Roadside pedestrian motion prediction using Bayesian methods and particle filter

Accidents between vehicles and pedestrians account for a large partition of severe traffic accidents. So, pedestrian motion prediction becomes major concern intelligent vehicles. However, current researches often neglect behaviour and/or intention in prediction. Meanwhile, related works are scattered divided into many small fields. No integrated system is proposed to connect the task perception decision. To solve these problems, model this paper. The method predicts based on combination crossing intention. Pedestrian recognized using Bayesian posterior model, by dynamic network. A modified particle filter behavioural used integrate effectiveness verified our provided BPI dataset with eight typical scenarios defined road type, vehicle velocity etc. results show that can give an accurate distribution pedestrians’ future trajectories.