Research On The Finite-time Consensus Problem Of Multi-agent Systems Based On Adaptive Backstepping Control

With the complexity of industrial system structure and diversification of functions,multi-agent systems have attracted the attention of experts and scholars,and the consensus control of multi-agent systems has become the focus of research.In order to achieve faster system convergence,scholars have proposed finite-time control theory.Compared with the asymptotic stability when time tends to infinity,finite-time control theory has the advantages of faster convergence speed and better robustness.In practical engineering applications,the occurrence of phenomena such as time delay and unknown nonlinear interference is commonplace,attributable to the influence of physical factors,including but not limited to,environmental and climatic conditions.At the same time,with the complexity of the physical model of the controlled object,the applicability of the linear control theory becomes limited,and more complex control methods and techniques are required to achieve the stability of the system.Backstepping,as a control method widely used in highorder systems,has the advantages of simple structure and strong robustness.Therefore,this thesis studies the influence of the above phenomena on the finite-time stability of nonlinear multi-agent systems by combining the backstepping method,and proposes a finite-time adaptive neural network controller.The main research work includes the following aspects:(1)A class of strict-feedback nonlinear multi-agent systems with state time-delays and unknown disturbances is studied in the finite-time consensus problem.Firstly,in order to overcome the influence of time delay on the system,Lyapunov-Krasovkii functional method and Young’s inequality are introduced in the analysis and design process.Secondly,the approximation characteristic of radial basis function neural network is utilized to handle the unknown term in the system.Based on the backstepping approach,a class of finite-time adaptive neural network controllers is designed,which ensures that the tracking error of the system converges to a minimal adjustable zero neighborhood in a finite time,and all the closed-loop signals in the system are bounded.Finally,the effectiveness and reliability of the proposed scheme are illustrated by simulation experiments.(2)Design and develop a graphical visualization interface based on MATLAB GUI,including four functional modules:MATLAB code simulation image,parameter input,running time setting,software startup and shutdown.Among them,the MATLAB code simulation module presents simulation graphics,which is convenient for visually displaying the system tracking effect.Taking the strict feedback nonlinear multi-agent system with state time delay and unknown disturbance as the control object,input different parameters to achieve the ideal tracking effect,customize the system running time,and control the start and stop of the system with the simulation and close buttons.The output trajectory of the leader-following system and the system tracking error curve are clearly visible in the simulation image,which verifies the effectiveness of the control scheme.