Research On Consensus For Multi-agent Systems And Formation Systems For Multiple Robots

Currently,multi-Agents and its cooperative control have been widely used in network resource optimization,traffic vehicle formation,environmental monitoring,target search and other fields.They have become a hot topic in the control discipline.The consistency problem is the foundation and core of the multi-agent system,and the convergence speed of the consistency algorithm plays a crucial role.Therefore,based on the existing research work,this paper uses control theory,algebraic graph theory and matrix theory and stability theory to design a multi-agent system consistency algorithm,improve the convergence speed of the algorithm,and apply the algorithm to multi-robot formation.The main contents are as follows:Firstly,for the continuous first-order multi-intelligent system,considering the time lag as a positive factor,a consistent algorithm with outdated state is proposed.The algorithm not only contains the current state of the agent,but also contains the obsolete state of the agent.At the same time,the theoretical analysis of the convergence and stability of the algorithm is carried out,and the influence of the delay parameter selection on the convergence speed is analyzed.Aiming at the selection of delay parameters,a grid greedy search algorithm is proposed to quickly determine the optimal delay parameters of the algorithm.When the obsolete state is chosen reasonably,the continuous consistency algorithm converges faster than the normal consistency algorithm.The effectiveness of the algorithm is verified by simulation.Secondly,for the continuous second-order multi-agent system,a two-layer neighbor continuous consistency algorithm with obsolete state is proposed.The algorithm not only contains the current state of a layer 2 neighbor,but also contains the obsolete state of a layer 2 neighbor,and theoretically analyzes the convergence and stability of the algorithm.When the obsolete state is chosen reasonably,the two-layer neighbor continuous consistency algorithm with obsolete state converges faster than the one-layer neighbor consistency algorithm.The effectiveness of the algorithm is verified by simulation.Finally,A continuous consistency algorithm with bounded input is proposed.This algorithm is applied to the first-order multi-robot formation system.Compared with the consistency algorithm without input constraints,the algorithm can optimize the initial motion state of the robot and reduce the system's The initial momentum,makes the formation process smoother and more reasonable.