Fault Tolerant Consensus Of Multi-Agent Systems With Actuator Fault

A multi-agent system consists of multiple agents that can interact with the envi-ronment and have certain information processing capabilities.By building large and complex systems into small,easy-to-manage systems that communicate and coordi-nate with each other,multi-agent systems can solve the complex problems that a single agent and single-layer system are difficult to solve.Therefore,multi-agent systems have a wide range of applications in many fields,such as telescope array,UAV formation,satellite attitude coordination and so on.The main research task of multi-agent system consistency is to design control algorithms to achieve the consistent state of each agent under certain network topology and single agent system dynamics conditions,such as consistent posture of speed and orientation in satellite formation and drone formation.In practical engineering,as the scale of multi-agent systems expands,the probability of failures in the system increases.In a multi-agent system,faults can be classified into network topology faults,actuator faults,sensor faults,and other component faults.The fault of the actuator can cause deviation between the actual output and the control signal,thereby reducing system performance and even making The system is unstable.In addition,compared with sensors,controlled objects,network communication faults,etc.,it is difficult and costly to implement hardware redundancy to improve reliability for actuator faults.Therefore,rational design of fault-tolerant controllers has high prac-tical value.In this thesis,the fault-tolerant consensus control for multi-agent system subject to actuator faults is studied.The technical methods adopted include adaptive sliding mode control,adaptive neural network control and active fault-tolerant control.The main research contents of the thesis include the following aspects:1.Adaptive sliding mode fault-tolerant control.Sliding mode control has a good performance of high control accuracy and is robust to system parameters and external disturbances.In this dissertation,a discontinuous sliding mode fault-tolerant consensus algorithm is firstly designed for the second order nonlinear multi-agent system.How-ever,the symbol function in classical sliding mode control will cause the actuator to oscillate,which will adversely affect the normal operation of the system.A continuous scheme optimizes the symbol function in the controller.On the other hand,when the system parameters and the degree of actuator fault are unknown,the control gain of the sliding mode control is not easy to determine.This dissertation introduces an adaptive algorithm so that the control gain can be adaptively adjusted.2.Neural network adaptive fault-tolerant control.Neural network adaptive control has a good processing ability for complex system dynamics and strong nonlinearity.This dissertation designs a neural network adaptive consensus algorithm for general nonlinear multi-agent system and non-affine nonlinear multi-agent system.The fault-tolerant consensus algorithm designed for the general nonlinear multi-agent systems contains a negative feedback term based on consensus error and an adaptive neural network term to compensate for unknown nonlinear dynamics.For non-affine nonlinear multi-agent system,the input of the adaptive neural network consensus algorithm is designed to contain both system state information and consensus error information.The negative feedback term in the neural network weight update law is multiplied by the absolute consensus error,which is beneficial to reduce the consistency final error and improve the consistency accuracy.3.Fault-tolerant consensus based on fault diagnosis information.In this thesis,a fault-tolerant control scheme with fault diagnosis is investigated for the consensus problem and containment problem of the second-order multi-agent systems.The fault-tolerant controller includes an ideal control algorithm and a fault diagnosis module.When no fault occurs,the multi-agent systems can achieve the goal under the action of ideal control algorithm.When a fault occurs,the controller adjusts its parameters according to the fault diagnosis information.Theoretical results show that the fault-tolerant algorithm can guarantee good system performance as the fault diagnosis accu-racy is within a certain range.