Research status and trend of trajectory prediction for intelligent connected vehicles
As a key technology on intelligent connected vehicles,trajectory prediction plays an important role in ensuring vehicle safety and preventing traffic accidents.Based on the existing trajectory prediction methods,the mathematical model,data set and evaluation index of trajectory prediction are presented in this paper.Then,the current research of trajectory prediction is introduced,and the data-driven trajectory prediction methods such as deep learning and reinforcement learning are reviewed.Meanwhile,the network model structure,advantages and applicable scenarios of existing research methods are compared and analyzed with the view of space-time modeling,interactive information and scene context.Finally,the research trend of multi-agent and multi-modal trajectory prediction and the future trend of trajectory prediction combined with large language model are further discussed.
automatic drivingtrajectory predictionagentdata setevaluation index
国家科技期刊平台
NSTL
万方数据
