Probabilistic intention prediction and trajectory generation based on dynamic bayesian networks
Ganglei He, Xin Yan Li, Ying Da Lv, Bingzhao Gao, Hong Chen · 2019
Lane keeping and lane change are the most general behaviors in highway driving. It’s crucial to predict the intention and trajectory of surrounding vehicles accurately as it can help intelligent vehicles make safe decisions and have a better understanding of the environment, thus improving the safety of autonomous driving and promoting cooperative driving. This paper focuses on the intention and trajectory prediction in expressway scenarios. A dynamic bayesian network is modeled, based on which the probability of each driving maneuver is inferred. Then, founded on maneuver estimation, lane change trajectory is roughly predicted and generated from the lane change starting point. This method is verified on real available public traffic data and the result demonstrates that this method has a good performance.