Modelling adolescent pedestrian crossing decision at unmarked roadway

Abstract

To address adolescent pedestrian safety problems at unmarked roadway, there is a need to understand adolescent pedestrian crossing decision behavior at unmarked roadway. cloud model, which is an uncertainty conversion model between qualitative knowledge description and quantitative value expression, was used to deal with adolescent pedestrian's cognitive uncertainty in the crossing decisionmaking process. Then the decision table for adolescent pedestrian crossing at unmarked roadway was established. Attribute reduction based on discernibility matrix and value reduction based on induction in the rough set theory were applied to reduce the decision table and extract the decision rules of adolescent pedestrian crossing at unmarked roadway. After the reduction, the conditional attributes were the vehicle speed and the distance between pedestrian and vehicle. Nine crossing decision rules were obtained, including five certainty rules. Finally, the proposed method was compared with the existing method. The results show that the prediction accuracy and area under the receiver operating characteristic (rOc) curve for the proposed method are 91.4% and 0.941 respectively, so it is superior to the logistic regression model, and the simple and intuitive pedestrian crossing decision rules can be obtained, which can lay the foundation for traffic safety simulation.