Vehicle speed estimation in driving mode for hybrid electric car using unscented Kalman filter

Abstract

As for the four-wheel drive hybrid electric car with complex and changeable driving modes, unscented Kalman filter (UKF) is proposed for vehicle speed estimation in consideration of obtainable driving torque and electronic stability program (ESP) senor signals. Simulation platform is established according to the four-wheel drive hybrid electric car, which integrates power-train system model, nonlinear seven degree of freedom vehicle dynamics model and the dynamic union tire model. The estimation result of UKF algorithm is compared with the simulated real car's velocity. After steering wheel angle sensor, yaw angular velocity sensor and acceleration sensors are all mounted in the prototype car and the signals of two front wheel angular speed are acquired as well as the torque information of driving wheel are introduced, UKF algorithm is tested on the real vehicle road experiments, which include 8-shape route driving case on pure electric drive mode, double-lane change driving case and S-shape route driving case on four wheel hybrid drive mode. Simulation and test results show that the proposed algorithm has not only high precision, but also strong adaptability.