Road Adhesion Coefficient Estimation Method Based on Image and Vehicle Dynamics Online Fusion

The road adhesion coefficient is a critical parameter in active safety control systems. However, achieving both accuracy and real-time performance in estimating this coefficient for nonlinear vehicle-tire systems remains a significant challenge. To address this issue, this paper proposes an online fusion method for estimating the road adhesion coefficient that integrates image-based and vehicle dynamic data. First, a threedegree- of-freedom vehicle dynamic model is employed, and correction factors are introduced into the Dugoff tire model to enable adaptive cornering stiffness calculation. Next, to leverage visual information effectively, the range of the road adhesion coefficients obtained from a residual network is incorporated as the upper and lower bounds of constraints in the moving horizon estimation method. Additionally, a detailed stability proof for the moving horizon estimation method is provided. Finally, the effectiveness of the proposed estimation method is validated through both software simulation and vehicle tests. The results demonstrate that the proposed estimation method achieves high estimation accuracy and rapid convergence.