On Adaptive Fault-tolerant Control For Multi-Agent Systems With Constraints

Multi-agent systems(MASs)are a distributed,autonomous system make up of a variety of intelligent agents.Studying MASs can not only improve human knowledge of natural events,but also provide a scientific basis for human implementation in industry.In real-world engineering applications,in order to guarantee the safety and performance standards of the system,it is typically necessary for the systems’ state,output,or transient performance to meet certain conditions.Moreover,affected by environmental and human factors,the likelihood of system failure rises,and the occurrence of failure will have an impact on the systems’performance and stability.In recent years,distributed adaptive fault-tolerant control of nonlinear MASs subject to restrictions and multiple uncertainties has become one of the hot issues in control theory research.This research is important from a theoretical point of view and could be useful in the real world.In this paper,several different control strategies for nonlinear MASs with constraints,actuator failures,sensor failures,and unmodeled dynamics are developed via command filter backstepping design.The main points of this paper are as follows:Firstly,a new distributed adaptive fault-tolerant control scheme for uncertain nonlinear MASs with full state time-varying constraints and actuator failures is proposed via command filtered backstepping design.By introducing two universally constrained functions and using certain mathematical transformation,the original constraint system is remade as a novel unrestricted system.Compared with the previous literature,the assumption that time-varying constraint functions need upper bounds is dropped.By employing Young’s inequality and the Gaussian function’s characteristics,the non-strict feedback term and MASs’ coupled issue are successfully treated.By constructing the auxiliary parameters and the adjustment laws,the influence of actuator failure is mitigated.By using the mean value theorem,the non-affine nonlinear functions including multiple input single output(MISO)fault models are tackled.Through Lyapunov stability analysis,it is demonstrated that all signals in the controlled system are semi-globally uniformly ultimately bounded(SGUUB)and all followers are capable of tracking the desired trajectory.Finally,the simulation result shows that the control scheme is effective.Secondly,a novel distributed adaptive event triggered fault-tolerant control manner for uncertain nonlinear MASs with time-varying output constraint,unmodeled dynamics and actuator failures is raised with command filtered backstepping design.By introducing two universally constrained functions and using certain mathematical transformation,the output constraint is effectively handled.Meanwhile,the bounds of synchronization errors are also guaranteed.The dynamic signals that are observable and generated by the first-order auxiliary system are used to mitigate the influence of unmodeled dynamics on the system.By utilizing radial basis function neural networks(RBFNNs)to approximate unknown functions.By introducing smoothing functions,the influences of event triggering errors and actuator faults are compensated.Through Lyapunov stability analysis,it is proved that all signals in the controlled system are SGUUB and all followers’ output can track the desired trajectory within the constraints.Finally,the simulation example also verify that the control scheme is valid.Thirdly,an adaptive fault-tolerant control strategy for uncertain nonlinear MASs with input unmodeled dynamics,prescribed performance,sensor failures and unknown control directions is developed via command filtered backstepping design.With the help of the property of Gaussian function,the barriers produced by the non-strict feedback terms and partial sensor faults are handled.Moreover,the prescribed performance on measured synchronization error is considered by designing a finite-time performance function.By introducing normalization signals and adjusting parameters,the problems caused by input unmodeled dynamics are resolved.The compensation signals and normalization signals are integrated into the whole Lyapunov function in the stability analysis.All the variables in the controlled system are turned out to be SGUUB.Meanwhile,the synchronization errors never exceed to predefined boundary.Finally,the simulation experiment also proves that the control scheme is available.