Fault Diagnosis And Fault Tolerant Control For Linear Multi-agent System

In the past few years,with the continuous progress of modern technologies such as communication,network and computer,multi-agent technology has gradually become a hot topic for researchers and has been widely used in multi-unmanned vehicle,multi-unmanned aerial vehicle.By comparison with single agent,multi-agent has more powerful functions,which enables it to complete increasingly complex tasks.However,while the functions of the multi-agent system become more powerful,the safety and reliability of the multi-agent system also become more vulnerable due to its own more complex topology.In the multi-agent system,agents interact with each other.Therefore,when a fault occurs,the occurrence of fault is likely to cause the output of the whole multi-agent system to be shifting and unable to achieve the desired trajectory,and in serious cases it may even cause the collision between the agents and lead to the collapse of the whole system.Therefore,it is very important to study the fault tolerant control scheme of multi-agent for improving the security and reliability of multi-agent systems.In this thesis,fault diagnosis and fault-tolerant control strategies for multi-agent systems are investigated.The main research contents are shown as follows:(1)The strategy to solve the problem of fault-tolerant control for uncertain multi-agent systems is investigated and the corresponding fault-tolerant control protocol is developed.The mathematical model of linear multi-agent with 1 leader and N followers is taken as the research object,and uncertain parameters and disturbances are considered in the mathematical model of multi-agent systems.When an actuator fault occurs in one of the followers,the fault is likely to affect the motion trajectory of the whole multi-agent.Then,a robust adaptive observer based on the local relative output estimation error is proposed to estimate actuator faults.based on the obtained fault estimation,an output feedback fault-tolerant control protocol based on relative output tracking error is designed to control the post-fault multi-agent system.Finally,a multi-agent system with four agents is used as a simulation object to prove the effectiveness of the proposed algorithm.(2)The solution of fault-tolerant control for multi-agent systems with time delay is discussed and the corresponding fault-tolerant control strategy is given.Due to the reason of data transmission,operation and so on,there is likely to be time delay in multi-agent systems.Firstly,a dynamic disturbance observer based on relative output estimation error is used to estimate the fault.Then,based on the estimated value of fault,a dynamic compensation control strategy based on relative output tracking error is proposed to offset the impact of the fault on the system.Finally,the multi-agent system with five multiple unmanned Aerials(UAVs)is used as a simulation example to demonstrate the effectiveness of the proposed algorithm.(3)When actuator fault and sensor fault occur at the same time,the problem of fault-tolerant control of multi-agent systems with undirected graph topology is investigated and the corresponding solution is given.Firstly,the multi-agent system based on local relative output information can be decoupled into several independent subsystems by coordinate transformation.Then,a fault diagnosis strategy based on unknown input observer is designed for the decoupled subsystem.Based on the estimated fault information,a dynamic compensated fault-tolerant control protocol based on the relative output tracking error is constructed to ensure that the output trajectory of post-fault multi-agent system is consistent with the expected trajectory.Finally,a simulation example is given and satisfactory simulation results are obtained,which proves the effectiveness of the algorithm.