Research On Fault-Tolerant Consensus Method For Multi-Agent Systems With Spatio-Temporal Characteristics

In recent years,many scholars have paid more attention to multi-agent systems,and many achievements have been made in the research of multi-agent systems in various aspects.However,with the continuous development of multi-agent systems,the complexity of the system is constantly upgraded.A simple failure may destabilize the whole complex system and cause irreparable loss of property and even personal safety,so when designing a complex system,the design of fault-tolerant protocol occupies a very important position.It is worthwhile to study how to ensure the safe operation of a complex system in case of subsystem failures,communication failures,or harsh environments.In this paper,the multi-agent systems with spatio-temporal characteristics and actuator failures are taken as the target system,and two types of multi-agent systems,leaderless and leader-following,are studied separately.To ensure the normal operation and safety and reliability of the target multi-agent systems,this paper applies an improved faulttolerant consensus strategy to the target system based on the research of a wide range of scholars to enhance the fault-tolerant of the target system.The main research contents and results include the following:(1)A partial differential equation approach is used to model the nonlinear multi-agent systems with spatio-temporal characteristics,which shows the spatial and temporal characteristics of the studied multi-agent systems well.A fault model for the actuator failure is developed and embedded into the dynamics equations of the studied nonlinear multi-agent systems to obtain the target system model.A fault-tolerant consensus protocol is selected and optimized for the leaderless and the leader-following multi-agent systems,respectively.Based on the idea of Lyapunov stability,a Lyapunov potential function is constructed and it is derived that the target system can eventually achieve consensus with this fault-tolerant consensus protocol when it encounters a set of actuator failures.Finally,the correctness of this method is demonstrated by a numerical simulation experiment.(2)For the multi-agent systems with actuator failures,adaptive fault-tolerant control protocols are used for both leaderless and leader-following multi-agent systems to compensate for the effects of actuator failures on the system.The dynamics of the target system are still modeled using partial differential equations,and the system model and actuator failures described in the previous section are improved to take into account the complexity of real systems.The newly constructed Lyapunov potential function demonstrates that the established adaptive fault-tolerant consensus protocol enables the system to achieve consensus and the goal of consensus of the target multi-agent systems.The final numerical simulation experiments also validate the correctness of such an approach.(3)The two fault-tolerant consensus protocols studied in this thesis can be tried to be applied to multi-unmanned vehicles,multi-unmanned aircraft,multi-unmanned boats,and various types of intelligent systems in subsequent research to increase the security of the systems.The fault-tolerant consensus protocols can solve the actuator failure problem of dynamic systems with changing time and position,and better guarantee the reliable operation of such dynamic systems for multi-agent intelligence systems with spatio-temporal characteristics.