Review of Vehicle Behavior Prediction for Intelligent Driving Systems in Highway Scenarios

Abstract

In recent years, particularly since 2015, autonomous driving systems have developed rapidly. With the increase of autonomous driving levels, more precise detection and environment cognition are necessary. In this paper, a systematic review on vehicle behavior prediction in highway scenarios is carried out, including the role of vehicle behavior prediction in autonomous driving system, the application scenarios and the mechanisms of widely used vehicle behavior prediction methods such as the Bayesian network (BN) based method and the long short-term memory (LSTM) network-based method. The future directions of vehicle behavior prediction are discussed.