Prediction of lane change trajectory of autonomous vehicles based on vehicle-environment data fusion

Abstract

Abstract. Vehicle lane changing is a current research hotspot in the field of autonomous driving. However, the current research does not consider the environmental data of a large number of vehicles as well as the lane change decision risk and lane change trajectory prediction as a system, and thus lacks the consideration of safety and trajectory accuracy in the lane change process. Based on this, a fusion model is established by two algorithms in this paper. First, the vehicle lane-changing behavior is analyzed, and the NGSIM trajectory dataset is used as the study data, and the data is processed using the SEMA method. Then, the XGBoost algorithm is used to build a vehicle lane change risk detection model. Finally, the LSTM algorithm is used to predict the lane change trajectories of vehicles. After experimental simulation, the designed model effectively improves the safety and accuracy of self-driving vehicles in the process of lane changing.