Vehicle speed estimation in driving case based on distributed self-adaptive unscented Kalman filter for 4WD hybrid electric car

Abstract

As for the four-wheel drive hybrid electric vehicle (4WD HEV), in order to improve the accuracy and robustness of speed estimation for electronic stability program (ESP), vehicle speed estimation based on distributed self-adaptive unscented Kalman filter (UKF) is proposed in consideration of obtained driving torque and ESP sensor signals. Firstly, powertrain and kinetic model are established according to the 4WD HEV, which includes power-train system model, seven degrees of freedom vehicle dynamics model and Burckhardt tire model. Secondly, considering the model is time-varying and strongly nonlinear, UKF algorithm is adopted to design the main/sub filter. On the one hand, self-adaptive UKF is designed for measurement noise to improve its robustness; on the other hand, main/sub filter results are fused and then the fused result is used to reset each filter, which improves the accuracy of speed estimation. Finally, off-line co-simulation of Carsim-Simulink and hardware-in-the-loop (HIL) platform for 4WD HEV are built, distributed self-adaptive UKF algorithm is tested in 8-shape route driving case and low speed double-line change driving case. The results show that the proposed speed estimation algorithm based on distributed self-adaptive UKF not only has high accuracy, but also has strong adaptability and robustness.