Modeling Right Turning Vehicles at Uncontrolled T-shaped Intersection Using Grid-based Cellular Automata



Modeling Right Turning Vehicles at Uncontrolled T-shaped Intersection Using Grid-based Cellular Automata

Authors

Jayant Sangole, Ph.D. Student, Dept. Of Civil Engg., IIT Bombay, Powai, Mumbai, Prof. Gopal R. Patil, Assitant Professor, Dept. Of Civil Engg., IIT Bombay, Powai, Mumbai

Description

Unsignalized intersections are very common in urban road network. Unsignalized intersections in developed countries are controlled using stop and yield signs, which decide priorities for different movements. However, in India, no priority is followed at most unsignalized intersections, thus the intersections remain uncontrolled. There exists a mix of various types of light and heavy vehicles with wide variation in sizes, characteristics, and operating conditions trying to cross the intersections as quickly as possible, resulting in a chaotic condition. Most of the approaches/models developed to analyze unsignalized intersections in developed countries cannot be directly applied in India. Most studies on evaluating unsignalized intersections use gap acceptance approach in which it is assumed that a vehicle on lower priority movement cross the intersection when a sufficient gap is available on a higher priority movement. This concept will not work at places where vehicles of all movements entering the intersection and compete to cross the intersections as quickly as possible. A better approach for such situations will be to simulate the intersection.
Cellular Automata (CA) has been widely used to model traffic flow and to simulate various traffic flow conditions. Conventionally, in such models the width of a cell is taken equal to the lane width. This is suitable if traffic has lane discipline and where multiple vehicles do not travel parallel on a single lane. In order to capture the realistic traffic behavior in India, we used grid-based cellular automata approach.

Abstract

Modeling right turning vehicles at uncontrolled T-shaped intersection using grid-based cellular automata

Gopal R. Patil1, Jayant P. Sangole2,
1Assistant Professor , 2Ph.D. Student, Civil Engineering Department IIT Bombay, Powai
Mumbai-400076, India
Email for correspondence: gpatil@iitb.ac.in

Unsignalized intersections are very common in urban road network. Unsignalized intersections in developed countries are controlled using stop and yield signs, which decide priorities for different movements. However, in India, no priority is followed at most unsignalized intersections, thus the intersections remain uncontrolled. There exists a mix of various types of light and heavy vehicles with wide variation in sizes, characteristics, and operating conditions trying to cross the intersections as quickly as possible, resulting in a chaotic condition. Most of the approaches/models developed to analyze unsignalized intersections in developed countries cannot be directly applied in India. Most studies on evaluating unsignalized intersections use gap acceptance approach in which it is assumed that a vehicle on lower priority movement cross the intersection when a sufficient gap is available on a higher priority movement. This concept will not work at places where vehicles of all movements entering the intersection and compete to cross the intersections as quickly as possible. A better approach for such situations will be to simulate the intersection.
Cellular Automata (CA) has been widely used to model traffic flow and to simulate various traffic flow conditions. Conventionally, in such models the width of a cell is taken equal to the lane width. This is suitable if traffic has lane discipline and where multiple vehicles do not travel parallel on a single lane. In order to capture the realistic traffic behavior in India, we used grid-based cellular automata approach. In this approach, an intersection is divided into grid of cell size 1.2m X 1.2m. Cell size is decided on the basis of the smallest vehicle size in Indian traffic, that is, two-wheeler. In India, vehicles are driven on the left side of a road. New crash avoiding rules are developed for right turning movements at T-shaped uncontrolled intersection, where both the through and turning vehicles have equal priority. This is implemented in programming language Java and is being calibrated to replicate the field conditions. Video data at uncontrolled T-intersections are collected from various cities in India (that is, Thane, Navi-Mumbai, Aurangabad, Kalyan). Different parameters including arrival rate, turning rates, type of vehicle, speed of vehicle, gap, conflict points, and congestion on the intersection are extracted from the data.
Once the model is calibrated and validated, it will be used for evaluating the performance of T-shaped intersections at different demand levels. Various parameters such as delay, capacity, speed will be studied by changing input parameters. This will help in developing models to estimate these parameters.

Publisher

Association for European Transport