Event triggered NN-embedding compensation fault tolerant control for high-speed trains with actuator saturation

Abstract

In this paper, an event-triggered neural network (NN) embedding compensation control scheme for a high-speed train (HST) with unknown dynamics, unknown disturbances, actuator faults, and asymmetric nonlinear actuator saturation (ANAS) is investigated. The adaptive PID sliding mode fault-tolerant control (PID-SMFTC) with mix basis function approximation (MBF) and finite-time nonlinear disturbance observer (FTNDOB) is proposed as base controller of event triggered NN-embedding compensation control. The MBF is employed to approximate the unknown dynamics term in HST system and eliminate the effect of ANAS. The FTNDOB is used to estimate unknown disturbances within a finite time. Then, NN-embedding compensation control scheme with two event triggered threshold strategies are proposed to optimize the performance of the base controller. Comparing with NN-embedding compensation controller, these methods reduce the consumption of communication and computation resources by optimizing the base controller’s performance only when events are triggered. Finally, simulation results using an actual train dynamic model are showcased to validate the effectiveness and feasibility of the proposed schemes.