Discrimination And Control Method Research Of Intelligent Vehicle Lateral Stability Based On SOFM And K-Means

Path tracking is one of the basic technologies for the development of intelligent vehicles.The driving environment of intelligent vehicles is complex and changeable in the process of tracking the path that poses a great challenge to the lateral stability of the intelligent vehicle.In this paper,the lateral stability of vehicle during track tracking has been studied,the main contents are as follows:Firstly,the database of cluster analysis is established.Deducing the ideal value of yaw rate in order to tracking aim path according to the optimal preview theory and the hypothesis of steady state circular motion.The active steering control based on the fuzzy PID theory is established to track the ideal yaw rate,due to there are always errors between ideal yaw rate and actual yaw rate.The parameters related to vehicle lateral stability are analyzed,and the clustering analysis database is built by setting up double lane change test condition for simulation and extracting simulation data.Secondly,a new method of judging the vehicle lateral stability is designed based on the SOFM neural network and K-means clustering analysis.Three offline clustering centers are obtained by offline clustering analysis of vehicle driving parameters through the combination of SOFM neural network and K-means clustering.According to the characteristics of vehicle driving condition,the results of clustering are analyzed to determine the instability level represented of three offline clustering centers.In order to improve the robustness of the combined clustering stability discrimination method,the mean value method was used to adjust the corresponding clustering center based on the offline clustering center to complete the online discrimination of vehicle lateral stability.Taking the tire force method as discriminant benchmark,the performance of the stability discrimination method based on SOFM and K-Means clustering is analyzed through Simulink/Car Sim joint simulation,and the effectiveness of the stability discrimination method is verified.Thirdly,a coordinated control strategy of lateral stability is designed based on the discrimination results.The direct yaw moment control is designed based on the fuzzy PID theory in order to avoid the problem that the stability deterioration in some limit cases of intelligent vehicles,which together with the active steering control constitutes the overall stability control strategy.The timing of intervention control and the control weight of stability control strategy are determined according to the results of stability discrimination to realize the coordinated control of vehicle lateral stability under different conditions.The effectiveness of the stability control strategy and the rationality of the stability discrimination result as the basis of the control strategy intervention control are verified through double lane change test condition and fishhook test condition simulation.Finally,based on the driving simulator platform,the rapid prototyping test is carried out.The stability discrimination method and stability control strategy are verified under the double lane change test condition.The experimental results further verified that the proposed lateral stability discrimination method can real time quantify vehicle stability and accurately guide the stability control strategy intervention control.While improving the tracking accuracy,it also ensures the lateral stability of the intelligent vehicle.