Sensorless optimal control of electronic wedge brake based on dynamic model and Kalman filter state multiple-estimation

Abstract

An electronic wedge brake is a brake-by-wire system developed by some major companies in the domain of the automotive industry. The inherent self-reinforcing feature of the electronic wedge brake is interpreted as reduced actuation force and hence lower actuation energy. Controlling the electronic wedge brake requires measuring the clamping force using a force sensor, which requires special techniques to be installed onto the system and maintained safe from extreme conditions. This way the system design will be more complex, and the cost is increased due to the sensor expense. In this paper, a new sensorless optimal control (Servo-Linear Quadrature Regulator) approach is proposed using system dynamic-based estimation and Kalman filter-based estimation for clamp force control of electronic wedge brake. The suggested control scheme eliminates the need for clamp force and motor speed sensors by state multiple-estimation using only the current measurement of the motor. First, the motor current is measured and used to estimate motor speed relying on system dynamic equations. Brake wedge speed and position are then approximately derived in terms of the motor state estimates. The resulting current-based state estimates are eventually introduced to the Kalman filter as state measurements and used with the system control input calculated by the optimal controller to correct and update state prediction in the Kalman filter.