Sized Variable Sized MSDU Other FSH PSH . . . CRC MSDU or Fragments SH MAC PDU frame carrying several variable length MSDUs packed together payload Other ARQ Feedback FSH SH MAC PDU frame carrying ARQ payload Variable Sized Other PSH ARQ Feedback PSH MSDU or Fragment SH MAC PDU frame carrying ARQ and MSDU payload ... CRC
GMH
GMH
CRC
MAC Management Message GMH CRC MAC management Frame CRC: Cyclic redundancy check GMH: Generic MAC Header FSH: Fragmentation Subheader PSH: Packing Subheader PDU: Packet Data Unit SH: Subheader
DL Subframe (1)
The first DL burst contains
DL map (DL MAP)
DL MAP always refers to current frame
UL map (UL MAP)

UL MAP may be broadcasted one frame ahead
DL channel descriptor (DCD) UL channel descriptor (UCD)
DL bursts are broadcasted in order of decreasing robustness BPSK> QPSK> 16QAM> 64QAM
A SS listens to all bursts it is capable of decoding
A SS does not know which DL burst (s) contain(s) information sent to it

TDD Downlink Sub-frame (2)

DL subframe starts with
Preamble FCH, Frame control header DIUC: downlink interval usage code TTG/RTG, this gap is an integer number of PS (physical slot = 4 modulation symbols) durations and starts on a PS boundary A portion of the DL subframe can be designated as zone for STC and AAS applications
DL PHY PDU Contention slot A Contention slot B UL PHY burst 1 UL PHY burst n
Preamble
FCH
DL burst 1
DL burst n
DL-MAP, UL-MAP DCD, UCD

MAC PDUs
Page 277
DL Map Message (3)
DL-MAP message defines usage of DL and contains carrier-specific data

DL allocation can be of broadcast, multicast and unicast
DL-MAP is the first message in each frame Decoding is very time-critical

Typically done in hardware
Entries denote instant when the burst profile change
Page 278
Typical Uplink Sub-frame (1)

Initial maintenance opportunities
Ranging (a procedure for MS to gain access to the BS) To determine network delay and to request power or profile changes Collisions may occur in this interval
Request opportunities
SSs request bandwith in response to polling from BS Collisions may occur in this interval as well
Data grants period

SSs transmit data bursts in the intervals granted by the BS Transition gaps between data intervals for synchronization purposes Any of these burst classes may be present in any given frame in any order and any quantity (limited by the number of available PSs) within the frame at the discretion of the BS UL scheduler as indicated by UL-MAP
UL Subframe Structure (2)

The SSs transmit in their assigned allocation using the burst profile specified by the UIUC (UL interval usage code) in the UL-MAP entry granting them bandwidth
UL subframe starts with Contention slot for initial ranging requests
Contention slot for bandwidth request messages

Tx/Rx transition gap (TTG)
Preamble
Initial maintenance opprtunities (UIUC = 2) Request contention opportunities (UIUC = 1)
SS transition gap
Preamble
SS 1 scheduled data (UIUC = i)
Rx/Tx transition gap (RTG)

SS N scheduled data (UIUC = j)
Preamble
Access burst
collision Gap
Access burst
Bandwidth collision request
Bandwidth request
Page 280
UL Transmission
UL is considered to be invited transmission and is more complicated than the DL
Transmissions in initial ranging slots
Ranging Requests (RNG-REQ)
Contention resolved using truncated binary exponential back-off algorithm
Transmissions in contention slots

Bandwidth requests Contention resolved using truncated binary exponential back-off algorithm
Each of these contention slots is further divided into minislots Bursts defined by UIUCs (UL interval usage code) by BS and the SS adapts and adjusts accordingly Transmissions allocated by the UL-MAP message All transmissions have synchronization preamble Ideally, all data from a single SS is concatenated into a single PHY burst
UL Physical Layer
The UL transmission convergence sub-layer is identical to the DL one. The UL PMD (physical media device) layer coding and modulation are as follows:
Three classes of bursts transmitted during the UL sub-frame:
Burst transmitted in contention opportunities reserved for initial maintenance Burst transmitted in contention opportunities provided by multicast and broadcast polls Bursts transmitted in intervals specifically allocated to individual SS
All UL transmissions are made according to the UL burst profiles, specified by the BS
Each UL burst begins with an uplink preamble
Page 282
UL Channel Descriptor
Defines uplink burst profiles

Sends regularly
All UL burst profiles are acquired

Burst profiles can be changed on the fly Establishes association between UIUC (UL interval usage code) and actual PHY parameters
Page 283
UL-MAP Message
UL-MAP message defines usage of the UL

Contains the grants Grants addressed to the SS Time given in mini-slots (A unit of UL BW allocation
equivalent to n physical slots, where n = 2^m, m is an integer ranging from 0 through 7)
unit of UL bandwidth allocation 2m physical slots
in 10-66GHz PHY physical slot is 4 modulation symbols long
Time expressed as arrival time at BS

UL Contention Resolution
Based on a truncated binary exponential backoff
The initial/maximal backoff window is controlled by the BS The SS shall randomly select a number within its backoff window This random value indicates the number of contention transmission opportunities that the SS shall defer before transmitting For bandwidth requests, if the SS receives a Unicast Request IE or Data Grant Burst Type IE at any time while deferring for this CID, it shall stop the contention resolution process
One Request IE Transmission Opportunity #1 Transmission Opportunity #2 Transmission Opportunity #3
Preamble (2 minislots)
BW Req Message (3 minislots)
SSTG (3 minislots)
Page 285
Downlink Preamble
DL subframe starts with two OFDM symbols containing preamble (called long)
Symbol 1 contains 50 subcarriers (every fourth subcarrier with no data or pilot subcarriers resulting in wider adjacent channel spacing ), QPSK Symbol 2 contains 100 subcarriers (every even subcarriers with no data or pilot subcarriers resulting in wider adjacent channel spacing ), QPSK Transmitted at 3 dB higher level than all others to make Rx to easily recover information Preamble followed by FCH then one or more data Symbols
All symbols in FCH and DL data burst are transmitted with equal power
Same modulation is kept within the burst but it may change from burst to burst Data initially starts out with low level modulation then gradually increases depending on RSL and CINR Generally inserts a few short mid-preamble in extremely long DL burst
OFDM Frame Structure Diagram

Variable number of subcarriers for OFDMA
Can contain DL MAP (if FCH is too small, DL BURST is used)

Long preamble 2 symbols
Preamble 1 symbol
Preamble 1 symbol
FCH DL DL 1 symbol Burst #1 Burst #2

BPSK
DL G Preamble Burst #n A 1 symbol P

100 CXR BPSK
Ranging
CDMA
BW
UL UL Burst #1 Burst #2
UL Burst #n
50 CXR BPSK 100 CXR BPSK
200 carriers, BPSK/ QPSK/16QAM/64QAM
200 carriers, BPSK/ QPSK/16QAM/64QAM
DL Subframe variable number of OFDM symbols
UL Subframe variable number of OFDM symbols
Frame #1
Frame #2
Frame #n
Page 287
Downlink/Uplink Preamble
UL subframe starts with short single OFDM Symbol that synchronizes the BS to the SS
Preamble (called short) consists of 100 even number subcarriers
Uses QPSK-1/2 modulation

Same power as data sub-carriers
Symbol contains no data or pilot subcarriers
Following DL preamble is a FCH (single OFDM Symbol of BPSK, 88 bits of overhead data that describes critical system decoding information such as BS ID and DL burst profile). DL burst contains one or more Symbols. Each symbol in the DL burst contains 12 to 108 bytes of payload data, depending on the modulation & coding types
Preamble Plot
Preamble
Data
8+8 = 16 us, Preamble

