Professional Documents
Culture Documents
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 61, NO. 11, NOVEMBER 2016
AbstractIn this technical note, symmetries in large-scale dissipative and passive control systems are considered. In the framework of dissipativity and passivity theory, stability conditions
for large-scale systems are derived by categorizing agents into
symmetry groups and applying local control laws under limited
interconnections with neighbors. Building upon previous studies
on stability of (Q, S, R)-dissipative large-scale systems, we show
that for cyclic and star-shaped symmetric systems there exists an
upper bound on the number of subsystems that can be added to
preserve stability of dissipative systems. In cyclic and star-shaped
symmetric systems, the subsystems can be heterogeneous as long
as they satisfy the same dissipative inequalities. Approximate
symmetry with respect to interconnections is also considered and
the robustness of the results is demonstrated.
Index TermsDissipativity, large-scale systems, passivity, stability of nonlinear systems, symmetry.
I. I NTRODUCTION
Symmetry is a basic feature of shapes and graphs and can be found
in many real-world networks, such as the Internet and power grids, as
a result of tree-like or cyclic growth process. Since symmetry is related
to the concept of a high degree of repetitions or regularities, common in
nature, the study of symmetry has been undertaken in many scientific
areas, such as mathematics (Lie groups), quantum mechanics, and
crystallography in Chemistry.
The concept of symmetry has been studied in the classical theory
of dynamical systems. For example, to simplify the analysis and
synthesis of large-scale dynamical systems, it is of interest to consider
smaller symmetric subsystems with simplified dynamics, which may
potentially simplify the analysis in control, planning or estimation [1],
[2]. When dealing with multi-agent systems with various information
constraints and communication protocols, under certain conditions
such systems can be expressed as or decomposed into interconnections
of lower dimensional systems, which may lead to better understanding
of system properties such as stability and controllability [3], [4].
Early research on symmetry in dynamical systems can be found
in [5][7]. Symmetry in the context of distributed systems containing
multiple instances of identical subsystems are studied in [8] and [9],
where the controllability of the entire class of systems was determined
by reducing the model and examining a lower order equivalent class.
Manuscript received April 20, 2015; revised January 30, 2016; accepted
February 3, 2016. Date of publication February 11, 2016; date of current
version October 25, 2016. This work was supported by the National Science
Foundation under Grants CNS-1035655 and CNS-1446288. Recommended by
Associate Editor C. De Persis.
The authors are with the Department of Electrical Engineering, University of
Notre Dame, Notre Dame, IN 46556 USA (e-mail: vghanbar@nd.edu; pwu1@
nd.edu; pantsakl@nd.edu).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TAC.2016.2528824
0018-9286 2016 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 61, NO. 11, NOVEMBER 2016
3677
Fig. 1. Interconnected multi-agent system. Hi1 , . . . , HiN are constant interconnection matrices and Hi is constant local feedback matrix.
Fig. 2. Star-shaped symmetry.
The technical note is organized as follows: In Section II, we introduce a background on dissipativity, passivity and symmetry in dynamical systems. Section III describes stability conditions for dissipative
and passive systems with star-shaped symmetry interconnections.
Section IV discusses stability conditions for dissipative systems with
cyclic symmetry, and finally, the concluding remarks are presented.
N
Hij yj ,
i = 1, . . . , N
(6)
j=1
(1)
where ui is the input to the subsystem i, yi is its output, uei is an external input, and Hij are constant matrices. The system inputs and outT
T T
] .
puts are stacked as u = [uT1 , uT2 , . . . , uTN ] , y = [y1T , y2T , . . . , yN
Defining a constant interconnection matrix, H = [Hij ], the linear
interconnected system can be represented by
u = ue Hy.
Lemma 1: [11] Let the ith subsystem have finite gain input-output
stable with gain i , for i = 1, . . . , N , and suppose that each subsystem
has only one input and one output. Define = diag{1 , . . . , N } and
A = H, where H can be obtained from (7). Then, if there exists a
diagonal positive definite matrix D such that
D AT DA > 0
t2
(u, y)dt V (x(t2 )) V (x(t1 )) .
(2)
t1
(4)
(5)
(7)
(8)
(9)
(10)
3678
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 61, NO. 11, NOVEMBER 2016
A
12
...
...
12
A
0
...
0
21
.
..
..
.
.
A= .
0
.
..
..
..
.
.
.
0
0
...
0
A
21
h0
h21
..
H =
.
.
..
h12
h0
h21
0
..
.
0
...
0
..
.
...
...
..
.
0
...
h12
0
..
.
(11)
h0
where block entries of H are feedback scalar gains since we consider each subsystem to be single-input single-output. Thus, the linear feedback interconnections are described by u0 = ue0 h0 y0
N
i=1 h12 yi , ui = uei h21 y0 h0 yi , i = 1, . . . , N .
