Estimation Method of Vehicle Centroid Sideslip Angle Based on DNN-EKF Fusion

Abstract

In order to solve the problem that vehicle sideslip angle is difficult to be measured online in the process of vehicle stability control, a vehicle sideslip angle estimation method based on DNN-EKF algorithm fusion is proposed. This method combines DNN algorithm and EKF algorithm to estimate sideslip angle of centroid. This method adopts DNN network to calculate the error between actual value and estimated value, then put error compensation for the status of the EKF, thus compensate and correct the estimated value of EKF. The experimental results show that the accuracy of the estimation method is 70% and 80% higher than DNN algorithm and EKF algorithm respectively, so the algorithm has good robustness.