Practical Pinch Detection Algorithm for Smart Automotive Power Window Control Systems

An improved pinch detection algorithm is proposed for low-cost antipinch window control systems. Apart from previous works, the proposed algorithm makes use of torque rate information to sense pinched conditions and to perform safety precautions. The motivation for this approach comes from the idea that the torque rate is less sensitive to motor parameter uncertainty than the torque or the angular velocity. The pinch estimator is designed by applying steady-state Kalman filter recursion to the augmented system model which includes the motor dynamics model and an additional torque rate state. The external torque rate is estimated using angular velocity measurements calculated from the Hall sensor output. A systematic way to set a reasonable threshold of the torque rate estimates under pinched conditions is suggested through deterministic estimation error analysis. Therefore, the proposed algorithm is able to prevent performance degradation due to the empirical threshold level as well as due to motor parameter variations. Experimental results show that our method satisfies EU legal requirements and guarantees robustness against parametric uncertainties.