Theorem 1: Consider a finite gain system extended by N starshaped finite gain symmetric subsystems with symmetric interconnections matrix H. If
N<
(A)
1 21 )
(12 A
(12)
0
1
h0
h21
h12
.
h0
A
12
, where 12 =
21
A
already has positive leading princi0 h12 and 21 = 1 h21 . Since A
12 A1 21 > 0
pal minors, A is an M -matrix [19] if and only if A
(Schurs theorem, [20, p. 7]). Then, based on Lemma 1, the extended
system is BIBO stable.
The new matrix A
can be written as A
=
H12
...
...
H12
H0
H21
H1
0
...
0
..
..
..
.
0
.
(13)
H = .
.
.
.
..
..
..
0
0
...
0
H1
H21
and therefore, the linear
feedback interconnections can be described by
u0 = ue0 H0 y0 N
i=1 H12 yi , ui = uei H21 y0 H1 yi , i =
1, . . . , N where H12 and H21 , are constant matrices containing feedback gains between the base system and subsystems. Also, H0 and H1
are constant local feedback matrices of base system and subsystems
respectively. We now have the following result:
Theorem 2: Consider a (Q, S, R)- dissipative base system extended
by N star-shaped symmetric (q, s, r)-dissipative subsystems as in
S,
R)-dissipative
with
Fig. 2. System s is (Q,
= Q 2SH + H T RH
Q
1
S = S {RH + H T R}
2
= R.
R
(14)
S,
and R
given by (14).
(Q,
with Q,
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 61, NO. 11, NOVEMBER 2016
(Q)
q
N < min
,
(15)
T
(H21
rH21 + (
q H)1 T ) H
where
= H T RH0 + SH0 + H T S T Q > 0
Q
0
0
q = H1T rH1 + sH1 + H1T sT q > 0
T T
T
= SH12 + H21
s H0T RH12 H21
rH1
T
H = H12 RH12 .
Proof: In Theorem 2, we have shown that the enlarged system is
S,
R) dissipative. For simplicity, we consider D = I in Lemma 2
(Q,
such that (10) holds. Based on the linear interconnections, we have
matrix of feedback interconnections, H, for star-shaped symmetry and
for asymptotic stability, we require
= SH + H T S T H T RH Q > 0.
Q
This implies
T
rH21
Q N H21
T
T
=
Q
..
...
(16)
>0
3679
(Q)
, i = 1, 2, . . . , N
max i qi1 iT
(19)
(17)
T
where
T T
T
s H0T RH12 H21
rH1
= SH12 + H21
T
= q IN H12
RH12 circ ([11 1]) .
A
B
> 0 if and only if
According to the Schurs theorem,
T
D
B
1 T
D > 0 and A BD B > 0. Thus, there is an upper-bound on
such symmetric extension which is given by (15).
Remark 2: The right-hand side of (15) consists of two parts, the first
part implies that the base system will become unstable if the sum of
a positive feedback gains from all subsystems exceed certain value,
while the second part implies that the subsystem will become unstable
if the positive feedback from the base system is too strong.
Notice that hitherto we have considered all subsystems being
(q, s, r)- dissipative. No restrictions were placed on the actual dynamics which may be different from each other. Thus, the above results
apply to heterogeneous systems as well, as long as they satisfy the
inequality (2).
(Q)
(18)
N<
1
( q T )
= ((H0 + H T )/2) > 0, q = ((H1 + H T )/2) > 0, and
where Q
0
1
T
)/2. The proof is similar to the proof of Corollary 1
= (H12 + H21
and is thus omitted here.
Note that passive systems are special cases of dissipative systems
where Q = 0, S = (1/2)I, R = 0 for the base system and q = 0, s =
(1/2)I, r = 0 for subsystems. One can directly get this result from
(15) in Corollary 1.
H0
H12
...
H12
H21
(20)
H .
..
H
H21
= circ([v0 v1 vN1 ]) is a circulant matrix with first row
where H
= P T HP
and H
can be written as H
=
[v0 v1 vN1 ]. Here, H
N1
. P is a matrix in companion form with
v0 I + v1 P + vN1 P
the last row [1, 0, . . . , 0].
We now consider subsystems with dynamics (5) and linear interconnected system (7) with (q, s, r)-dissipative subsystems, connected via
a cyclic interconnection as shown in Fig. 3.
Corollary 4: Consider a (Q, S, R)- dissipative system extended
by N cyclic symmetric (q, s, r)- dissipative subsystems as in Fig. 3.
