Sideslip Angle Estimation of Uncertain Vehicle System Based on Robust Extended Kalman Filter

Abstract

Accurate estimation of the sideslip angle (SA) is crucial for active safety systems of vehicle. However, model-based estimation methods cannot avoid the impact of parameter uncertainty. To improve the accuracy of SA estimation under model parameter deviation, a robust extended Kalman filter (REKF) method that compensates for virtual noise to model error is proposed and validated through practical application. First, a three-degree-of-freedom vehicle dynamic model is established, and a linearized and discretized parameter error model is obtained. Then, a REKF based on virtual noise compensation is used to estimate the SA and the robustness of this method is demonstrated. Finally, simulation and real vehicle experiments validated the effectiveness of this method. The results show that compared to existing methods, the proposed method can achieve better results in the presence of model parameter deviation, with an improvement in estimation accuracy of at least 40%.