10x0.8 = 8 us t1 t2 t3 t4 t5 t6 t7 t8 t9 t10 GI2 2x0.8 + 2x3.2 = 8 us T1 T2
0.8+3.2 = 4 us
0.8+3.2 = 4 us 0.8+3.2 = 4 us
GI Signal GI Data 1 GI Data 2
Coarse Freq Signal Detect, AGC, Offset Estimation Diversity Selection Timing Sync
Channel and Fine Frequency Offset Estimation
Rate Length
OFDM Service+Data
Page 289
Training Symbol Structure

Flexible usage in OFDMA and MIMO
OFDM Packet (time domain)
Preamble User Data
1 OFDM symbol Preamble-based
{
3 OFDM symbol Pilot-Based 2D TimeFrequency Interpolation 1D Frequency Interpolation 1D Time Interpolation Time Time Training Symbol
Frequency
Frequency
Data Symbol
Page 290
DL Physical Layer, (1)

Available bandwidth in DL direction: physical slots
Available bandwidth in UL direction: mini-slots (mini slot length = 2^m physical slots where m is an integer ranging from 0 through 7) Number of physical slots with each frame is a function of symbol rate (20 Mbps: 5000 PHY. Slots within 1 ms frame)
DL frames can be TDD (the subframe contains preamble for synchronization and equalization, frame control section to see where bursts begin, and data) and FDD (preamble, frame control section and TDM portion organized into bursts transmitted in decreasing order of burst profile robustness)
DL & UL Physical Layers, (2)
Physical layer allows for flexible spectrum usage and support, both TDD and FDD
Burst transmission format is framed to support adaptive burst profiling (modulation and coding schemes can be adjusted individually to each SS)
The UL physical layer is based on a combination of:

TDMA (time division multiple access) DAMA (demand multiple access)
Page 292

UL channel is divided into a number of time slots
Its various number is controlled by the MAC layer in the BS
DL channel is a TDM (information for each subscriber is multiplexed onto a single stream of data) The downlink physical layer includes a transmission convergence sub-layer which helps the receiver to identify the beginning of a MAC frame. The PHY layer performs randomization, FEC encoding and modulation (QPSK, 16-QAM or 64-QAM)
Page 293

The UL PHY layer is based upon TDMA burst transmission
Each burst is designed to carry variable length MAC frames
PMD layer performs randomization, FEC encoding and modulation

Frame duration: 2.5 to 20 ms Each frame contains a DL sub-frame and an UL sub-frame In the TDD case, UL and DL transmissions share the same frequency but are separated in time In FDD case, both transmissions occur at the same time but the channels are on separated frequencies
Protocol Architecture
IEEE 802.16 Protocol Architecture has 4 layers: Convergence, MAC, Transmission and physical, which can be mapped into two lowest OSI layers: physical and data link
Bridged LAN Digital telephony ATM Digital audio/ IP Back haul video multicast Virtual point to point Frame ralay
TCP/IP model
TCP/IP protocols NFS HTTP FTP SMTP Name Server XDR
OSI Reference Model Layers 7 6 RPC 5

Session Transport Network Data Link Physical Application Presentation
MAC Convergence sublayer
ATM Ethernet, 802.1Q Internet Protocol
Application
Transport Internet
Transmission Control Protocol TCP
User Data Protocol UDP
4 3
OSI data layer
MAC
Packing, Fragmentation, ARQ, QoS
Internet Potocol IP Ethernt IEEE 802.3 twisted pair optical fiber Token Ring coaxial cable DQDB
MAC Transmission, Privacy sublayer OSI physical layer
Authentication, Key Exchange, Privacy (encyption)

OFDM, ranging, power control, DFS, Tx, Rx
2 1
Network
Physical Layer
Medium
Protocol Structure
CS: All functions that are specific to a higher layer protocol
Receives and adapts higher layer PDUs to MAC CPS Classifies SDUs based on MAC address, VLANs, priorities Assigns service flow ID (SFID) and connection identifier Maps data to a CID
MAC MAC Convergence sublayer (CS)
CPS: Provides the core MAC functionality

Fragmentation and reassembly of large MAC SDUs Packing and unpacking of several small MAC SDUs QoS control and scheduling
MAC Common Part Sublayer (CPS)
MAC Transmission, Privacy sublayer
Bandwidth request and allocation

Automatic repeat Query (ARQ)
PHY
Physical Layer (PHY)
Page 296
Security & PHY Sub-layer

Provides authentication, secured key exchange, encryption support ARQ scheme
Supports two protocols: Encapsulation protocol for data encryption
Defines cryptographic suites such as pairings of data encryption and authentication algorithms
The rules for applying those algorithms to a MAC payload
MAC
MAC Convergence sublayer (CS)
MAC Common Part Sublayer (CPS)
Privacy key management protocol

Describes how the BS distributes to SS
PHY Sub-Layer
Single carrier, 11-66 GHz
MC, NLOS, below 11GHz, ARQ, AAS & MIMO S-OFDMA, NLOS, H-ARQ, Fast feedback, Handover
MAC Transmission, Privacy sublayer
PHY
Physical Layer (PHY)
Page 297
Unicast Polling
1. BS allocates sufficient space for the SS in the uplink subframe 2. SS uses the allocated space to send a BW request 3. BS allocates the requested space for the SS (if available) 4. SS uses allocated space to send data
BS SS
Poll(UL-MAP)
Alloc(UL-MAP)
Request
Data
Page 298
ATM Convergence Sub-layer
Supports for:
VP (Virtual Path) switched connections VC (Virtual Channel) switched connections
Support for end-to-end signaling of dynamically created connections
SVCs
Soft PVCs ATM header suppression Full QoS support
Packet Convergence Sub-layers
Initial support for Ethernet, IPv4 and IPv6

Payload header suppression
Generic plus IP specific
Full QoS support Possible future support for:

PPP
MPLS etc.
Page 300
MAC Addressing
SS has 48-bit IEEE MAC address

BS has 48-bit Base Station ID
Not a MAC address 24-bit operator indicator
16-bit Connection ID (CID)

Used in MAC PDUs (packet data units)
Page 301
MAC PDU Transmission

MAC communicates using MAC protocol data units (MPDUs) that are carried by the PHY MAC PDUs (packet data units) are transmitted in PHY bursts A single PHY burst can contain multiple Concatenated MAC PDUs The PHY burst can contain multiple FEC blocks MAC PDUs may span FEC block boundaries The TC (transmission conversion) layer between the MAC and the PHY allows for capturing the start of next MAC PDU in case of erroneous FEC blocks
The TC PDU format allows resynchronization to the next MAC PDU if the previous block had irrecoverable errors Without the TC layer, a receiving SS or BS would potentially lose the entire remainder of a burst when an irrecoverable bit error occurred Performs conversion of variable length MAC PDUs into fixed length FEC blocks (plus possibly a shortened block at the end) of each burst
MAC PDU Transmission
MAC Message Fragmentation MAC PDUs PHY Burst PDU 1 PDU 1 PDU 2
SDU 1 Packing PDU 3 PDU 4
SDU 2
PDU 5
Concatenation P FEC 1 MAC PDUs FEC 2 Preamble FEC 3 Shortened FEC block
Page 303
MAC PDU Format

There are types of MAC header (generic or BW request)
Both generic and BW request MAC headers are fixed length and 6 bytes long
One or more MAC sub-headers may be part of the payload The presence of sub-headers is indicated by a type field in the Generic MAC header Size varies from 6 byte to 2047 bytes Flexibility creates transmission inefficiency
msb
6 bytes
0 to 2041 bytes
4 bytes
Generic MAC Header

