MIXED TRAFFIC HEADWAY MODELLING ON URBAN ROADS USING NEURAL NETWORK

Abstract

Traffic flow is a complex phenomenon involving several parameters, one of them is headway. Headway distributions are key building blocks for microscopic traffic flow characteristics which involves the safety, level of services, driver's behaviour and capacity of transportation system. Headway models help to understand the arrival pattern, driver's behaviour and safety on roads and intersections. Headway modelling by conventional methods may not be suitable in all the situation due to some limitations: Therefore digital simulation technique (Artificial Neural Network) may be used which can prove to be a better modelling techniques. In the present study, data collected at one section of urban roads in Delhi has been used to predict headway at different conditions of traffic using Artificial Neural Network. Neural planner 4.1 is used to predict headway for defined set of problem. The effect of traffic composition and traffic volume on headway between two vehicles has been investigated. The capacity of the road section is estimated as 2092 PCU/hr for a 100% car situation.