Modelling the Crossing Behavior of Pedestrian at Uncontrolled Intersection in Case of Mixed Traffic Using Adaptive Neuro Fuzzy Inference System

Abstract

 Abstract-Critical gap is the most important parameter associated with gap acceptance study especially in case of delay and capacity estimation. Many studies are reported on gap acceptance behavior of pedestrian but most of them confined to developed country where traffic is homogeneous and traffic rules are strictly followed. Uncontrolled intersections in case of developed countries control the traffic and pedestrian movements based on priorities but in India, no one follows the stop and yield signs, even the rules of zebra crossings. It creates more conflicts and increases delay to both. This paper systematically analyzes the behavior of pedestrian during the gap acceptation at two four legged TWSC intersection located at Ahmedabad in state of Gujarat, India. Data on gap acceptance behavior is obtained by video recording technique and analyzed the various parameters relate to crossing behavior of pedestrian. The critical gaps for pedestrians estimated by raff method and results shows that values of critical gap estimated are as low as 3.20 sec. which is smaller than reported in other studies. An Adaptive Neuro Fuzzy Inference System (ANFIS) has been established to estimate the possibilities of accepting a given gap or lag considering the combinations of different parameters which affects the crossing behaviors of pedestrians. Nine different combinations are considered. However, the age of the pedestrian is found to be most important variable in crossing behavior compared to other.

Social-Aware Driver Assistance Systems for City Traffic in Shared Spaces

“…In particular, fuzzy approaches have shown their utility in dealing with qualitative reasoning on them. For instance, on modelling the crossing behavior of pedestrians from their critical gap [49] or of cyclists from their social force [50]. Nevertheless, and as with many mathematical models, these approaches have focused on specific decisions that depend on functions and thresholds over certain inputs.…”