A differentiated decision‐making algorithm for automated vehicles based on pedestrian feature estimation

One critical difficulty to high-level automated driving is the decision-making process of vehicles in complicated traffic environments, especially situations mixed pedestrians and vehicles. This paper proposes a differentiated algorithm promote passing capability efficiency conditions. First, behavioural characteristic pedestrians, denoted as pedestrian feature index, estimated by multi-layer perception module input with quantitative analysis action. Based on estimation results, decision merges index into intelligent driver model adjusts corresponding parameters, which used be unchangeable so that ego-vehicle can make differential decisions according various features. Validation PIE dataset shows accuracy ensured. A simulation scenario established utilizing cellular automata, results indicate proposed greatly improve under safety manoeuvrability prerequisite.