Distributed Containment Control Of Multi-Agent Systems With Velocity And Acceleration Saturations

Recently,multi-agent control has aroused extensive attention in control field.One of the basic issues is the consensus control of multi-agent,whose basic goal is to make the state of the group converge to a common point.The containment control is the generalization of the consensus issue,and its goal is to make the states of the group converge to a convex set formed by all leaders.Velocity and acceleration constraints are inevitable in the practical applications of distributed control algorithm,but most of the existing results of containment control are based on linear control algorithm,and the influence of these two basic nonlinear constraints has not been considered.A simple example given in this paper shows that even for a simple multi-agent system consisting of three agents,the coupling between dimensions needs to be considered if there are velocity constriants,otherwise the goal of containment control will be difficult to achieve.The main contribution of this paper is to solve the distributed containment control problem for multi-agent system modeled by double-integrator under velocity and acceleration constraints.Meanwhile,the topologies describing the information exchange are allowed to be time-varying as long as the basic connectivity conditions are satisfied.The work includes the following aspects:1.proposed a class of distributed circular saturation controllers to ensure that the velocity and acceleration meet the expected constraints,and proved that the model of closed-loop system composed of controlled agent and circular saturation controller is unchanged after coordinate rotations,and then the high-dimensional containment control problem is transformed into set convergence problem in one dimension.2.gave the invariance properties of every controlled agent in any dimensions,and proved that if the topologies of multi-agent system satisfy some basic connectivity conditions,the invariance property of single agent can derive the invariance properties of agent group and the set convergence properties of outputs of all agents.3.extended the case of static leader to the case of movable leader,and proved the final set convergence of multi-agent system,that is,the outputs of all agents can converge to the set the leaders converge to.The main results of this paper are based on theoretical analysis and verified by simulation experiments.