Fault Tolerant Control For Nonlinear Multi-Agent Systems And Its Applications To Multiple Underwater Vehicle

With the extensive application of multi-agent system in the fields of military,industry and agriculture,the research about multi-agent system has deepened gradually.People have higher and higher requirements for the accuracy of machinery equipment.The structures of complex system are becoming larger and larger such as multi-agent system,making physical components widely distributed in the space.These physical components may become faulty which can cause system performance deterioration and lead to instability that can further produce catastrophic accidents.Therefore,it is meaningful to study fault tolerant control of multi-agent system.In this paper,for several high order nonlinear multi-agent systems,the signal transmission faults,actuator faults and sensor faults are considered.Based on adaptive control,dynamic surface technique and other technique,several cooperative fault tolerant control schemes are proposed.This thesis has completed the following research work:Firstly,for a class of nonlinear leader-following systems with the information transmission faults,the problem of cooperative adaptive fault tolerant control is studied.The communication network is a weighted directed graph and external disturbances are considered.By designing a distributed sliding mode adaptive controller for each follower,a distributed fault tolerant control scheme is proposed.This scheme can ensure that the states of all followers asymptotically synchronize the states of the leader.The synchronization errors converge to a small adjustable neighborhood of the origin.Finally,the simulation results demonstrate the effectiveness of the design scheme.Secondly,the problem of cooperative adaptive fault tolerant control is studied for a class of nonlinear leader-following systems with actuator faults.The unmodeled dynamics with prescribed performance for each follower are considered.By introducing a dynamic signal to deal with unmodeled dynamics.Combing dynamic surface technique with backstepping design method,reduce the difficulty of controller design.This proposed scheme can ensure that the output of all followers asymptotically synchronize the output of the leader.The synchronization error meets the prescribed performance.Finally,the simulation demonstrate the effectiveness of the design scheme.Thirdly,for a class of nonlinear multi-agent systems with sensor faults,the problem of cooperative adaptive fault tolerant control is studied.Each follower has different unknown nonlinear dynamics and the system states are unmeasurable.Firstly,a novel hierarchical decomposition method is presented to preprocess the original graph by splitting it into hierarchical structure.By designing an adaptive state observer based on neural network to estimate the unmeasurable states.Based on backstepping design method,a new cooperative fault tolerant control scheme is proposed.The fault tolerant controllers for the followers in different layers can be generated in a sequential order.This scheme can ensure that the outputs of all followers asymptotically synchronize the output of the leader.Finally,the simulation demonstrate the effectiveness of the design scheme.