Advances in sensor technology have made it possible to directly measure the torque of electric vehicle (EV) driveshafts. However, no previous studies have utilized torque sensors for torsional vibration suppression control of the driveshaft. This study is the first to propose the use of directly measured torque for vibration suppression control for EV powertrains. By eliminating the modeling difficulties associated with backlash and tire nonlinearities inherent in the speed-sensor-based approaches, the torque-sensor-based method achieves a significant simplification of the control structure. The influence of torque-sensor-specific drift is also eliminated by setting the steady state gain of the controller to zero. Moreover, the controller is designed to have no impact on the primary torque delivery and to function as an optional controller, thereby minimizing its influence on upper-level control systems and enhancing flexibility in control system design. The controller was validated through both vehicle simulations and experiments using a downscaled experimental setup. In downscaled experiments with a backlash of 0.02 deg., the root mean square error (RMSE) of the shaft torque was reduced by 29% compared with feedforward (FF) control alone. Simulations are conducted using a vehicle model that incorporates a suspension system and tire lift-off. Compared with FF control alone, the proposed controller reduces the RMSE of the shaft torque by 82% on a wavy road and by 80% on a periodic low-μ road. Using a novel method, the periodic low-μ condition is reproduced in the scaled-down experimental setup, where the RMSE of the shaft torque is reduced by 81% relative to FF control.
Ueno et al. (Thu,) studied this question.