A Novel Fuzzy Method to Traffic Light Control Based on Unidirectional Selective Cellular Automata for Urban Traffic Moein Shakeri 1 , Hossein Deldari 1 , Alireza Rezvanian 2 , and Homa Foroughi 1 1 Department of Computer Enginnering, Ferdowsi University of Mashhad, Mashhad, Khorasan, Iran 2 Department of Computer Enginnering, Azad University of Qazvin, Qazvin, Iran Abstract — Vehicular travel which demands on the concurrent operations and parallel activities is increasing throughout the world. In this paper to control urban traffic, we study the optimization of traffic light controller in a city and present a fuzzy algorithm based on cellular automata. In existent system factors like priority of streets of intersection and width/length of streets are assumed equal. However, in real situations parameters like time during the day, density of vehicles of street and number of shopping centers have determinant effects on amount of traffic of streets. To overcome such limitations we propose a three leveled fuzzy system. At first level priority of street is computed based on fuzzy rules. At second level real velocity of vehicles is calculated. We use CA for simulating vehicles’ density transmission. At third level by regarding priority of street and amount of density, decision for changing status of traffic light is done. Index Terms — Cellular Automata Model, Fuzzy Control, Traffic Control, Traffic Light Control, Unidirectional Selective Cellular Automata, Urban Traffic. I. INTRODUCTION Transportation research has the goal to optimize transportation flow of people and goods. As the number of road users constantly increases, and resources provided by current infrastructures are limited, intelligent control of traffic will become a very important issue in the future [6]. Optimal control of traffic lights using sophisticated sensors and intelligent optimization algorithms might therefore be very beneficial. Optimization of traffic light switching increases road capacity and traffic flow, and can prevent traffic congestions. In the recent years there were strong attempts to develop a theoretical framework of traffic science among the physics community. Consequentially, a nearly completed description of highway traffic [7, 11], e.g., the “Three Phase Traffic” theory, was developed. This describes the different traffic states occurring on highways as well as the transitions among them. Also the concepts for modeling vehicular traffic are well developed. Most of the models introduced in the recent years are formulated using the language of cellular automata (CA) [3,4,12,14-17]. Unfortunately, no comparable framework for the description of traffic states in city networks is present. In contrast to the highway networks, where individual highway segments can be treated separated, the structure elements of city networks exert an immense influence onto the traffic dynamics [7]. In existent urban traffic systems priority of intersection streets and also width and length of streets are assumed equal; so they have no role in making decision for changing the status of the traffic light. To overcome such limitations, we propose a novel fuzzy system that momently computes priority of the street based on fuzzy rules and regarding to environmental factors. Furthermore, in the proposed system length of streets are not considered the same. Subsequently a cellular automaton is employed for modeling both density transmission and type of movement of vehicles. In fact each street forms a lattice and each member of this structure is regarded as a separate cell in cellular automata space. The paper is organized as follows: in section II cellular automata is described concisely and then unidirectional selective cellular automata is introduced for further simulations. In section III we briefly review some existing traffic light controller systems. Our proposed system and its technical details are described in section IV. Simulation results are represented in section V and finally we draw conclusions in section VI. II. CELLULAR AUTOMATA Cellular automata (CA) are a class of discrete dynamical systems [13], consisting of an array of nodes (cells) of some dimension, n. Each cell can be in one of k different states at a given tick of the clock. At each discrete tick of the clock, each cell may change its state, in a way determined by the transition rules of the particular CA [1,5]. The transition rules describe precisely how a given cell should change states, depending on its current state and the states of its neighbors. Let n be the dimension of the lattice, k the number of states, T the transition rule function, C t ( i 1 , … , i n ) the state of the cell at position ( i 1 , … , i n ) at time t, N t ( i 1 , … , i n ) the values (given in a specific order) of the neighboring cells to this location at time t. Then the dynamics of the CA is completely specified by the initial states of all the cells, C 0 , along with the recursion rule (Equation 1). (1) Proceedings of 11th International Conference on Computer and Information Technology (ICCIT 2008) 25-27 December, 2008, Khulna, Bangladesh 1-4244-2136-7/08/$20.00 ©2008 IEEE 300