Segmented Estimation of Road Adhesion Coefficient Based on Multimodal Vehicle Dynamics Fusion in a Large Steering Angle Range

Real-time estimation of the road surface friction coefficient is crucial for vehicle dynamics control. Under large steering angles, the accuracy of existing road surface friction coefficient estimation methods is unsatisfactory due to the nonlinear characteristics of the tire. This paper proposes a segmented estimation method for the road adhesion coefficient, which considers different steering angle ranges and utilizes multimodal vehicle dynamics fusion. The method is designed to accurately estimate the road adhesion coefficient across the full steering angle range of the steer-by-wire system. When the front wheel angle is small (less than 2.8°), an improved Unscented Kalman Filter (AUKF) algorithm is used to estimate the road surface friction coefficient. When the front wheel angle is large (greater than 3.2°), a rack force expansion state observer is constructed using the dynamics model of the steer-by-wire actuator to estimate the rack force. Based on the principle that the rack force varies with different road surface friction coefficients for the same steering angle, the rack force is used to distinguish the road surface friction coefficient. When the front wheel angle is between the two ranges, the average value of both methods is taken as the final estimate. The method is verified through Matlab/Simulink and CarSim co-simulation, as well as hardware-in-the-loop experiments of the steer-by-wire system. Simulation results show that the relative error of road surface friction coefficient estimation is less than 10% under different steering angles. The segmented combination estimation strategy proposed in this paper reduces the impact of tire nonlinearities on the estimation result and achieves high-precision road surface friction coefficient estimation over the entire steering angle range of the steer-by-wire system, which is of significant importance for vehicle dynamics control.