Optimal Clutch Torque Prediction for Shifting Process of Dual Clutch Transmission Based on Support Vector Regression

摘要

To overcome the difficulty of real-time optimization during shifting process for dual clutch transmission, a clutch optimal torque prediction method based on Support Vector Regression is proposed. Firstly, a shifting dynamic model of dual clutch transmission system is established. Afterwards, the maximum jerk, friction work and shifting time are weighted and summed as an objective function for the optimization problem, weighting factors of which are determined by driving intention. Meanwhile, the clutch torque is formulated by a Fourier series, coefficients of which during shifting process are optimized by Genetic Algorithm. Subsequently, the data-driven controller is trained by Support Vector Regression to predict the optimal clutch torque in real time during shifting process. Finally, the prediction accuracy of the Support Vector Regression method is verified by simulation and experiment. The results show that the Support Vector Regression algorithm has high accuracy in predicting the optimal clutch torque during shifting process. Consequently, the online optimal control can be realized with the assistance of the optimal torque prediction.