Payload (optional)
CRC (optional)
Page 304
GMH, Generic MAC Header

The GMH is used for transmit data or MAC messages and may optionally have one or more appended sub-headers
Fragmented Sub-header (2 bytes, optionally 1 byte)
Packing (3 bytes, optionally 2 bytes)

Grant Management (2 bytes) Mesh Sub-header (2 bytes)
Fast-Feedback-Allocation (1 byte)
Extended Sub-header (variable length) The subheader can occur only once per MAC PDU except for the Packing subheader, which may be inserted before each MAC SDU packed into the payload
Page 305
BWH, Bandwidth Header
The BWH is used by the SS to request more bandwidth on UL

ARQ Fast Feedback and Grant Management subheaders are used to communicate ARQ and bandwidth allocation states between the BS and SS Fragmentation and Packing sub-headers are used to utilize the bandwidth allocation efficiently
Page 306
PHSF, Payload Header Suppression Format
If PLHS is enabled at MAC connection, each MAC SDU is prefixed with a PHSI (payload header suppression index), which references the PHSF (payload header suppression field). The classifier (located at the sending entity) uniquely maps the packets to its associated PHS Rule. The receiving entity uses the CID and the PHSI to restore the PHSF.
Payload Header
Useful portion
PHSF
Payload
PHSI
Useful portion
Payload
Page 307
Header Suppression for VoIP over WiMAX
The protocols used in addition to WiMAX are RTP, UDP and IPv6
Application Layer Voice Payload RTP UDP IPv6 MAC PHY

Voice Payload Voice Payload Voice Payload Voice Payload Voice Payload
Page 308
Header Suppression for VoIP over WiMAX, 2
Header sizes of each of these layers:

between 12 to 72 bytes for RTP 8 bytes for UDP
40 bytes for IPv6

the total length of RTP/UDP/IPv6 header is between 60 and 120 bytes
PHS suppresses repetitive (redundant) parts due to the higher layers in the payload header of the MAC SDU The receiving entity restores the suppressed parts Its is the responsibility of the higher-layer service entity to generate a PHS Rule
Page 309
A Payload Header Suppression Index (PHSI), an 8-bit field which references the Payload Header Suppression Field (PHSF) that has been used for header suppression
The PHS rule has also a Payload Header suppression Mask (PHSM) option to allow the choice of bytes of PHSF that can not be suppressed
B 0 X D 0 X B
C 1 C
D 0 X
E 1 E
=verify =assign Packet Transmission PHSF PHSM
Payload
PHSI 1 byte MAC header
Sender
PHSM PHSF
1 A B
Air Interface PHSM Receiver PHSF
Payload
1 A A
1 C C
0 X D
1 E E
A-E=curent in A-E=cached X=dont care PHSS=5
Packet Header Reconstruction (using PHSI and CID)
Payload
Page 310
Number of suppressed bytes per header:

IPv6: 37 bytes UDP: 4 bytes RTP: 4 bytes.
The RTP/UDP/IPv6 Header drops from 60 bytes to 15 bytes (45 Header bytes or less are suppressed)
Page 311
Fragmentation
Partitioning a MAC SDU into fragments then transporting in multiple MAC PDUs
Longer packet increases probability of losing a packet and hence initiate retransmission Allows better packing of MAC SDUs into the available OFDM freq-time resources by using all data subcarriers in each OFDM symbol Each connection can be in only a single fragmentation state at any time Contents of the fragmentation sub-header: 2-bit Fragmentation Control (FC) Unfragmented, Last fragment, First fragment, Continuing fragment 3-bit Fragmentation Sequence Number (FSN) Required to detect missing continuing fragments Continuous counter across SDUs
Fragmentation is an optional feature that improves the link efficiency
Page 312
Packing, (1)
A process of combining multiple MAC SDUs (or fragments thereof) into a single MAC SDU
Allows better packing of MAC SDUs into the available OFDM frequency-time resources by using all data subcarriers in each OFDM symbol
Can, in certain situations, save up to 10% of system bandwidth
On connections with variable length MAC SDUs

Packed PDU contains a sub-header for each packed SDU (or fragment thereof)
On connections with fixed length MAC SDUs

no packing sub-header needed
Packing and fragmentation can be combined

Packing Fixed-Length SDUs, (2)
k MAC SDUs
MAC Header LEN = n*k+j
fixed length fixed length fixed length MAC SDU MAC SDU MAC SDU length = n length = n length = n
....
fixed length MAC SDU length = n
Page 314
Packing Variable Length SDU, (3)
2 Byte packing sub-header before each SDU

Length of the SDU: 11 bits Fragmentation control (FC): 2 bits
Fragmentation sequence number (FS): 3 bits
k MAC SDUs
MAC Header LEN = j Type = 00001Xb
PSH Length = a+2
PSH Length = b+2
variable length MAC SDU length = a
variable length MAC SDU length = b
....
PSH Length = c+2
variable length MAC SDU length = c
Page 315
MAC Header LEN=y5 Type=00001xb PSH FC=01, FSN=x+s+1 Length=b+2

Last fragment of MAC SDU length = a
MAC Header LEN = y2 Type = 00010xb FSH FC = 11, FSN = x+y Length = e+1
MAC Header LEN = y1 Type = 00001Xb PSH FC = 01, FSN =x' Length = a+2
Continuing fragment of MAC SDU legth =e
PSH FC=D0, FSN=x+s+2 Length=1+2
Packing with Fragmentation
PSH FC=D0, FSN=x+s+3 Length=1+2
PSH FC=00, FSN=x+s+2 length=k+2
Last fragment of MAC SDU length=h
MAC Header LEN = y3 Type = 00010xb FSH FC=11, FSH = x+y1 Length = f+1
PSH FC = 00, FSN = n+1
unfragmente d MAC SDU length = 0
r MAC SDUs
unfragmente d MAC SDU U length=f unfragmente d MAC SDU length=f Unfragmente d MAC SDU length=k
r MAC SDUs
PSH FC = 00, FSN = n+2 Length = c+1
s-I+1 MAC SPUs
MAC Header LEN=y45 Type = 00010Xb FSH FC=11, FSN=x+5 Length = q+1
Continuing fragment of . . . MAC SDU length=f continuing frragment of MAC SDU length=g
PSH FC = 10, FSH = x+y+1 Length = d+2
unfragmente d MAC SDU length = c first fragment of MAC SDU length = d'
Page 316
OFDMA, Typical TDD Time Frame

Pilot, null and DC subcarriers are not shown
OFDMA Symbol Number
k+0 k+1 k+2 k+3 k+4 k+7 k+9 k+11 k+13 S+0 S+1 S+2 ... k+17 k+20 k+23 ... ...
Time
k+31 k+33
FCH DL burst #2
UL-MAP
Ranging Subchannels UL burst #1
FCH
Preamble
DL burst #4
UL burst #3
Preamble
RTG
Page 317
DL-MAP
DL burst #1 DL burst #5 UL burst #4
S+N DL TTG
UL burst #5
UL
UL-MAP
DL-MAP
Chain Transmission
Randomization and FEC coding in UL are identical to the corresponding in the DL The type of modulation and the power adjustment rules are set by the BS QPSK, 16QAM and 64QAM are mandatory and the 256QAM is optional The mapping of bits to symbols are identical to those in the DL Systems shall use Nyquist square-root raised cosine pulse shaping (role off factor 0.25) A frame duration of 5 ms is typically used as the compromise between transport efficiency and latency Must be able to compensate at most 20 dB/s for 40 dB range
Actual power control algorithm is left to vendors 0.25 dB resolution
Page 318
Down Link Transmission

