부산대학교 도서관

For enhancement of the drivability and lifetime of the components, reduction in vehicle driveline oscillations has been addressed by advanced active control algorithms. However, most of the existing works involve subjective determination of their critical controller parameters, imposing heavy adjustment tasks on designers. This research presents an efficient tuning algorithm of the model-based driveline vibration controller that explicitly considers the adverse influences due to nonlinear backlash. First, a driveline dynamics model with a dead-zone effect of backlash is established. A dynamic output feedback $H_{2}$ controller is designed as a baseline controller to attenuate the low-frequency resonance of a driveline. A simple control mode switching algorithm is combined with the controller to deal with the backlash nonlinearity. The optimal values of their important design parameters included in the active control system are automatically searched by a computationally efficient algorithm, i.e., the simultaneous perturbation stochastic approximation (SPSA). The proposed active oscillation controller tuned by the SPSA is validated via several simulation tests. The robustness is evaluated for various patterns of driveline dynamics fluctuations such as the model parameters, the length of backlash, and the driving conditions. Moreover, the proposed controller is compared to traditional active controllers including a proportional-integral-differential (PID) controller tuned by the Ziegler-Nichols method. As a result, the improvement of the vibration suppression performance as well as its robustness originating from the compensation for backlash is revealed.