Fundamental relationship investigation using cellular automata

Abstract

Cellular automata (CA) have been used by many researchers for homogeneous traffic flow modelling. In
country like India, non lane based traffic prevails; hence, designing control systems for such situations is a challenging
task. Traffic simulation helps the analyst to model the behaviour of such complex systems. To represent multiple vehicle
types, a multi cell representation was adopted. The position and speed of vehicles are assumed to be discrete in
developed model. The speed of each vehicle changes in accordance with its interactions with other vehicles and is
governed by some pre-assigned (stochastic) rules. In this paper, mid block section two-lane heterogeneous traffic flow
model(MBTFM) using CA is presented. A conventional CA model is modified to handle heterogeneous traffic. While
addressing heterogeneity, essentials features of the CA like uniform cell size, simple rules, and the computational
efficiency achieved in homogeneous models are retained to a great extent by modifying cell size and developing certain
rules which control vehicles as well as the road characteristics. The cell size is reduced to accommodate more ranges of
speed, acceleration, size of vehicles and deceleration as well as to represent aggregate behaviour of vehicles. Further, to
address the issue of non lane based movement, lateral movement rules were proposed.
Calibration and validation of the model is also carried out at macroscopic level using real-world data. The model is
verified for the microscopic characteristics with a conventional car-following simulation model. The model performed
reasonably well. The results show that developed model can be used as an effective tool for heterogeneous traffic flow
modelling considering the computational benefit and simplicity in developing CA models.