Left-turn treatments are essential to signal capacity and operational safety at signalized intersections. left-turn warrants and guidelines are sets of evaluation procedures designed to maximize levelof service, minimize approach delay, and reduce left-turn-related accidents. currently, three left-turn phasings are used to allow vehicles to make left turns on a green arrow or a circular green indication: permissive, exclusive, and exclusive/permissive. an experimental expert system design for recommending alternative left-turn phase selection on microcomputer systems was investigated. the goal of this study was to computerize left-turn phase selection by using artificial intelligence languages and knowledge engineering. this study focuses on investigating expert systems programming using prolog and the insight 1 system in an ibm pc/xt/at microcomputer environment. three experimental systems were developed using the pd prolog system, the turbo prolog system, and the insight 1 knowledge engineering system. the background of the study, the artificial intelligence concept, the basic systems design, and the practical experience gained arediscussed. potential advantages and disadvantages of developing expert systems using different artificial intelligence languages and the knowledge engineering for traffic engineering applications are evaluated. the results of this limited study indicate that it is feasible to combine artificial intelligence and traffic engineering technologies for alternative traffic signal analysis. this paper appeared in transportation research record no. 1145, expert systems for transportation applications. for covering abstract see irrd no 817754.
Samenvatting