An Optimal Traffic Stream Model for Indian Traffic Conditions

Traffic stream models provide relationships among the three basic macroscopic traffic stream variables namely traffic speed, volume and density. Traffic stream models are the basic building blocks of traffic flow modeling, design of road systems and traffic management systems. A number of research papers are available proposing different traffic stream models under homogeneous traffic conditions. These models are based on homogeneous traffic conditions and may not be directly suitable for the Indian traffic condition, which is heterogeneous and less lane disciplined in nature. Not much research has been done on this important aspect so far in India. The present study is an attempt to develop a traffic stream model suitable for the heterogeneous traffic flow condition, such as the one existing in India, taking Chennai as a case study. Traffic data for the present study were collected using videographic technique. Flow, speed and occupancy were manually extracted from collected videos and density estimated using the relation connecting occupancy and density. All the traffic variables were measured without considering the traffic lanes in order to take in to account the lack of lane discipline. The heterogeneity was incorporated by converting the data into standard Passenger Car Unit (PCU) equivalent values. An optimal non-linear speed-density relationship which minimizes the sum of squares of residuals using a conventional optimization algorithm based on the Trust Region Method was developed. Using this relation and the fundamental equation connecting speed, flow and density, the relationships between speed-flow and flow-density were derived.