Estimation of Tire-Road Friction Coefficient Under Non-Zero Mean Gaussian Noise Interference

Accurate acquisition of tire-road friction coefficient(TRFC) is crucial for vehicle safety and stability. However, current estimation methods heavily rely on the precision of sensor data and are unable to address the high- accuracy estimation problem when faced with non-zero mean measurement noise interference, besides, they always ignore the impact of model mutation on estimation accuracy. This paper presents a seven-degree-of-freedom vehicle dynamics model, incorporating a parameter-adaptive Brush tire model to enhance model accuracy. Moreover, an Adaptive Unscented Kalman Filter (AUKF) is proposed to estimate the mean of the measurement noise in real-time. The accuracy and robustness are verified through simulation tests on different road conditions. Results show that AUKF outperforms UKF in both accuracy and robustness, confirming the feasibility of AUKF.