Two kinds of bursts: TDM and TDMA
All bursts are identified by a DIUC
Downlink Interval Usage Code
TDMA bursts have resync preamble

allows for more flexible scheduling
Each terminal listens to all bursts at its operational IUC, or at a more robust one, except when told to transmit Each burst may contain data for several terminals SS must recognize the PDUs with known CIDs DL-Map message signals DL usage
Page 319
Downlink Channel Descriptor
Used for advertising DL burst profiles

Burst profile of DL broadcast channel is well known
All others are acquired

Burst profiles can be changed on the fly without interrupting the service
Not intended as super-adaptive modulation
Establishes association between DIUC and actual PHY parameters
Page 320
Burst Profiles
Each burst profile has mandatory exit threshold and minimum entry threshold
SS allowed to request a less robust DIUC (DL interval usage code) once above the minimum entry level SS must request fall back to more robust DIUC once at mandatory exit threshold Requests to change DIUC done with DBPC-REQ (DL burst profile change Req.) or RNG-REQ (range Req.) messages
Burst Profile Z Overlap Burst Profile Y Overlap Burst Profile

Page 321
C / (N+I)
0
UL-AAS Beam Response Message

Message contains a total of 48 bits
Management message type, 8 bits Frame number, 8bits
Feedback request number, 3 bits

measurement report Types, 2 bits Resolution parameter, 3 bits
Beam bits mask, 4 bits

reserved, 4 bits RSSI mean value, 8 bits
Quantized in 2 dB increment in range from -48 to -110 dBm
CINR mean value, 8 bits

Quantized in 1 dB increment in range from 10 to 53 dB
Admission Control, Scheduling and Link Adaptation

Admission Control
Ensure that new flows do not degrade the quality of established flows
Scheduling
BS schedules usage of the air link among the subscribers per specific QoS Packet schedulers at the BS and subscribers gives transmission opportunities to multiple connection queues
Link Adaptation
BS determines the contents of the DL and UL portions of each frame
BS determines the appropriate burst profile (code rate, modulation level and so on) for each subscriber BS determines the BW requirements of the individual subscriber based on the service classes of the connections and on the status of the traffic queues at the BS and SS
Page 323
The QoS Object Model
PDU
SFID [Sevice Class] [CID] Payload N 1
Service Flow
SFID Direction [CID] [Provisioned QoS ParamSet] [AdmittedQoSParamSet] [ActiveQoSParamSet] N 0, 1 1 0, 1
Connection
Connection ID QoS Parameter Set
Service Class
Sevice Class Name QoS Parameter Set
Page 324
QoS Mechanisms
Classification
Mapping from MAC SDU fields (e.g., destination IP address or TOS field) to CID and SFID
Scheduling
Downlink scheduling module
Simple, all queues in BS
Uplink scheduling module

Queues are distributed among SSs Queue states and QoS requirements are obtained through BW requests
Algorithms not defined in standard

QoS Control
.
Applications
Control Plane Conn_req Conn_rsp Control Plane AC
Connection Classifiers
BE CIDs nrtPS CIDs rtPS CIDs ertPS CIDs UGS CISs
UL Scheduling BW request UL Map

Ctrl/mng channels
UGS Non UGS BS
Priority Scheduler SS
UL Data packets (data channels)
Page 326
QOS Mechanism
Subscriber Station
MAC CS
New Connection Connection Request
Base Station
Connection Request Connection Response Implicit Request Piggyback Request Unicast Polling Contention Based Polling
MAC CPS Connection Request Generator
MAC CPS
MAC CS
Packet Re-Construction
Admission Control
CID#6 CID#7
TDM/ Voice VoIP MPEG TFTP, FTP HTTP E-mail
Packet Classifier
CID#6 (UGS) CID#7 (ert-PS) CID#8 (rt-PS) CID#9 (nrt-PS)
TDM/ Voice VoIP MPEG TFTP, FTP HTTP
UL BW Request Generator
Slot Allocation
CID#8 CID#9
CID#10 (BE)
UL BW Grant Processor
Data Traffic
UL BW Grant Generator
CID#10
E-mail
UL-MAP
TDM/ Voice VoIP MPEG TFTP, FTP HTTP E-mail
Packet Construction
CID#1 CID#2 CID#3 CID#4
DL Traffic Processor
Data Traffic
DL Traffic Processor
CID#1 CID#2
Packet Classifier
TDM/ Voice VoIP MPEG TFTP, FTP HTTP
DL-MAP Generator
DL-MAP
CID#3 CID#4
CID#5
CID#5
E-mail
Page 327
5-Types of Scheduling Services

Unsolicited Grant Service (UGS)
for constant bit-rate (CBR) or CBR-like service flows (SFs) such as T1/E1
Extended real-time Polling Service (ertPS)

for real time variable bit rate in an unsolicited manner and has less request/grant overhead than the rtPS, VoIP services with silent suppression
Real-time Polling Service (rtPS)

for rt-VBR-like SFs on periodic basis such as MPEG video
Non-real-time Polling Service (nrtPS)

for nrt SFs with better than best effort services such as bandwidth-intensive file transfer (FTP)
Best Effort (BE)

for best-effort traffic with no minimum service level required
Page 328
UGS, Unsolicited Grant Service

Supports services that generate fixed size data packets on periodic basis
T1/E1 services or voice over IP without silence suppression
No need for explicit BW requests

Low overhead
Offers fixed size grants on a real time periodic basis, which eliminates overhead and latency of SS requests
No unicast request opportunity provided May include a grant Management (GM) sub-header containing
Slip indicator: indicates that there is a backlog in the buffer due to clock skew or loss of maps
Poll-me bit: indicates that the terminal needs to be polled (allows for not polling terminals with UGS-only services)
Page 329
ertPS, extended-real-time Polling System (for 16e)
An extended real-time polling service (ertPS) combines UGS & rtPS

Supports VoIP with silence suppression
Periodic unsolicited grants similar to UGS for data transmission or for requesting additional BW
Unlike UGS, allocations are not fixed and may change over time (on/off UGS) Default size is based on maximum sustained traffic rate MS may request a change in allocation size, using grant management sub-header or other means
Page 330
rtPS, real-time Polling System

Supports real time flows with variable size data packets on periodic basis such as MPEG video Provides periodic request opportunities
SS specifies the frame size in the BW request in response Unicast request opportunities which meets the flows real time needs and allows SS to specify the size of desired grant
Prohibited from using any contention requests More overhead, but more flexible and provides optimum data transport efficiency than UGS Terminal polled frequently enough to meet the delay requirements of the SFs Bandwidth requested with BW request messages (a special MAC PDU header) May use Grant Management sub-header
new request can be piggybacked with each transmitted PDU (protocol data unit)
nrtPS, non-real-time Polling System

Works like rt-polling except that polls are issued less frequently Combines periodic and contention request opportunities Base station issues unicast polls on the order of a second or less SS may also use contention request opportunities Can be used for delay tolerant traffic
No delay or jitter guarantees
Intended for non-real-time service flows with better than best effort service
e.g., bandwidth extensive file transfer
May use Grant Management sub-header

New request can be piggybacked with each transmitted PDU
BE, Best Effort

For best-effort traffic in the UL
Generic Data e.g., HTTP, SMTP, etc.
No QoS guaranteed
Leftover or unused allocation may be used by SSs SS/MS allowed to use contention request opportunities
BS may allocate unicast opportunities

