Event-triggered Adaptive Prescribed Performance Fault-tolerant Control

The cyber-physical system(CPS)plays an essential role in modern industrial applications,which has been one of the focuses in academic research in recent years.CPSs meet plenty of practical requirements such as distributed control,high reliability,low cost,increased flexibility,easy configurability and maintenance.However,the networks in CPSs inevitably introduce various network-induced constraints,overload of which may degrade the system performance.These constraints include limited communication band-width,restricted computation and energy resources,which brings challenges for controller design.This gives rise to event-triggered mechanism that conserves the communication resources while ensuring the desired system performance.Besides,adaptive control,fault-tolerant control and prescribed performance control have undergone great development and improvement in both academic research and practical applications.Prescribed performance control can directly predefine and ensure the transient performance of the system.However,there are few works investigating the adaptive event-triggered tracking control problem for unknown nonlinear systems with prescribed performance.Fault-tolerant control and unknown control directions also need to be further considered.Additionally,due to the existence of uncertainties,most existing works only guarantee that the tracking errors converge to a bounded set,and cannot ensure the asymptotic convergence.Thus,for adaptive event-triggered tracking controller,how to guarantee prescribed performance,asymptotic convergence and finite-time convergence remains an open problem.This paper studies the adaptive event-triggered control problem for a class of block-triangular MIMO strict-feedback unknown nonlinear systems.The goal is to ensure the transient performance of the system,while decreasing the communication and calculation burdens.Various factors including fault-tolerant control,prescribed performance control,finite-time control,unknown control directions are considered.The controller is of high robustness,low structure and calculation complexity.It is approximation-free without using fuzzy logic systems or neural networks to deal with unknown nonlinearities,and it avoids the explosion of complexity.Besides,the cost of computational resources is at low level as the analytic calculation in common backstepping-based works is not needed.Combined with the event-triggered mechanism,the controller is suitable for CPSs which have limited communication and computation resources.Besides,some assumptions such as boundedness or globally Lipschitz used in the related literatures are relaxed.This paper investigates several simulations and experiments to evaluate the effectiveness of the proposed controllers.