Real-Time Traffic Density Estimation without Reliable Side Road Data

Abstract

One of the most popular intelligent transportation systems (ITS) applications is to provide real-time road traffic congestion information to the users. Traffic density is a major congestion indicator, and because its measurement is difficult, it is usually estimated from other readily measurable parameters. Several studies have explored various approaches for density estimation for homogeneous and lane-disciplined traffic conditions. However, Indian traffic is different with its heterogeneity of traffic and absence of lane discipline. Another characteristic is the lack of access control, making automated measurement of net entry into a study section difficult. An added difficulty is that the roads in India are not yet equipped with traffic sensors, leading to limitation in data collection. The present study mainly addresses the issue of estimating traffic density in the absence of automated sensors at the side roads/ramps on Indian roadways. A lumped parameter macroscopic traffic flow model has been formulated, and using this model, a model-based estimation scheme has been designed based on the Kalman filtering technique. The only data required for implementing this method in the field are the flow passing the entry location and the spot speeds of vehicles passing through the entry and exit locations. The proposed method was corroborated using data measured from a road stretch in Chennai, and the performance was found to be satisfactory.