Research On Fault-tolerant Consensus Control For Heterogeneous Multi-agent Systems

The premise of multi-agent systems(MAS)function implementation is that each agent can operate normally.Once a single or multiple actuators fail during operation,tasks such as MAS consensus cannot be satisfied.Therefore,MAS fault-tolerant control research has important practical significance.However,in practical applications,heterogeneous multi-agent systems(HMAS)are more common due to the more or less heterogeneous characteristics of MAS,such as the inconsistent parameters or different model structures of each agent in systems such as hybrid UAV formation control and air-ground cooperative control.In view of this,this paper proposes different fault-tolerant consensus algorithms for HMAS,solving the fault tolerance consensus control problem of different types of HMAS.The main research contents are as follows:1)For the first-order and second-order HMAS with actuator additive timevarying faults and external bounded energy disturbances,the problem of active fault-tolerant consensus control is studied.Firstly,the external dis turbance of the system is estimated by establishing a time-varying gain disturbance observer.Then,the integral sliding mode surface is constructed by using the system consensus error variable,and the adaptive sliding mode fault-tolerant controller is designed by combining the estimated value of the disturbance observer.The actuator additive fault is estimated by the adaptive algorithm to obtain the fault boundary value.The estimated external disturbance value and fault boundary value information are used to complete the reconstruction design of the controller,and finally the HMAS realizes fault-tolerant consensus control.2)For the same-order different-dimensional linear / nonlinear HMAS with actuator additive time-varying faults and external bounded energy disturbances,the problem of the fault-tolerant consensus control problem is studied.For linear HMAS,firstly,considering the unknown actuator fault information of the actual system and the influence of external disturbance on the observer,a robus t sliding mode observer with adaptive law correction is proposed to reconstruct the actuator fault.Then,the robust fault-tolerant controller is designed by combining the fault reconstruction value,and the original heterogeneous complex dynamic high-dimensional system is transformed into two lowdimensional closed-loop error systems by decoupling method.Finally,the controller gain solution method is obtained by Lyapunov and Ricatti theory,and the sufficient conditions for fault-tolerant consensus control of linear HMAS are obtained.For nonlinear HMAS,a nonlinear adaptive robust sliding mode observer is designed to reconstruct the actuator fault.Secondly,the fault reconstruction value is introduced into the distributed fault-tolerant control protocol,and the influence of actuator fault on the system is solved by compensation.Then,in order to solve the gain of the controller,the system is transformed into a low-dimensional system by decoupling principle,and the nonlinear system is linearized by robust principle.Finally,the controller gain is obtained by mathematical theory,and the sufficient conditions for fault-tolerant consensus control of nonlinear HMAS are obtained.3)For complex HMAS with actuator additive time-varying faults and external bounded energy disturbances.the problem of fault-tolerant consensus control is studied.Firstly,the first-order agent fault observer is designed and the second-order agent actuator fault observer is designed by using the observer design method of 1)to estimate the actuator fault.Secondly,a joint distributed fault-tolerant controller is designed by combining the fault estimation value.Then,in order to obtain the controller gain matrix,the original complex system is decomposed into two same-order subsystems by graph theory knowledge,and the two subsystems are augmented by the augmented matrix principle.Finally,sufficient conditions for fault-tolerant consensus control of complex HMAS are obtained.