Robust control and estimation of clutch-to-clutch shifts

Clutch-to-clutch shifts are ubiquitous in automatic transmissions, motivating the need for formal and robust methods for controlling these shifts. Limited sensing in production transmissions poses a severe hurdle for feedback control of these gearshifts. In the current study, nonlinear estimation methods are developed to compensate for limited sensing, and enable model-based closed loop control of the torque and inertia phases of shifts by manipulation of clutch pressures. During the torque phase, the offgoing clutch is controlled to emulate a one-way clutch, which ensures smooth coordination of the two clutches and reduced overall variation in the output shaft torque during the gearshift. During the inertia phase, the oncoming clutch is controlled to ensure smooth engagement at lock-up, resulting in reduction of shock and subsequent driveline oscillations. Controller performance is evaluated through numerical simulation of the proposed observer based controller on an experimentally validated high order model of a stepped production automatic transmission. The results show that shift control objectives were met by the proposed estimation and control strategy in the presence of appreciable model uncertainty and speed sensor noise, thus validating the robustness and practical effectiveness of the controller. Also, the proposed model-based controller was shown to be effective in controlling gearshifts at different power-levels (at different throttle openings), which establishes effectiveness of the same over a wide range of operating conditions.