Depending on policy No guarantees
May use Grant Management sub-header

New request can be piggybacked with each transmitted PDU
QoS
Convergence Sublayer
Downlink Data Translator
Queques without Latency
Downlink Determine UL/DL subframe 1st Phase Proportionating
Uplink
Assign slots for queques Second Phase Proportionating Two-Phase Proportionating
UGS ertPS
Translator
rtPS nrtPS BE
Assign slots to SSs Write in DL-MAP
Assign slots to SSs Write in UL-MAP Uplink Frame
Bandwidth Requests
Two-Phase Proportionating
PHY Layer
Downlink Frame
Page 334
Service Flows and QoS, example

Priority + EDF + WFQ + RR - combined model
UGS CISs ertPS CIDs rtPS CIDs Fixed Bandwidth Fixed Bandwidth with Silent Detect
Priority Scheduler
Earliest Deadline First (EDF)
nrtPS CIDs
Weighted Fair Queuing (WFQ) or WRR
UL Map
BE CIDs
Round Robin
Page 335
QoS Summary
QoS Category
UGS
Unsolicited Grant Service
Applications
VoIP, T1/E1
Fixed data rate
QoS Specifications
Maximum sustained rate
Maximum latency tolerance Jitter tolerance
ertPS Extended Real Time Packet Service
Voice with activity detection (VoIP) Variable data rate
Maximum sustained rate Minimum reserved rate Maximum latency tolerance Jitter tolerance, Traffic priority
rtPS Real Time Packet Service
Screaming Audio and MPEG Video
Minimum reserved rate Maximum sustained rate Maximum latency tolerance
Committed burst size, Traffic priority

nrtPS Non-Real Time Packet Service BE Best Effort Service
File Transfer Protocol (FTP)
Minimum reserved rate Maximum sustained rate
Traffic priority
General data transfer, Web Browsing, etc... Maximum sustained rate Traffic priority
Page 336
Scheduling Types
Scheduling Type UGS Piggy Back Request Not Allowed Not Allowed PM bit is used to request unicast poll for bandwidth needs of nonUGS connections Scheduling only allows unicast polling Scheduling only allows unicast polling Scheduling may restrict a service flow to unicast polling via the transmission/request policy; otherwise all forms of polling are allowed All forms of polling allowed
Page 337
BW stealing
Polling
ertPS rtPS nrtPS
Allowed Allowed Allowed
Allowed for GPSS (Grant per SS) Allowed for GPSS Allowed for GPSS
BE
Allowed
Allowed for GPSS
Request / Grant Scheme

Self Correcting
No acknowledgement All errors are handled in the same way, i.e., periodical aggregate requests
Bandwidth Requests are always per Connection Grants are either per connection (GPC) or per Subscriber Station (GPSS)
Grants (given as durations) are carried in the UL-MAP messages SS needs to convert the time to amount of data using information about the UIUC
Page 338
BW Requests
Comes from the Connection
Several kind of requests:
Implicit requests (UGS)
No actual messages, negotiated at connection setup

BW request messages
Uses the special BW request header Requests up to 32 kb with a single message Incremental or aggregate, as indicated by MAC header
Piggybacked request (for non-UGS services only)

Presented in GM sub-header and always incremental
Up to 32 kb per request for CID
Poll-Me bit (for UGS services only)

Used by the SS to request a bandwidth poll for non-UGS services
BW Allocation and Burst Placement

.
Flow 1
Flow Queque in DL
Flow 2 Flow 3 Flow 4 Flow 5
Flow 1
Flow 2 Flow 3
Flow 4
Flow 5
Flow 1
Flow 5 UL Subframe
Flow 4
Flow 2
Flow 3
Page 340
Bandwidth Request and Allocation, (1)
SSs may request BW in 3 ways:

Uses contention request opportunities interval upon being polled by the BS (multicast or broadcast poll)
Contention is resolved by using back off resolution
Sends a standalone MAC message called BW request in an already granted slot

Due to the predictable signaling delay of the polling scheme, contention-free mode is suitable for real time applications
Piggybacks a BW request message on a data packet
Page 341
BS grants/allocates bandwidth in one of two modes:

Grant Per Subscriber Station (GPSS)
BS scheduler treats all the connections from a single SS as one unit and grants BW to the SS. An additional scheduler is employed at the SS which determines the service order for its connections in the granted slot More scalable and efficient than the GPC
Grant Per Connection (GPC)

BS scheduler treats each connection separately and BW is expilicitly granted to each connection SS transmits according to the order specified by the BS
Decision based on requested BW and QoS requirements vs. available resources Grants are realized through the UL-MAP

DL-MAP and UL-MAP indicate the current frame structure
BS periodically broadcasts DL Channel Descriptor (DCD) and UL Channel Descriptor (UCD) messages to indicate burst profiles (modulation and FEC schemes)
Page 343
GPSS vs. GPC

Bandwidth Grant per Subscriber Station (GPSS)
Base station grants bandwidth to the subscriber station Subscriber station may re-distribute bandwidth among its connections, maintaining QoS and service level agreements Suitable for many connections per terminal; off-loading base stations work Allows more sophisticated real time reaction to QoS needs Low overhead but requires intelligent subscriber station Mandatory for P802.16 10-66 GHz PHY
Bandwidth Grant per Connection (GPC)

Base station grants bandwidth to a connection Mostly suitable for a few users per subscriber station Higher overhead, but allows simpler subscriber station
Page 344
Maintaining QoS in GPSS
Semi-distributed approach
BS sees the requests for each connection; based on this, grants bandwidth (BW) to the SSs (maintaining QoS and fairness) SS scheduler maintains QoS among its connections and is responsible to share the BW among the connections (maintaining QoS and fairness) Algorithm in BS and SS can be very different; SS may use BW in a way unforeseen by the BS
Page 345
SS Initialization Steps
Scans for DL channel and establish synchronization with the BS
Obtains transmit parameters (from UCD message) Performs ranging
Negotiates basic capabilities

Authorizes SS and performs key exchange Performs registration
Establishes IP connectivity
Establishes time of day Transfers operational parameters
Set up connections
Page 346
Ranging, (1)
Procedure for MS to gain access to the BS Four types of ranging can be defined
Initial Ranging for Network entry Periodic Ranging for synchronization Bandwidth requests HO ranging
Single Ranging channel (multiple sub-channels) using 1 to 8 subcarriers defined by the system specified in the UCD Ranging process accomplished through PN codes assigned spreading for specific Ranging types
Also known as CDMA-like (a maximum of 256 sets of 144-bit wide pseudo-noise code) ranging for OFDMA
Page 347
Ranging, (2)
For UL transmissions, times are measured at BS
At start up, SS sends a RNG-REQ message in the contention slot reserved for this purpose
SS looks for initial ranging opportunities (UL-MAP) information present in every frame
BS measures arrival time and signal power; calculates required advance and power adjustment BS send adjustment in RNG-RSP SS adjusts advance and power, sends new RNG-REQ
Loop between BS & SS is continued until power and timing is ok

Channel Acquisition
SS scans for suitable BS DL signal

SS Sync to this signal and searches the first DL burst of the DL PHY PDU
Reads the DL channel descriptor (DCD) Reads the UL channel descriptor (UCD)
Learn the modulation and coding schemes used on the carrier
Page 349
Negotiation of Capabilities
BS sends
Power adjust information Timing adjust information
CID for the basic management connection

CID for the primary management
SS reports its PHY capabilities on the primary management connection

Modulation Coding scheme Half-duplex or full-duplex operation (FDD)
BS may deny the use of any capability reported by the subscriber station
SS Authentication
SS must pass authentication

