Longitudinal Velocity and Road Slope Estimation in Hybrid/Electric Vehicles - Development and evaluation of an adaptive Kalman filter

Abstract

An accurate and efficient method of velocity and slope estimation is presented in this thesis. An adaptive Kalman filter is proposed to deal with the over-slip wheel, which is a challenging problem during velocity estimation. The research object of this work is the hybrid/electric vehicles with electric motors. In the adaptive Kalman filter, we control the gain matrix directly based on the over-slip flag. If all of the four wheels over-slip at the same time, the velocity results are replaced by the integration of acceleration. To reduce the integration error, the longitudinal acceleration is modified by road slope estimation results. The slope estimation is based on typical Kalman filter and the observation variable is velocity estimation result. Then, the over-slip criterion and wheel speed selection method are involved aiming to estimate velocity accurately when all the four wheels are over-slip. Besides the wheel speed and pre-estimation of velocity, the wheel torque, provided by electric motor, is also used to find out the over-slip wheels. Nevertheless, some abnormal measurements that cannot be detected by the over-slip criteria affect the accuracy of the estimation. Thus wheel speed selection is put forward to reduce the influence of measurements error as well as the calculation quantity. After selection, only one wheel speed is selected as the observation variable of Kalman filter. At last, the algorithm is verified on both high and low friction road