Systems & Control Letters 59 (2010) 587–595
Contents lists available at ScienceDirect
Systems & Control Letters
journal homepage: www.elsevier.com/locate/sysconle
Distributed rotating formation control of multi-agent systems
✩
Peng Lin
a,∗
, Yingmin Jia
b
a
Institute of Astronautics and Aeronautics, University of Electronic Science and Technology of China, Chengdu 610054, PR China
b
Seventh Research Division, Beihang University (BUAA), Beijing 100191, PR China
a r t i c l e i n f o
Article history:
Received 19 October 2009
Received in revised form
9 March 2010
Accepted 26 June 2010
Available online 15 September 2010
Keywords:
Formation control
Consensus
Second-order dynamics
Complex systems
a b s t r a c t
This paper investigates collective rotating motions of second-order multi-agent systems. We first consider
rotating consensus problems. Using local relative information, we propose a protocol and give a necessary
and sufficient condition for rotating consensus of the system. Then, we consider rotating formation
control problems. With the help of Lyapunov theory for complex systems, we propose rotating formation
protocols and give sufficient conditions to make all agents move with a specific structure in a circular
channel. Finally, simulation results are provided to demonstrate the effectiveness of our theoretical
results.
© 2010 Published by Elsevier B.V.
1. Introduction
Recent years have witnessed a growing interest in the study
of the distributed control of multi-agent systems, due to re-
cent technological advances in communication and computation,
and important practical applications, including formation control
of unmanned aerial vehicles (UAVs) and spacecraft, cooperative
robotics, and sensor networks. In contrast to the conventional
control systems, where dominant centralized control is the core,
multi-agent systems are concerned with both mobile individual
dynamics and interconnection topologies. The problems related
to multi-agent systems can be roughly categorized as consensus
problems and formation control problems. ‘‘Consensus’’ means to
reach an agreement regarding a certain quantity of interest while
‘‘formation’’ means to achieve and maintain a pre-specified struc-
ture that depends on the specific task. Numerous results have been
obtained to deal with these two problems in many different back-
grounds [1–18]. For example, Jadbabaie et al. and Olfati-Saber et al.
studied discrete-time and continuous-time first-order multi-agent
systems using graph theory or matrix theory in [1,2] respectively.
In [8], Ren introduced many second-order consensus algorithms
for state consensus of systems with the consideration of the satu-
ration of the actor and limited available information while in [9],
✩
This work was supported by the NSFC (60672029, 60334030, 60774003) and
the National 973 Program (2005CB321902).
∗
Corresponding author.
E-mail addresses: Lin_peng0103@sohu.com (P. Lin), ymjia@buaa.edu.cn (Y. Jia).
Tanner et al. introduced a set of control laws that enable the
second-order agents to generate stable flocking motion.
In nature and society, there exist a kind of collective motions
widely such as the motion of celestial bodies, flocks of birds fly-
ing around a closed circuit course and schools of fish swimming
along an approximately circular orbit, which can find important
potential applications in formation flight of satellites around the
earth, spacecraft docking, circular mobile sensor networks, and so
on. However, most of the existing results cannot be straightfor-
wardly applied to imitate or explain such motions. Only recently,
in [10], Sepulchre et al. formulated a new rotating formation con-
trol problem from the use of autonomous underwater vehicles
(AUVs) to collect oceanographic measurements and gave a design
methodology to stabilize relative equilibria in a model of identical,
steered particles moving in the plane at unit speed. In the mean-
time, in [11], Pavone et al. studied this kind of collective motions
and proposed a cyclic pursuit policy by letting one vehicle pur-
suit another vehicle along the line of sight rotated by a common
offset angle, and Ren extended the results of [11] by introducing
a rotation matrix to an existing consensus algorithm for double-
integrator dynamics in [12].
Motivated by the work of [10], we investigate collective
rotating motions of second-order multi-agent systems. As most of
collective rotating motions in reality do not maintain a constant
magnitude of velocity and usually start from unregular initial
velocities, the control input of each agent considered in this paper
is taken as the derivative of the velocity, different from [10], where
the control input is the angular acceleration and the magnitude of
the velocity of each agent remains unchanged. This difference leads
to completely distinct analysis. The analysis in this paper relies on
0167-6911/$ – see front matter © 2010 Published by Elsevier B.V.
doi:10.1016/j.sysconle.2010.06.015