Design Of Group-Consensus Control Protocols For Multi-Agent Systems

Multi-agent system is an class of important research branch of distributed con-trol system, which describe that multi-agents accomplish a giant complex task to-gether by sharing of information, cooperation, coordination and competition. Fault tolerance, robustness and reliability of the multi-agent system, which has features of high distribution, resource sharing, strong coordination and independence, effec-tively improved. Therefore, multi-agent systems drawn great wide attention among researchers. A large amount of theoretical results have been presented in the con-sensus problems and group consensus problems, some of them have been widely implied to the various fields, such as, intelligent transportation system, wireless sen-sor network, distributed intelligent decision, formation control of mobile robots, and unmanned air vehicles.This paper will study the analysis and design of the controler on the weight-ed group-consensus problems, quasi-average group-consensus problems. The main contents mainly consist of the following four parts:The first part introduces some preliminaries on algebraic graphs and Lyapunov systems which are important for the research of multi-agent systems.This second part investigates weighted group-consensus problem for multi-agent systems with/without time-delay on a connected bipartite graph. For the sys-tem without time-delay, A new controller is designed, using the methods of the matrix theory and the graph theory, it is proved that the multi-agent systems can converge to the arbitrary prescribed weighted state. For the system with time-delay, based on gerschgorin theorem and nyquist’s criterion, a upper bound on the max-imum time-delay that the system can tolerated is obtained. Numerical examples confirm the effectiveness of the proposed method.This third part investigates weighted group-consensus problems of multi-agent systems with/without time-delay on a directed bipartite graph. For the system with-out time-delay, a new controller is designed for weighted group-consensus problems of multi-agent system, based on algebraic graph theory, a necessary and sufficient condition for weighted group consensus is presented. For the system with time-delay, a less conservative upper bound of time-delay is obtained by nyquist criterion and gerschgorin theorem. Finally, two illustrative examples with simulations are studied to support our new results.This fourth part investigates quasi-average group-consensus problems of multi-agent systems with constant time delay and time-varying delay. Switching topolo-gies that have a spanning tree and are balanced are analyzed. For those systems with constant time delay and time-varying delay, based on a reduced dimension model, the method of Lyapunov stability and linear matrix inequality are used to obtain t-wo sufficient conditions for quasi-average group-consensus of multi-agent systems, a less conservative upper bound of time-delay is obtained. One illustrative example with simulations is studied to confirm the effectiveness of the proposed method.