考虑驾驶风格的网联自主车辆换道意图预测方法OA

The behavioral intention of Connected Autonomous Vehicles(CAVs)results from the interaction of multidimensional elements in complex traffic scenarios,and the prediction of lane changing intention is crucial for behavioral safety control and trajectory planning.This article proposes a lane change intention prediction method which not only considers the vehicle's motion trajectory but also its driving style.First-ly,vehicle trajectory data that meet specific criteria are extracted and classified using k-means clustering algorithm into conservative,general,and aggressive types based on driving style.Secondly,a Convolu-tional Long Short-Term Memory Network(ConvLSTM)model is established to predict vehicle lane change intention.The prediction is compared with trajectory data corresponding conservative,general,and aggressive styles for validation analysis.The results indicate that driving style has an impact on the accuracy of prediction.Compared with the prediction of lane changing intention without dividing driving style,the accuracy is improved by 5.17%,which can achieve accurate prediction of lane changing inten-tion.