Improved Intersection Crossing Behavior with Active Perception Approach

Nowadays, micro-simulation is a common approach to study the behaviors of drivers in the road traffic. The main concern of most microscopic simulators is the network efficiency evaluation. The micro-simulation approach relies on models such as car following, lane changing and intersection crossing. Each of these models has a theoretical basis, and corresponds to a specific road section and a specific driver intention. In addition to traffic flow, the micro-simulation approach would be an interesting tool to investigate an accident or near accident situation. Some approaches tackle the individual behavior in these micro-simulations. For these approaches, a more detailed behavioral model, such as the nanoscopic simulation, is required. In this paper, the authors focus on the intersection crossing behavior of drivers. Although various researches have already addressed this subject, existing approaches seem inadequate to simulate accurately drivers' behavior in the conflict area (the center of intersection) or in the crossroads exit. The authors are developing an active perception model following a nanoscopic approach. The aim of this paper is to make a qualitative comparison between their approach and the existing gap acceptance model. Their model allows a simulation of the interaction between drivers at the center of intersection. The authors have implemented their model in an unsignalized intersection. They have explored the impacts of the perception limit parameter σ comparing the mean of the traffic performance (flow), the number of accidents, the run time and the number of deadlocks in the intersection on 100 simulations of one hour traffic.