Intelligent Vibration Control for Semiactive Suspension Systems Without Prior Knowledge of Dynamical Nonlinear Damper Behaviors Based on Improved Extreme Learning Machine

Abstract

Aiming at enhancing the vehicle comfort and handling performances, this article concerns with the development of the vibration control method for the semiactive suspension systems installed with electrohydraulic dampers. To reject the external disturbances induced by the irregular road profile, sliding mode was intensively investigated for vibration control owing to its robust feature of insensitivity to the parametric uncertainty and external disturbances. However, the unknown nonlinearity of damper behaviors leads to model mismatch to cause the high-frequency switching of sliding-mode controllers. Consequently, the severe chattering phenomenon produces. Although saturated function is available to alleviate the chattering problem of sliding-mode control, there is always a tradeoff problem between tracking accuracy and chattering suppression. To solve this problem, this study provides a new intelligent robust control method for simultaneous improvements of tracking accuracy and chattering suppression. Given the computational efficiency, an improved extreme learning machine (ELM) is proposed to intelligently approximate and compensate the unmodeled dynamics with unknown nonlinearity to restrain the chattering problem, where a new adaptive learning law is designed in the premise of Lyapunov stability. To validate the effectiveness and efficiency of the proposed ELM-based robust control, a quarter-car test rig was set up for the hardware-in-the-loop test. Experimental results show that the proposed controller outperforms the sliding-mode controller with saturated function in depressing the sprung mass acceleration and tire deflection, showing its significance in both control performance enhancement and chattering elimination.