A Unified Event-Triggered Control Approach for Uncertain Pure-Feedback Systems With or Without State Constraints.

Abstract

Existing schemes for systems with state constraints require the bounds of the constraints for controller design and may result in conservativeness or even become invalid when they are applied to systems without such constraints. In this paper, we study the problem of event-triggered control for a class of uncertain nonlinear systems by considering the cases with or without state constraints in a unified manner. By introducing a new universal-constrained function and using certain transformation techniques, the original-constrained system is converted into an equivalent totally unconstrained one. Then, an event-triggered adaptive neural-network (NN) controller is designed to stabilize the unconstrained system and compensate for the control sampling errors caused by event-triggered transmission of control signals. Unlike some existing control schemes developed for systems with state constraints, which need to check whether each virtual control meets certain feasibility conditions at every design step, our proposed unified method enables such feasibility conditions to be relaxed. In addition, a suitable event-triggering rule is designed to determine when to transmit control signals. It is theoretically shown that the designed controller can achieve the desired tracking ability and reduce the communication burden from the controller to the actuator at the same time. Simulation verification also confirms the effectiveness of the proposed approach.