Adaptive Control Investigation For Nonlinear Multi-agent Systems

Nowadays,the adaptive control technology of multi-agent system has good development prospects in modern control theory and many practical applications,and has become one of the hot researches in the field of control today.With the continuous development and popularization of drone formation,robot collaboration,smart grid and other fields,multi-agent systems are gradually favored by researchers.As an effective way to solve problems in the field of control,adaptive control technology can adapt to the dynamic changes of the system by adjusting the parameters of the controller online,so as to realize the optimal control of complex systems.With the continuous increase of practical application requirements,the research of adaptive control gradually matures and innovates continuously.In multi-agent systems,consensus control is one of the important research areas,and its purpose is to enable multiple agents to maintain consistency while completing tasks cooperatively.Command filtering backstepping technology,as an emerging control method,can effectively solve the computational explosion problem in traditional backstepping design processes.Therefore,this research uses the command filtering backstepping technology as the research basis,and is committed to solving the adaptive consensus tracking control problem of nonlinear multi-agent system with input delays,unmeasurable states,full state constraints,and unmodeled dynamics.The main research work of the paper is summarized as follows:(1)The problem of adaptive control of nonlinear multi-agent systems with input timedelay and unmeasurable state parameters is studied.Firstly,the state observer is established through the radial basis neural network to realize the estimation of the unmeasurable state parameters of the agent.When dealing with the input time-delay problem,the Pade approximation technique is considered to convert the input timedelay into a rational function form,and an auxiliary variable is introduced to be used in the design of the controller.Secondly,this paper uses command filtering backstepping technology to propose a new distributed adaptive consensus control method,which overcomes the "complexity explosion" problem and compensates the filterin gerror caused by the filter.Using the Lyapunov stability theory,it is proved that u n-der the action of the designed control law,the consistent tracking error of the syste mcan converge to a sufficiently small neighborhood of the origin,and all signals inthe closed-loop system are bounded.finally.Simulation results further validate t he effectiveness of the obtained theoretical protocol.(2)The problem of adaptive finite-time control of nonlinear non-strict feedback multiagent systems with full-state constraints and unmodeled dynamics is studied.First,for non-strict feedback systems,scaling is considered during the design process by exploiting the properties of basis functions in radial basis neural networks.At the same time,dynamic signals are introduced by means of auxiliary systems to achieve effective processing of unmodeled dynamics.Secondly,a finite-time command filtering backstepping control scheme is proposed,which not only effectively avoids the"complexity explosion" problem caused by multiple differentiations,but also over-comes the interference of filtering errors.Based on the Barrier Lyapunov Function this paper proves that the state of the system can be constrained within the preset value range.Using the finite-time stability theory,it is proved that the consistent tracking error of the system can converge to the expected small neighborhood of the origin in a finite time,and all signals of the closed-loop system are guaranteed to be bounded.Finally,the feasibility of the adopted strategy is verified by simulation.