System c is asymptotically stable if
(Q)
q
N < min
(21)
, T
T
T
(H21
rH21 + N 1 N
) H12
RH12
3680
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 61, NO. 11, NOVEMBER 2016
where
ACKNOWLEDGMENT
sH1 +
q IN
circ ([11 1])
+ s (H)
+ (H)
q
H)
q = r(H)(
T T
T
= SH12 + H21
s H0T RH12 H21
rH
N = [ ]
=
=
(H)
H1T rH1 +
T
H12
RH12
N
N1
N1
j=0
j=0
vj ji =
H1T s
vj e
2ij
N
R EFERENCES
(Q)
T
(N 1 N
)
(22)
where
=
T
H1 + H1T
H12 + H21
, =
, N = [ ].
2
2
N
[1] J. Cortes, S. Martinez, J. P. Ostrowski, and H. Zhang, Simple mechanical control systems with constraints and symmetry, SIAM J. Control
Optimiz., vol. 41, no. 3, pp. 851874, 2002.
[2] Y. Wang and M. Morari, Structure of hierarchical linear systems with
cyclic symmetry, Syst. Control Lett., vol. 58, no. 4, pp. 241247, 2009.
[3] D. D. Siljak, Large-Scale Dynamic Systems: Stability and Structure.
New York, NY, USA: North-Holland, 1978, vol. 310.
[4] A. N. Michel, On the status of stability of interconnected systems, IEEE
Trans. Autom. Control, vol. AC-28, no. 6, pp. 639653, Jun. 1983.
[5] F. Fagnani and J. C. Willems, Representations of symmetric linear dynamical systems, SIAM J. Control Optimiz., vol. 31, no. 5, pp. 12671293,
1993.
[6] J. Grizzle and S. Marcus, The structure of nonlinear control systems
possessing symmetries, IEEE Trans. Autom. Control, vol. AC-30, no. 3,
pp. 248258, Mar. 1985.
[7] S. Y. Zhang, The structures of symmetry and similarity of complex
control systems, Control Theory Appl., vol. 11, no. 2, pp. 231237, 1994.
[8] M. B. McMickell and B. Goodwine, Reduction and nonlinear controllability of symmetric distributed robotic systems with drift, in Proc. IEEE
Int. Conf. Robotics Autom., 2002, vol. 4, pp. 34543460.
[9] M. B. McMickell and B. Goodwine, Motion planning for nonlinear symmetric distributed robotic formations, Int. J. Robot. Res., vol. 26, no. 10,
pp. 10251041, 2007.
[10] J. C. Willems, Dissipative dynamical systems part i: General theory,
Arch. Ration. Mechan. Anal., vol. 45, no. 5, pp. 321351, 1972.
[11] P. Moylan and D. Hill, Stability criteria for large-scale systems, IEEE
Trans. Autom. Control, vol. AC-23, no. 2, pp. 143149, Feb. 1978.
[12] S. Hirche and S. Hara, Stabilizing interconnection characterization for
multi-agent systems with dissipative properties, in Proc. 17th IFAC
World Congr., 2008, pp. 15711577.
[13] M. Arcak and E. D. Sontag, A passivity-based stability criterion for
a class of interconnected systems and applications to biochemical reaction networks, in Proc. 46th IEEE Conf. Decision Control, 2007,
pp. 44774482.
[14] E. D. Sontag and M. Arcak, Passivity-based stability of interconnection
structures, Recent Advanc. Learn. Control, vol. 371, pp. 195204, 2008.
[15] M. Arcak, Diagonal stability on cactus graphs and application to network stability analysis, IEEE Trans. Autom. Control, vol. 56, no. 12,
pp. 27662777, Dec. 2011.
[16] F. Castanos, R. Ortega, A. J. van der Schaft, and A. Astolfi, Asymptotic
stabilization via control by interconnection of port-hamiltonian systems,
Automatica, vol. 45, no. 7, pp. 16111618, 7 2009.
[17] R. Olfati-Saber, J. A. Fax, and R. M. Murray, Consensus and cooperation in networked multi-agent systems, Proc. IEEE, vol. 95, no. 1,
pp. 215233, 2007.
[18] P. Wu and P. J. Antsaklis, Symmetry in the design of large-scale complex
control systems: Some initial results using dissipativity and lyapunov stability, in Proc. 18th Mediterranean Conf. Control Autom. (MED), 2010,
pp. 197202.
[19] M. Araki, Application of m-matrices to the stability problems of
composite dynamical systems, J. Mathemat. Anal. Appl., vol. 52, no. 2,
pp. 309321, 1975.
[20] S. P. Boyd, L. El Ghaoui, E. Feron, and V. Balakrishnan, Linear Matrix
Inequalities in System and Control Theory. Philadelphia, PA, USA:
SIAM, 1994, vol. 15.
[21] P. Tabuada, A. D. Ames, A. Julius, and G. J. Pappas, Approximate
reduction of dynamic systems, Syst. Control Lett., vol. 57, no. 7,
pp. 538545, 7 2008.