Accurate State Estimation for Electro-Mechanical Brake Systems

Electro-mechanical brake (EMB) system is an electric motor based braking force generation module, and it requires various sensors such as motor position, motor current and clamping force sensor for stable vehicle deceleration control. Because fault in these sensors can lead to degradation of the system performance, system monitoring is essential. To build a model based state estimator for the braking system, there are some requirements: the mathematical model presenting the nonlinearity and disturbance of the real system, fast response time and the accurate estimation. To solve this problem, this paper proposes a new EMB model which clamping force term is divided into the linear and nonlinear compensation part, and Kalman filter algorithm is applied to design the state estimator. The proposed model is simple and linear, and Kalman filter algorithm is robust to system noise and guarantees the fast computation time. Additionally, the braking direction aware and contact point aware clamping force estimation techniques are introduced, and they help to improve the accuracy of the state estimation. Lastly, the proposed approach is verified through experiments on the EMB test bench.