Longitudinal Speed Estimation of Multi-axle Distributed Drive Vehicle Based on Federal Kalman Filter

When the wheels of distributed drive vehicle slip, the effectiveness of longitudinal speed estimation based on dynamics model is poor. Therefore, a multi-model fusion estimation method based on Federal Kalman Filter (FKF) is proposed in this paper. In order to tackle the nonlinear problem of vehicle dynamics model, an Extended Kalman Filter (EKF) is used to estimate the longitudinal speed. Additionally, considering the large mismatch between the dynamics model and the real vehicle when the wheels slip, the acceleration and linear Kalman Filter (KF) framework are used to construct a speed estimation algorithm based on kinematics. Then a FKF is designed to fuse both dynamics and kinematics methods. The slip ratio observer is introduced as the information distribution principle of two methods, and information fusion and time update are completed in FKF. After running for a period of time, the fusion result is used to reset the filter of the kinematics method to reduce the accumulated error. Finally, Trucksim/Simulink co-simulation platform is utilized for simulation analysis. Results show that the proposed speed estimation method has good effect and adaptability in the slip condition.