Automatic Clutch Control Based on Estimation of Resistance Torque for AMT

Abstract

Automatic clutch control is critical in automated manual transmission systems because clutch actions have a significant influence on the vehicle performance, including safety, comfort, reliability, and shifting quality. However, clutch control is a challenging problem due to nonlinearity of throw-out force and uncertainty in clutch wear. For precise position control of the clutch, this study proposes a control scheme using model predictive control method with the correction of clutch wear based on the estimation of resistance torque. First, a novel and detailed clutch model is introduced, mainly including a driven plate model and a diaphragm spring model. Next, through theoretical analysis of the clutch model and the experimental data of a real clutch, the characteristics of the nonlinear throw-out force are obtained with different clutch wear situations. Then, a reliable clutch actuator is designed, whose model is also built to analyze its dynamics. At last, an automatic clutch controller is designed, and in order to apply it in clutch control unit for real-time control, it is simplified to reduce the online computing burden. The results both of simulations and bench experiments show that the proposed control scheme has a satisfying control performance.