Decision model for pedestrian interacting with traffic at uncontrolled intersections

With the continuous advancement of automated vehicles (AVs), there is growing concern about how AVs interact with vulnerable road users, particularly pedestrians. As a result, this emerging concern has given rise to extensive research into computational pedestrian behaviour models. However, very limited existing approaches modelled pedestrian crossing behaviour from an anthropomorphic perspective. Therefore, this study proposed a decision model for pedestrians interacting with traffic at uncontrolled intersections based on a human perception theory. Meanwhile, for the first time, we specified a kind of traffic flow-induced effect where pedestrians optimise their decisions by comparing the perceived risk of different traffic gaps. Furthermore, based on a deconstructed road-crossing decision process, we modelled pedestrian decisions and their timing in detail. A dataset collected in a CAVE-based simulator was applied to calibrate and validate the model. The results indicated that the proposed model fitted the data well and reasonably predicted pedestrian crossing decisions across a range of traffic flow scenarios. The model provides insights into the understanding and model pedestrian-AV interactions.