Multi-sensor Data Fusion and Application in Longitudinal Velocity Estimation of Electric Vehicles ☆

vehicle which cannot be measured directly. Longitudinal velocity of the vehicle is one of the most important information which is needed for motion and autonomous control. To take full advantage of traction electric motor, the longitudinal velocity must be estimated at an appropriately fast frequency. In this paper, we propose an optimal estimation of velocity of the EV based on multi-sensor data fusion. The estimator does not need any dynamic parameters but only information from sensor system that is equipped on the vehicle. The performance of the estimation algorithm is validated by experiment on a real electric vehicle and compared with a commercial data acquisition system. The results of the paper can be used for futher control purpose as well as estimation of other state of electric vehicles.