SS contains an X.509 digital certificate and the certificate of the manufacturer SS sends these certificates to BS BS examines certificates and authenticates (or deny) the SS If authentication is successful, the BS sends the authorization key
The AK is used both by SS and BS for securing further information flow (subsequent key derivation)
Registration
Registration is a form of capability negotiation

SS sends a list of capabilities and parts of the configuration file to the BS in the REG-REQ message BS replies with the REG-RSP message
Tells which capabilities are supported/allowed
SS acknowledges the REG-RSP with REG-ACK message
Page 352
SS registration
After successful authentication, the SS registers with network

Response from BS contains CID for a secondary management connection
Secondary management connection is secured
SS and BS determines
Capabilities related to connection set up Parameters required for MAC operation IP version used
Page 353
MAC Management Connections

Upon entering the network, the SS is assigned three management connections in each direction Basic management connection
Exchange of short, time-critical MAC, radio link control management messages with minimal delay Used to quickly adapt to wireless environment
Primary management connection

Exchange of longer, more delay tolerant MAC management messages Authentication and connection setup
Secondary management connection (higher layer)

Exchange of delay tolerant IP-based messages (DHCP, SNMP, TFTP, ToD)
Page 354
IP Connectivity and Configuration File Download
IP connectivity established via DHCP or static IP server

SS establishes the time of the day via the Internet Time Protocol DHCP server provides the address of the TFTP server Configuration file downloaded via TFTP Contains provisioned information
Operational parameters
Page 355
Connection(s) set up
Secondary management connection is also used for setting up one or more transport connections
Transport connections carry the actual user traffic Service flows defines unidirectional transport of packets between the subscriber station and BS
service flows are characterized by a certain set of QoS parameters
Service flows are established using a three-way handshaking establishment procedure
Page 356
Initial Connection Setup
BS passes Service Flow Encodings to the SS in multiple DSA-REQ (dynamic service addition Req) messages
SS replies with DSA-RSP messages Service Flow Encodings contain either
Full definition of service attributes (omitting defaultable items if desired)
Service class name
ASCII string which is known at the BS and which indirectly specifies a set of QoS Parameters
Page 357
Privacy and Encryption

Secures over-the-air transmissions Authentication (SIM, Universal SIM, removable user identity module RUIM)
X.509 certification with RSA PKCS
Strong authentication of SSs (prevents theft of service) Prevents cloning
Data encryption
Currently 56-bit DES in CBC mode IV based on frame number Easily exportable
Message authentication
Key MAC management messages authenticated with one way hashing (HMAC with SHA-1)
Designed to allow new/multiple encryption algorithms Protocol descends from BPI+ (DOCSIS)
Security Associations
A set of privacy information

Shared by a BS and one or more of its client SSs share in order to support secured communications
Includes traffic encryption keys and CBC IVs
Security Association Establishment

Primary SA established during initial registration Other SAs may be provisioned or dynamically created within the BS
Page 359
SS Authorization
Authentication and Authorization

SS manufacturers X.509 certificate binding the SSs public key to its other identifying information
Trust relation assumed between equipment manufacturer and network operator

Possibility to accommodate root authority if required
Authorization Key Update Protocol

The SS is responsible for maintaining valid keys Two active AKs with overlapping lifetimes at all times
Re-athorization process done periodically

AK lifetime (7 days) & grace timer (1 hr)
Traffic Encryption Key Management
Two-level key exchange protocol

Key Encryption Key (symmetric) established with RSA Traffic Encryption Keys (TEK) exchanged with symmetric algorithm negotiated at SA establishment (currently only 3-DES supported) Two sets of overlapping keying material maintained
No explicit key acknowledgements

Key synchronization maintained by 2-bit key sequence number in the MAC PDU header
Traffic Encryption Key Exchange Protocol

Defined by the TEK FSM transition Matrix
Data Encryption
DES in CBC mode with IV derived from the frame number
Hooks defined for other stronger algorithms, e.g. AES Two simultaneous keys with overlapping and offset lifetimes allow for uninterrupted services
Rules for key usage
AP: encryption (older key), decryption (both keys) AT: encryption (newer key), decryption (both keys)
Key sequence number carried in MAC header Only MAC PDU payload (including sub-headers) is encrypted Management messages are unencrypted
IP addressing
Unique address that identifies the network and host
IPv4 consists of 32 bit wide address
4 decimal numbers separated by period A valid address ranges 0.0.0.0 to 255.255.255.255 Class A, B, C, classless, restricted address (0 broadcast, any address starting with 127 is a loopback, a host with binary all 1s is broadcast over the specific network. A host with 0 points to itself. Network address 0 points to its own network) Netmask allows to separate network/host part from address
Performs bit wide AND function
IPv6, consists of 128 bit wide address

PHY Sub-block Diagram, Example

With redundant circuit implementation for STC transmitter
PHY to support three different modes: SC, OFDM and SOFDMA
Digital Domain Subcarrier Allocation + Pilot Insertion Channel Encoder + Rate Matching
Interleaver
Analog Domain
IFFT
DAC Antenna 1
Symbol Mapper
Space Time Encoder Subcarrier Allocation + Pilot Insertion Frequency Domain IFFT DAC Antenna 2 Time Domain
Page 364
PHY
TDD and FDD Adaptive modulation and coding (CC with puncture & RS)
subscriber by subscriber, burst by burst, uplink and downlink Optional Turbo-coding to increase coverage/capacity at the expense of latency and complexity
Point to multipoint Support for adaptive antennas and space-time coding Slot allocation and framing Dynamic frequency selection to detect and avoid interference 256 sub-carriers (192+28+27+8+1) for OFDM Configurable CP length of 1/4, 1/8, 1/16 or 1/32 depending on expected delay Optional signaling support for Adaptive antenna Optional transmit diversity support (Space time block codes)
PHY Coding Rates

If R bps is the input data rate, Nused of FFT, M Modulation order, FEC, then each data subcarrier would carry {(R/Nused)* (2/1)*M} bit rate
Uncoded Modulation Blocks (bytes) RS Code CC Code BPSK 12 (12, 12, 0) 1/2 4-QAM 24 (32, 24, 4) 2/3 4-QAM 36 (40, 36, 2) 5/6 16-QAM 48 (64, 48, 8) 2/3 16-QAM 72 (80, 72, 4) 5/6 64-QAM 96 (108, 96, 6) 3/4 64-QAM 108 (120, 108, 6) 5/6
Coded Blocks Overall (bytes) Coding 24 1/2 48 1/2 48 3/4 96 1/2 96 3/4 144 2/3 144 3/4
Page 366
Latency
Traffic delays through equipment due to processing and propagation Increased delay results in annoying voice echo Voice over IP applications Video conferencing
Simulcast applications
Time out issues for some data applications Issues to reliably controlling remote devices in real time Dynamically adjustment for certain protocols Limited alignment performed by BS prior to mobile handover Latency decreases as the symbol rate increases Latency increases for longer frame size
Higher latency with interleaver

Round trip total latency must be 100ms? (should be 20 ms round trip for VoIP without echo canceller) Latency accumulates linearly with increased number of tandem back haul hops
Handoff (HO)
Operator X backbone network
Gateway
Backhaul connection
Operator Y backbone network
Sector sw C
Handoff (HO) Schemes

Mobile WiMAX performs mobile communication but no mesh mode
Hard handover (HHO) - mandatory Micro-Diversity handover (MDHO) - optional Fast BS switching (FBSS) - optional
BS1
RSL
BS2
RSL
With Hysteresis Rx Threshold Noise Floor
No hysteresis
Page 369
Hard Handoff (HHO)

