Simultaneous Perturbation Algorithm Tuned Oscillation Controller for Automotive Driveline With Kalman Filtering-Based Compensation for Dead-Zone Nonlinearity

Abstract Advanced active control algorithms have addressed the reduction of vehicle driveline oscillations with the purpose of improving component lifespan and drivability. However, the majority of the current works require the important controller parameters to be determined subjectively, placing a significant burden on designers to make adjustments. This study offers an effective model-based driveline vibration controller tuning technique that specifically takes nonlinear backlash’s negative effects into account. Initially, a driveline model is created that incorporates a dead-zone impact of backlash. To reduce a driveline’s low-frequency resonance, a baseline controller reflecting H2 control theory is used. To handle the backlash nonlinearity, the controller is coupled with a control mode switching-based compensation. For identifying whether the system is in the backlash mode or the contact mode, the state vector is estimated by a Kalman filter. The Kalman filter-based state estimation enables switching of the two control modes systematically and stably. A computationally efficient approach, namely the simultaneous perturbation stochastic approximation (SPSA), finds out the best values of their significant design parameters that are incorporated in the active control system. Multiple comparative simulations confirm the effectiveness of the proposed active oscillation controller adjusted by the SPSA. The resilience is assessed for several patterns of driveline dynamics variations. Consequently, it is discovered that the improved vibration suppression performance is stemming from the backlash compensation.