Experimental validation of adaptive grey wolf optimizer-based powertrain vibration control with backlash handling

Abstract

The controller optimization task is rarely spotlighted despite its importance for vehicle drivetrain mechanisms although many studies have been dedicated to developing vibration control strategies. Based on the adaptive grey wolf optimizer (AGWO), this research develops a fast-optimization scheme for a drivetrain oscillation control system that simultaneously addresses effects of nonlinear backlash. A drivetrain system model governed by a backlash nonlinearity is presented, and a baseline controller is derived for damping low-frequency drivetrain resonance based on the optimal $H_2$ synthesis. The introduction of a time-dependent-switched Kalman filter realizes a solution for dealing with the nonlinear backlash issue, relying on straightforward controller-switching-based compensation for the backlash and contact modes. Optimal solutions for the control system parameters are efficiently obtained using AGWO. AGWO exhibits both global search capability and superior computational efficiency because of its systematic stopping criteria and adaptive exploration/exploitation parameter. This study improves the efficiency of optimizing active drivetrain vibration control by introducing the adaptive mechanism into the controller parameter tuning. Comparative experiments demonstrate that the AGWO-based scheme provides a sufficiently good controller with the fastest time.