Handover allows MSs to handover between neighboring BSs while moving across the corresponding coverage areas
This may also be triggered by BS to do an optimal traffic load balancing
BS periodically broadcasts the neighbor advertisement message (MOB_NBR-ADV). Once the handover decision is made, handover process is carried out in two steps Handover preparation: MS or BS may initiate the handover by using the MOB_MSHO-REQ / MOB_BSHO-REQ, the serving BS replies with MOB_BSHO-RSP message containing recommended BSs after negotiation with candidate BSs Handover execution: MS sends MOB_HO-IND message to the serving BS and cuts all communication with serving BS. MS then switches the link and executes ranging with target BS. Then MS negotiates basic capabilities, performs authentication and finally registers with the target BS
Micro-diversity Handoff (MDHO)

Multiple BS serve the MS within the same frame, i.e., Multiple BS transmit the same packet to the MS within the same frame so that MS can perform the diversity combining MS scans the neighbouring BS and maintain a set of BSs that are involved on MDHO - the diversity set
MDHO begins when an MS decides to transmit or receive unicast messages and traffic from multiple BSs in the same time interval
MS communicates with all BSs in the diversity set for UL and DL unicast messages and traffic
For DL MDHO, two or more BSs provide synchronized transmission of data to MS such that diversity combining can be performed at the MS
For UL MDHO, MS transmission is received by multiple BSs such that selection diversity of the received information could be performed
When the long-term CINR of a serving BS in diversity set is less than a threshold, the MS shall send the MOB-MSHO-REQ to delete this BS and update the diversity set
Allows for true soft-handover (make before break) Highly complex, requires synchronization and scheduling above BS layer
Micro-diversity Handoff (MDHO)

UL & DL Comm including Traffic Diversity Set Active BS
Active BS
Area of Neighboring BSs
Neighbor BS MS Active BS Anchor BS Neighbor BS
Only RSL measurement No Traffic

Fast BS Switching (FBSS)

A state where the MS may rapidly switch from one BS to another Multiple BS are ready to serve the MS The Diversity Set is maintained as for MDHO MS communicates with single BS within given OFDMA frame
An Anchor BS is defined within the Diversity Set that MS is registered, synchronized, communicates with for all UL and DL traffic including management messages
MS continuously monitors the signal strength of the active BS and select one to be the anchor BS A FBSS handover begins with a decision by a MS to switch to another Anchor BS using the MOB_MSHO-REQ message The anchor BS can be changed from frame by frame. This means every frame can be sent via different BS in Diversity Set Required synchronization among group of BS using a common timing source Allows for version of soft-handover (communication is never interrupted)
Fast BS Switching (FBSS)

UL & DL Comm including Traffic Diversity Set Active BS
Active BS
Area of Neighboring BSs
Neighbor BS Active BS MS Neighbor BS Anchor BS
Data are transmitted & received but not processed in BS or MS

Only RSL measurement No Traffic

Page 374
Handoff Summary
BS informs neighbouring BSs via MAC Messages
Handover initiated by MS & BS
Process optimized for FBSS (fast BS switching) & MDHO (macro diversity handoff) MS sync with other BSs to estimate associated channel conditions
Handover process allows a MS to switch to another BS in order to improve its QoS
All quality of service and services access are maintained during handovers
Hard HOs use a break before make approach and are typically sufficient for data services. Soft HOs, while complex to implement and administer, are beneficial for applications that require low-latency such as VoIP
Idle Mode / Paging
Allows the MS to traverse a cellular environment and become periodically available for DL broadcast without UL transmission
For MS: save power and operation resources For BS: provide a simple and timely method for alerting the MS to pending MS-directed DL traffic
Page 376
Network Entry Process, SS

SS network entry process
Power ON Negotiate Basic Capabilities Scan for DL Channel SS Authorization and key Exchanage Synchronize with DL of Serving BS Register with Network
Get Time of Day
Transfer Operational Parameters
Obtain UL Parameters Obtain IP Address Ranging
Establish Provisioned Parameters
Network Entry Complete
Page 377
Network Initialization, BS
BS starts by sending beacon
SS first listens for a beacon and then sends a ranging request in the ranging period BS then sends a ranging response. In the ranging response, the BS assigns the SS two connection-IDs called the primary CID and the basic CID. The primary CID is used for further exchange of management messages while the basic CID is used for further periodic ranging exchanges Registration process is required prior to any connection formation. The process involves a registration request from the SS, followed by a registration response from the BS After registration, the SS can request for a connection. A connection request from an SS to the BS elicits a connection response from the BS to the SS The BS and SS are now ready to exchange data with each other
BS Scanning
BS starts with a known channel. Scan all possible channels, until a valid channel is found. PHY Sync is the first step. MAC acquires channel control parameters for DL i.e.; DL channel descriptor (DCD) containing BS ID, modulation, coding, interval. Obtain UCD information containing back off, modulation, coding and message length
Page 379
BS MAC Layer Sequence of Steps

Downlink Period
BS prepares the UL-MAP and allocates the slots to different SSs by keeping in mind the scheduling policy BS sends a beacon which contains the UL-MAP along with the preamble, UCD and the DCD BS sends any pending ranging, registration or connection responses BS inspects its four different queues (one each for UGS, ertPS, rtPS and nrtPS) and sends packets one by one until the DL period finishes The packets are sent in order of their priority i.e., UGS followed by ertPS, rtPS and nrtPS The incoming packets from the link layer are added to the queues according to the flow type
Uplink Period
BS receives the packets sent to it by the SS and passes it on to the upper link layer. The BS does not have any other task to perform in this period
Page 380
SS MAC Layer Sequence of Steps

Downlink Period
SS receives packets sent to it by the BS. Since the packets are broadcasted it checks for the destination in the packet header Incoming packets from the link layer are added to the queues according to the flow type. These packets are sent in the UL frame whenever the SS is allocated a slot.
Uplink Period
SS checks if any of the ranging, registration and or connection requests are still pending
SS reads the UL-MAP and identifies the slots assigned to it SS starts sending the packets in the slots assigned to it in the order of the priority of the packets. This is done by inspecting the four different queues (one for UGS, ertPS, rtPS and nrtPS)
UL-MAP Preparation
BS accesses the queues of all the SSs for four different flows and hence gets to know about the requirements of all the SSs
Accessing the queues of the SSs provides information on current piggybacking and the BW requirements
BS starts filling the UL-MAP as per the bandwidth requirements of the SS
For UGS, ertPS & rtPS flows: the slots are assigned equal to the number of slots required if the total UL slots are not over. UGS flows are given the highest priority
For nrtPS flows: left over slots are divided equally among the SSs which have bandwidth requirement for nrtPS kind of traffic
Network Reference Model, Typical

R2 BS R8 MS R1 BS R8 BS R8 BS R6 ASN R6 Gateway R4
Internet or any IP Network Internet or any IP Network
R2 R6 ASN Gateway R6 Home NSP CSN Visiting NSP CSN R5
Private IP Service Tunneling
ASN
R3
R o a m i n g
Another ASN
NAP
NSP
Page 383
Multi-operator Roaming Framework
Page 384
WiMAX Reference Point

Logical reference interfaces between WiMAX network equipment R1: MS-ASN Implements the air-interface specifications and management plane protocol R2: MS-CSN Authentication, service authorization, IP host configuration and mobility management R3: ASN-CSN QoS policy enforcement, mobility management R4: ASN-ASN Roaming between ASNs R5: CSN-CSN Roaming between CSNs R6: BS-ASNGW Mobility tunnel management, intra-ASN path and inter-ASN tunnels R7: ASNGW-DP & ASNGW-EP An optional protocol for coordinating between two groups identified in R6 R8: BS-BS Control Introduction to WiMAX Technology plane protocol between BSs to ensure fast and seamless HO Page 385
ASN, Access Service Network

