Development and Application of a Simulation-Enhanced Platform for Pedestrian Crossing Behaviors Experiment

The issue of disorder, high risks and low efficiency at signalized intersections is mainly attributed to illegal pedestrian crossing. Faced with the challenge of complicated psycho-physical factors, virtual experimental method using microscopic traffic simulation tools shows its advantage over theoretical analysis and questionnaires in studying pedestrian crossing behaviors. In this paper, the Pedestrian Crossing Behavior Virtual Reality System (PCBVRS) was developed based on the VISSIM and Virtual Reality (VR) technology. Through the communication between master and two slaves, the VISSIM simulation was synchronized and the left, ahead, right view of pedestrians was captured respectively in three computers. The participant can watch the simulation scenes and make the crossing decisions through the human-computer interface, meanwhile an avatar crossing the virtual street and corresponding information were recorded in PCBVRS. By comparing the experimental data with the field data using the concept of reliability and validity, the pedestrian crossing behaviors could be better reproduced in PCBVRS. Finally, 59 pedestrians were recruited and 449 crossing samples were gotten to analysis 5 pedestrian red countdown signals using PCBVRS experiments. According to the pedestrian average waiting time and compliance rate, the red countdown was recommended to start within the range of 45 to 60 seconds. Compared to traditional VR systems, PCBVRS is low-cost and high-portability and will contribute substantially to the study of pedestrian crossing behaviors.