Advances in computer technologies have been changing transportation field. Based on those, Intelligent Transportation System has been proposed and being developed for the next generation transportation system. This system requires more utilization of Artificial Intelligence, such as Knowledge Based Expert System, than current one. Although Knowledge Based Expert System has not been robustly studied in the field of transportation, it is suggested to be implemented in Traffic Adaptive Signal Control System as a part of Intelligent Transportation System. Therefore, it is fundamental to overview that what the current position of Knowledge Based Expert System in transportation is. It will help to construct a specialized knowledge on and understand that system in Transportation. Previous applications of Knowledge Based Expert System in transportation field are briefly reviewed, and potential implementations are discussed by focusing on Traffic Adaptive Signal Control System in this paper.
Key words: Knowledge Based Expert System, Intelligent Transportation System, Traffic Adaptive Signal Control System.
Advances in computing have dramatically changed the transportation field in the last decades. These enhance the speed and credit of traffic analysis and allow transportation engineers to study various problems that would not be possibly answered without computer techniques, such as simulations. On the other hand, transportation system have been experienced congestion during its peak hours. To solve this, its facilities have been expended in order to increase their capacity by means of adding additional lanes on existing roadways and building new links between points where high traffic demand exits. Being experienced with negative environmental impacts and high land cost from those processes, transportation engineers noticed the need of new technologies to maintain transportation system to be enable to handle traffic without constructing new facilities.
Intelligent Transportation System(ITS) is an ongoing movement, which starts in the late 1980s, intending to maximize the utilities of transportation facilities. It is based upon the utilization of advanced computer technologies in communication between hardware control systems and in decision making processes on various tasks. Many advanced studies have been achieved under the name of ITS, and they include the application of Artificial Intelligence, such as Knowledge Based Expert System (KBES) which will be dealt with in this report.
KBES has implemented in limited fields of transportation; it is mainly used in pavement management system including road maintenance and road rehabilitation. Previous implementation of KBES in transportation field are briefly reviewed in the following section. In addition, potential applications of KBES are discussed by focusing on the Traffic Adaptive Signal Control System, which is under the ITS project and is suggested to be designed to use KBES by many researchers.
KBES is an intelligent computer program that uses the knowledge and inference procedures of human experts to solve difficult problems, and it is not robustly studied in the field of transportation. Some KBES projects are achieved by the several state Department of Transportation (DOT). Their concerns are mainly in Pavement Management System and some others [1]. Connecticut DOT has developed the Pavement Rating and Analysis System. It is a network-level pavement condition rating tool that uses laser videodisc. Cracks and other type of distress are rated based on closeups of images. PARADIGM (Pavement Rehabilitation Analysis and Design Mentor) was developed by Ritchie et al in California. It is a microcomputer-based and integrated set of interacting expert systems(1) and algorithmic models for flexible pavements, emphasizing asphalt concrete surfaces and overlays. Minnesota DOT has contracted to develop the ROUTBUILDER that guides the administrator through the process of collecting the necessary information, find an appropriate legal route between two points for lead in question and automatically issue the permit to the trucking company. New Jersey DOT developed CHINA (Computerized HIghway Noise Analysis) which is a tool developing a series of routines that could be used to effectively design a noise barrier. This system is developed to address a nationwide lack of experience in the acoustic design of highway noise barriers. It is intended to develop a series of routines that could be used to effectively design a noise barrier, test these routines independently one another, and develop a control structure that would coordinate their activities. It is designed by EXSYS, which is an expert system design shell written in C. There are many other KBES applications accomplished by other DOT, but their implementations are still limited in such fields of transportation.
Traffic Adaptive Signal Control system is the core part of Automated Traffic Management System(ATMS), which is one of technology-oriented ITS areas consisting of three parts: AITS (Advanced Traveler Information System), AVCS (Automated Vehicle Control System) and ATMS.
One of the main goals of ATMS is to manage the traffic flow in real-time by predicting future traffic flow in short time period [2]. In Traffic Adaptive Signal Control system, KBES has been suggested to be used in the process of determining optimal strategy to control traffic signal under various traffic condition [3] [4]. At five different research centers, new methodologies for that particular system are produced. Most of them suggest that a hierarchical structure of traffic control phases should be standardized and that KBES be used in real time(2) [3] [5] [6]. The implementation of KBES in this problem can be justified with the following two reasons. First, there are a limited number of transportation engineers who can work for transportation management system in real time. Large number of intersections are operated and monitored in urban transportation network, but it is practically impossible that well trained transportation engineers can be assigned to manage those in Transportation Management Center twenty-four hours a day. Second, there is a best solution that is from existing engineering methodologies and that can be obtained by human experts. Since the current traffic control system dominantly uses historical traffic volume data, which is equivalent to the assumption that traffic is uniformly generated over the period, the output from the current system is not the best one in realistic. In addition, traffic actuated control system can only be dealt with an isolated intersection, so it can not provide the optimal in system wide. Implementation on Traffic Adaptive Signal Control System will provide following advantages. First, it can mainly reduce delay time and average stops of vehicles and provide better performance of traffic flow. Second, KBES assists operators in Transportation Management Center to work efficiently, when unexpected incidents occur, by utilizing experts' knowledge accumulated in the system. Real-time KBES has studied in a similar case. Stack and Ritchie has researched real-time freeway incident management by using G2, which is an expert system shall designed for the development and execution of complex applications that require intelligent monitoring and control [7]. This system has tested in Transportation Management Center in California and have good responses from the system operators.
Current position of KBES in the field of transportation and a potential implementation of KBES in ATMS are reviewed in this paper. Previously, KBES has implemented in limited transportation fields, such as pavement management system and transportation logistics. Further application of KBES is considered to be embedded in the Traffic Adapted Signal Control System in ATMS. It determines the optimal traffic control strategy under various traffic condition and also assists operators in Transportation Management Center to work efficiently when unexpected traffic incidents occur. Its success is significant in the field of traffic control because it can mainly reduce delay time and provide more effective performance of traffic flow.
1. Cohn L. F. and Harris R. A. "Knowledge Based Expert Systems in Transportation", NCHRP Synthesis 183, Transportation Research Board, National Research Council, Washington D.C.,September 1992.
2. Head K. L. "An Event-Based Short-term Traffic Flow Prediction Model", Paper No 950838, Transportation Research Board's 74th Annual Meeting, Washington D.C., January 1995.
3. Gartner N.H., Stamatiadis C. and Tarnoff P. J., "Development of Advanced Traffic Signal Control Strategies for IVHS: A Multi-Level Design", Paper No 950974, Transportation Research Board's 74th Annual meeting, Washington D.C., January 1995.
4. Morris M. A. and Potgieter L. J., "Expert System for Aspects of the TSM Process", Transportation Research Record No. 1280, Transportation Research Board, National Research Council, Washington D. C., 1990.
5. Ritchie S. G., "A Knowledge-Based Decision Support Architecture For Advanced Traffic Management" Department of Civil and Environmental Engineering and Institute of Transportation Studies, University of California, Irvine, CA September 1989.
6. Dell'Olmo P. and Mirchandani P. "An Approach for Real-time Coordination of Traffic Flows on Networks", Paper No 950837, Transportation Research Board's 74th Annual meeting, Washington D.C., January 1995.
7. Stack R. and Ritchie S.G., "Real-Time Expert System for Freeway Incident Management" Institute of Transportation Studies, University of California, Irvine, CA, December 1993.
2.
In addition, it has resulted that each method from those five research centers is applicable in different traffic condition also.