Research On Containment Control And Dynamic Targets Surrounding For Multi-agent System

In recent years,the multi-agent systems develop rapidly and have become an important tool for modeling and analyzing the complex system.The formation control is an important research direction in the multi-agent systems.The research contents regarding formation control at home and abroad mainly cover formation production,formation maintenance,formation transformation and formation obstacle avoidance.An important object of researching multi-agent systems to realize formation control is containment control and surrounding control.This paper studies the containment control and surrounding for dynamic targets of the multi-agent system pursuant to the related knowledge(such as Algebraic Graph Theory,matrix theory and Lyapunov stability).The main research contents are as follows:(1)As for the single target dynamic tracking and surrounding control of the first-order nonlinear multi-agent system,this paper puts forward a joint tracking and surrounding algorithm.The sequential multi-hypothesis tracking algorithm is adopted for the tracking process.After the targets state is obtained in the clutter environment,a distributed controller is designed based on consensus theory.The agent state information minus target state information is used to transform the surrounding problem into the stability of system errors,the Lyapunov stability theory is used to obtain the conditions of multi-agent system and dynamic single target errors tending to be constant.The dynamic tracking and surrounding for single target by the multi-agent systems are realized.(2)As for the dynamic surrounding problem of second-order linear/non-linear multi-agent systems with continuous time-varying topologies,a distributed controller with position and velocity items is designed based on containment control theory.The problem of surrounding is transformed by a system model into the problem of stability of system errors.The Algebraic Graph Theory and Lyapunov stability theory are used to obtain the conditions that the stability of system errors,which is related to the eigenvalue of the Laplace matrix of its communication topology.Then,the dynamic surrounding of multiple targets by the multi-agent systems is accomplished by designing the deviation vectors.(3)As for the containment control of multi-agent systems with time-varying delay and communication noise under the semi-Markov switching topology,a distributed control protocol with time delay is purposed.Because the communication channel between agents is affected by communication noise,the robust control methods are used to reduce the effect of communication noise on the system.On the basis of upper bound and lower bound of semi-Markov process jump probability,Jensen inequation,free weight matrix and linear matrix inequality,this paper gives the sufficient condition for the follower to converge to the target's convex hull,and obtains the method for designing the gain matrix of the controller.