816 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 3, MAY 2001 IDUTC: An Intelligent Decision-Making System for Urban Traffic-Control Applications M. Patel and N. Ranganathan Abstract—The design of systems for intelligent control of urban traffic is important in providing a safe environment for pedestrians and motorists. Artificial neural networks (ANNs) (learning systems) and expert systems (knowledge-based systems) have been extensively explored as approaches for decision making. While the ANNs compute decisions by learning from successfully solved examples, the expert systems rely on a knowledge base developed by human reasoning for decision making. It is possible to integrate the learning abilities of an ANN and the knowl- edge-based decision-making ability of the expert system. This paper presents a real-time intelligent decision making system, IDUTC, for urban traffic control applications. The system inte- grates a backpropagation-based ANN that can learn and adapt to the dynamically changing environment and a fuzzy expert system for decision making. The performance of the proposed intelligent decision-making system is evaluated by mapping the the adaptable traffic light control problem. The application is implemented using the ANN approach, the FES approach, and the proposed integrated system approach. The results of extensive simulations using the three approaches indicate that the integrated system provides better performance and leads to a more efficient implementation than the other two approaches. Index Terms—Artificial neural network, expert system, integra- tion, intelligent vehicle highway system (IVHS), urban traffic con- trol (UTC). I. INTRODUCTION P OPULATION growth has increased the number of vehi- cles and passengers on the country’s freeways and high- ways. Since the current transportation infrastructure has not kept pace with the growth in traffic demand, research to develop modern transportation systems has become important. With in- adequate space and funds for the construction of new roads, and the growing imbalance between traffic demand and transporta- tion resources, intelligent highway vehicle systems (IVHSs) are gaining interest. The study of IVHS solutions has evoked sub- stantial interest in Europe, Japan, and the United States. An IVHS system would perform tasks that are typically done by human operators and use advanced technologies from various fields such as image processing, computer vision, intelligent controls, and artificial intelligence (AI). The IVHS systems are classified into four categories: 1) advanced traffic management system (ATMS); 2) advanced driver information system (ADIS); 3) freight and fleet control system (FFCS); 4) automated vehicle control system (AVCS). Manuscript received November 1, 1997; revised October 1, 2000. M. Patel is with the Honeywell Space Systems Commercial Systems Divi- sion, Clearwater, FL 33764 USA. N. Ranganathan is with the Computer Science and Engineering Department, University of South Florida, Tampa, FL 33620 USA. Publisher Item Identifier S 0018-9545(01)03954-8. The ATMS systems perform tasks such as surveillance, control, and management of freeway and arterial networks. Such appli- cations include intersection traffic light control and congestion and incident management. The ADIS systems are responsible for such tasks as origin–destination calculation and motorist ad- visories. They provide information such as efficient alternatives for reaching destinations based on time and road conditions. The FFCS systems manage cargo and freight traffic. Since cargo and freight vehicles are massive in number and size, efficient systems are required to ease fuel consumption, congestion, and road wear and tear caused by such transports. The AVCS sys- tems include vehicle platooning, obstacle avoidance, and au- tonomous vehicle guidance. The system concentrates on intel- ligent guidance systems for vehicles. They are capable of either driving the vehicle in a fully automatic manner or giving the human driver useful advice. Such systems would allow drivers to operate at high speeds and simultaneously reduce the proba- bility of having accidents or collisions. A. Intelligent Decision Making Systems Artificial Intelligence techniques used in IVHS systems in- clude artificial neural networks (ANNs) [7], [3], expert systems [23], [2], [27], and fuzzy logic controllers [29], [24], [5], [3], [4], [15], [26]. There are two common approaches for intelligent de- cision making: one based on learning systems, such as the ANNs, and the other based on expert systems. In a learning system, the decisions are computed using the accumulated experience or knowledge from successfully solved examples. The learning sys- tems use various methods and mathematical models to exploit the computational power of a computer, regardless of the inferencing power of humans. ANNs can be used to compute solutions for complex problems. They possess an adaptive feature that allows each cell within the network to modify its state in response to experience. The neural network can then learn or self-modify. Often, ANNs have been used to mimic expert systems. In an expert system, problems are solved using a computer model of human reasoning. Various implementations of expert systems can be found in the literature. Several system architectures for im- plementing neural networks and a few schemes for implementing expert systems exist in the literature. The problem of integrating a neural network and a fuzzy expert system and its application to urban traffic control is the main focus of this work. An intelligent decision-making system must 1) be able to solve problems of practical nature and size, 2) always arrive at correct solutions, and 3) adapt to the changes in the application environ- ment. The typical human is constantly faced with making impor- tant decisions and almost always uses prior knowledge or experi- ence in determining them. Experts have accumulated knowledge 0018–9545/01$10.00 © 2001 IEEE