A Review of Cellular Automata Model for Heterogeneous Traffic Conditions

Abstract

Over the years various microscopic traffic models were developed to predict vehicular behaviour from mid-block section of road to the network level. Cellular Automata (CA) was found to be a promising approach to meet this challenge in the recent past. A CA approach to traffic simulation is potentially useful in order to achieve a very high computational rate in microscopic simulation, and to facilitate distributed computing. Because of this CA models are becoming increasingly popular for their potential to simulate large scale road network using macroscopic traffic characteristics like flow and density. Despite an increase in computational power over the past decade, limited efforts have gone in evaluating the model at microscopic level using characteristics like lane keeping and lane change. These characteristics along with traffic composition and density have significant influence on the amount of interaction between different vehicle types. This paper provides a brief review of CA models developed for heterogeneous traffic conditions and provides insights for improvement. Model performance is evaluated at macroscopic and microscopic level using characteristics like speeds and positions obtained from vehicle trajectories. The data was collected on arterial roads in Ludhiana, India for this study.