Gate way equipment between the BS and the Internet
AAA proxy: transfer of device, user, and service credential to selected NSP AAA and temporary storage of user profiles Provides fast a& efficient radio resource management, QoS policy enforcement and applications per specific subscriber basis Provides Mobility related functions such as handover, location management, paging within ASN and support for mobile IP with foreignagent functionality
May include redundancy and load-balancing among several ASN-GWs

Relay functionality for establishing IP connectivity between the MS and the CSN Admission control functions Cache SS profiles and encryption keys Establishes mobility tunnels with BS and other resources
CSN, Connectivity Service Network

IP address allocation to the MS for user sessions AAA proxy or server for user, device and services authentication, authorization and accounting Policy and QoS management based on the SLA/contract with the user Subscriber billing and inter-operator settlement Inter-CSN tunneling to support roaming between NSPs Connectivity infrastructure and policy control for such services as Internet access, access to other IP networks, ASPs, location-based services, peer-to-peer, VPN, IP multimedia services, law enforcement, and messaging Inter-ASN mobility management and mobile IP home agent functionality
Authentication
Support for device, user, and mutual authentication between MS/SS and the NSP, based on PKMv2
Support for authentication mechanisms, using variety of credentials, including shared secrets, subscriber identity module (SIM) cards, universal SIM, universal integrated circuit card, removable user identity module, and X.509 certificate as long as they are suitable for EAP methods satisfying RFC 4017 Support for global roaming between home and visited NSPs in mobile scenario, including support for credential reuse and consistent use of authorization and accounting through the use of RADIUS in the ASN and the CSN Accommodation of mobile IPv4 and IPv6 security associations management Support for policy provisioning at the ASN or the CSN by allowing for transfer of policy related information from the AAA to the ASN or CSN
Validation & Interoperability

IEEE P802.16c
Published in Jan03 Specifies particular combinations of options Used as basis of compliance testing MAC Profile: ATM and Packet PHY Profile: 1.25-20, 25 & 28 MHz, TDD & FDD Test Protocols: IEEE Standards 802.16/Conformance-0X PICS Test Suite Structure & Test Purposes Radio Conformance Tests Two levels of mobile certifications (Wave 1 & wave 2)
Wave 1 includes basic PHY and MAC functions Wave 2 includes MIMO operation
Page 389
Interoperability Conformance
PHY Tests
Emission, spectral mask, power control and accuracy Interference tolerance at CCI & ACI Relative constellation errors (RCE) vs. symbol & sub-carriers Spectral flatness, crest factor, peak, average & min EVM Error rates, RSSI, SNR, CINR, PCINR, SINR and Rx threshold Frequency error, DynFF
Page 390
IOT, Release 1 Mobile PHY Profile and Certification

Function Wave 1 Wave 2 PUSC PUSC w/ All Subchannels DL Subcarriers Allocation FUSC AMC 2x3 PUSC UL Subcarriers Allocation AMC 2x3 Initial Ranging Handoff Ranging Ranging & Bandwidth Request Periodic Ranging Bandwidth Request 6-bits Fast-Feedback Repetition Randomization Convolutional Coding (CC) Channel Coding Convolutional Turbo Coding (CTC) Interleaving Preamble ID BS Configuration DCD, UCD Packing Fragmentation PHS MAC PDU Manipulation IPv4 IPv6/IPv4 with ROHC BS Initiated Service Flow Initiation SS Initiated Chase Combining H-ARQ BS-BS Time/Freq Synchronization N/A N/A BS-BS Frequency Synchronization N/A N/A Synchronization MSS Synchronization Closed-loop Power Control Power Control Open-loop Power Control Power, Frequency error Transmitter Measurements Trasmit constellation error, Spectrum Release 1 PHY Profile
Comments
Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3
Required in Wave 1 for Band Class 3
Page 391
IOT, Release 1 Mobile PHY Profile and Certification (conti.)

Release 1 PHY Profile Receiver Measurement Function Wave 1 Wave 2 Physical CINR Using Preamble Physical CINR Using Pilots Effective CINR Using Pilots RSSI Measurements Ping Support Ack/Nack Support AWGN RF Amplitude DL 4-QAM DL 16-QAM DL 64-QAM UL 4-QAM UL 16-QAM UL 64-QAM (Optional) Normal MAP Compressed MAP Sub DL-UL-MAP UGS erPS rtPS nrtPS Best Effort 2nd Order Matrix A/B Collaborative Spacial Multiplexing Fast Feedback on DL Mode Selection Feedback w/ 6-bits MIMO DL-UL Chase PUSC w/ Dedicated Pilots AMC 2x3 w/ Dedicated Pilots UL Sounding 1 (Type A) UL Sounding 2 CINR Measurement (group Indication) MIMO Computation Feedback Cycle Comments Required in Wave 1 for Band Class 3
Impairments
Modulation
MAP Support
Data Delivery Methods
MIMO (IO-MIMO for BS)
AAS/BS (IO-BF for BS)
Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3 PUSC, Required in Wave 1 for Band Class 3 Required in Wave 1 for Band Class 3
Page 392
Interoperability MAC Conformance
MAC Test:
802.3 Frame format Protocol Scheduling Admission control
QoS
MIMO Link adaptation
Page 393
IOT, Others
SS Access Point and MSS Access Point: SS/MS connectivity, provisioning and admission control Over the air and end to end security Mobility management Device management UL and DL data exchange Authorization and tunneling for specialized IP services Application layer end to end signaling Power management, compression and data reliability CN1: Control, data and management plane between the RANs and operators core network CN2: control , management and service planes to ASP networks RNSN: Control, data and management plane interfaces between two RNSNs RNSNAP: Control, data and management plane interfaces between an AP and an RNSN Mobility Management: Provisioning, multi-sector handover and end to end mobility management
System Synchronization
For TDD system, the Tx and Rx time frames among BSs/SSs must be synchronized to avoid interference and the SS transmission do not overlap each others as they arrive to BS Timing and frequency offset can influence the performance
Mitigated by reserved pilots & increased CP duration
Frequency offset can influence orthogonality of sub-carriers

Loss of orthogonality can lead to inter-carrier interference Loss of synchronization causes hits during handover Tracking and estimating the position of the frame is necessary for reliable data delivery Timing sync through GPS (cost effective solution but difficult to access open sky if in the basement) For interference mitigation, system-wide synchronization is essential when using tFRAME1 TDD BS-1 DL-TX UL-RX DL-TX UL-RX DL-TX UL-RX
time tFRAME2 DL-TX
BS-2
UL-RX
DL-TX
UL-RX
DL-TX time
Page 395
Network Synchronization
For TDD system, the Tx and Rx time frames among BSs/SSs must be synchronized to avoid interference
Long FEC coding, interleaving and frame structure lead to jitter and wander accumulation Clock accuracy, timing and synchronization is essential for reliable MS handover, MS/SS operation and to minimize interference affect in MIMO configuration GPS timing to aid in synchronizing the network IEEE 1588 timing over IP/Ethernet backhaul
Synchronization distributed from IEEE1588 master clock in the network Less accurate than the GPS
WiMAX network is entirely IP and there is no option of recovering timing signal as there is with TDM application
Page 396
Q&A
Thank you for your attention!
Your feedback and comments are greatly appreciated
Page 397

Introduction to WiMAX Technology: An Overview

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Introduction to WiMAX Technology: An Overview

Uploaded by

Copyright:

Available Formats

Introduction to WiMAX Technology