Adaptive Finite-Time Neural Constrained Control for Nonlinear Active Suspension Systems Based on the Command Filter

Abstract

This article focuses on the adaptive finite-time command filter control problem for nonlinear quarter active suspension systems with actuator failure. In the previous designs of active suspension systems, a second-order system for the vehicle body movement is employed to achieve the controller design and a zero-dynamic analysis is needed, thus the designed controller only contains a part of system variables and there is a difficulty in selecting a proper Lyapunov function in the zero-dynamic analysis. In order to overcome the aforementioned problems, a novel active suspension system is shown based on the Butterworth low-pass filter. The neural networks (NNs) are used to identify the unknown functions of active suspension systems. Meanwhile, the command filter is proposed to handle the “explosion of complexity” problem caused by the adaptive backstepping technique. Then, the adaptive practical finite-time control scheme combined with the command filter and neural identify technique is constructed in a unified framework, which can handle the “singularity” problem. Through the practical finite-time stability analysis, it is easily to obtain that all signals of closed-loop systems are bounded in a finite time. Finally, simulation results for the nonlinear quarter active suspension systems are provided to verify the effectiveness of the NN control scheme.