Neural network-based adaptive finite-time tracking control of switched nonlinear systems with time-varying delay

Abstract

This work investigates the adaptive finite-time tracking control problem for switched nonlinear systems, in which backlash-like hysteresis and time-varying delay are taken into account. The nonlinear estimation ability of radial basis function neural networks is employed to relax the restriction on unknown nonlinear functions. The dynamic surface technology and the finite-time control approach avoid the “curse of dimensionality” and “singularity” problems existing in the backstepping design procedure, separately. By Lyapunov–Razumikhin function scheme, the proposed finite-time signal guarantees superior tracking performance under the average dwell time switching. Finally, to testify the practicability of the presented strategy, two simulation examples are shown.