Experimental verification of active oscillation controller for vehicle drivetrain with backlash nonlinearity based on norm-limited SPSA

Abstract

To address vehicle drivetrain vibrations that cause discomfort and poor drivability, this study proposes a new active damping strategy with simple backlash compensation based on the simultaneous perturbation stochastic approximation (SPSA) with norm-limited update vector. First, an experimental device developed for a simplified drivetrain mechanism is demonstrated. A mechanism for reproducing both the contact mode and the backlash mode is included in the device. For the contact mode, a model-based [Formula: see text] controller is employed as the baseline damping strategy. Further, to mitigate the backlash effect, a simple algorithm based on mode-switching-based compensation is used with the [Formula: see text] controller. In particular, for the critical controller parameters, this article presents a systematic design approach to search for their optimal values. The key parameters, which are needed for the backlash and contact mode controllers, are simultaneously auto-tuned using norm-limited update vector-based SPSA, which ensures the stability in the iterative tuning. The novelty of this study is that both the backlash mode controller and the contact mode controller are simultaneously optimized by the improved version of SPSA, thus realizing a comprehensive auto-tuning design of an active drivetrain damping system. Finally, the active controller is experimentally verified using the actual test device. Comparative studies show that the proposed approach significantly reduces drivetrain vibrations and is robust against fluctuations in the backlash.