A dynamic lane-changing decision and trajectory planning model of autonomous vehicles under mixed autonomous vehicle and human-driven vehicle environment

Abstract

How to complete a lane changing process considering various variables has always been a critical issue in the field of autonomous driving. Developing a lane-changing decision model with full consideration of the surrounding vehicles and related decision-based trajectory planning model that comprehensively weighs safety and efficiency are conducive to the driving of autonomous vehicles (AVs) under mixed autonomous vehicle and human-driven vehicle (AV–HV) environment. Under the mixed AV–HV environment, we optimize a multi-player dynamic game model considering the status of surrounding vehicles to ensure the accurate execution of lane-changing decision of AVs. Lane changing trajectory of AV is planned based on polynomial curves, which can be dynamically updated according to the real-time status of vehicles and game results. Then, a computational experiment basing on the lane changing vehicles data from NGSIM (Next Generation Simulation) is performed with proposed models. The simulation results show that the lane-changing decision and trajectory planning model developed in our research have good adaptability to lane changing process in different scenarios, which can effectively measure the driving intention of surrounding vehicles and dynamically plan a smooth trajectory line considering safety and efficiency.