Highway Research Record

Heterogeneous Traffic Flow Modelling for India

Date of Start:July 2003
(i) Indian Institute of Technology, Delhi (R,I)
(ii) Department of Science and Technology (S)

The objective of this study is to develop a modelling framework that would simulate the heterogeneous traffic patterns in urban road networks. The system will be forming an evaluation tool for forecasting the traffic behaviour in various scenarios: those due to external factors like adding new lanes, new parking regulations and internal traffic aspects like changes in vehicular mix, reducing or separating the non-motorized traffic etc. Thus the system developed will not only serve as a tool for replicating the existing traffic behaviour but also help to predict changes in the future, for example, the impact of new car technologies on the road system. The model will also help in estimating the congestion indices on the overall network and will be able to suggest measures to improve the situation.

Cellular Automata (CA) models are discrete dynamical systems that have the capability to describe continuous dynamical systems, capable of simulating complex traffic flow phenomena with minimum computational effort. Time, space and states are discrete and many traffic characteristics are modelled implicitly in CA model, which led to its computational efficiency. In this approach, a lane is represented by one-dimensional lattice.
One lane model is developed in Java Programming language, the results of which are in agreement with several empirical features observed in homogeneous traffic conditions such as capacity drop, hysteresis and meta-stability. A two-lane CA model is also developed for homogeneous traffic conditions with proper use of car following and lane changing rules. To model the heterogeneous traffic conditions, the empirical features of heterogeneous traffic are being analysed. Later this model will be extended to typical elements of network such as T-junctions and intersections.

Data Collection

Data required for this modelling approach is collected from a mid-block of an arterial road in Delhi using video cameras. Different parameters of this model namely breaking probability, time step. etc, are obtained based on empirical flow- density diagram and other field observations. Validation of the model is done by comparing different parameters of the model with the field observations.

Findings and Conclusions
With a proper validation at link level, this model can be of great help in developing a network model, which can be used to study the flow properties at macroscopic level. This will help in regulating the traffic at network level so that congestion problems in urban areas can be minimised.

1. Ch. Mallikarjuna and K. Ramachandra Rao, (2005) “Traffic Flow Modelling on Highways using Cellular Automata: A Review”’     Proceedings of International Conference on ‘Structural and Transportation Engineering (START- 2005), N.Delhi, pp        912- 919.
2. Amit Utkarsh Sinha, Arun Kochar and K. Ramachandra Rao, (2005) “Traffic Flow Modelling on Highways using Cellular     Automata: A Review”’ Proceedings of International Conference on ‘Structural and Transportation Engineering (START-     2005), N.Delhi, pp 928-935.