KNOWLEDGE-BASED GEOGRAPHIC INFORMATION SYSTEM FOR SAFETY ANALYSIS AT RAIL-HIGHWAY GRADE CROSSINGS

The development of a knowledge-based geographic information system for managing and analyzing safety-related information for rail-highway grade crossings is discussed in this paper. The allocation of federal funding for safety improvements at public, at-grade rail-highway crossings is made based on the performance of the states with respect to accident reduction. The motivation behind the work was to establish guidelines and to develop an integrated system that would ultimately result in accident reduction through better access and management of safety information. This was accomplished by using geographic information system (GIS) technology and decision support tools through integration of the GIS application with a statistical model and a knowledge-based expert system (KBES). The work continued an ongoing project that resulted in the integration of rail-highway grade crossing safety data from various sources, such as the FRA and Delaware Department of Transportation (DelDOT), into a data base management system. The selection and integration of the U.S. Department of Transportation (USDOT) accident prediction model into the system was also required. This paper describes the conversion of rail-highway grade crossing safety attribute data into a GIS-acceptable format, the development of the GIS application including the spatial analysis, visual display, and query capabilities, and the development of a KBES to account for site-specific and qualitative factors. The KBES is also capable of suggesting safety upgrade action(s) at the crossing. Interfacing of the KBES and the program for the USDOT model with the GIS and the framework of the complete package developed for safety analysis at rail-highway grade crossings are also discussed in the paper.

Language

  • English

Media Info

  • Features: Figures; References; Tables;
  • Pagination: p. 91-100
  • Monograph Title: Artificial intelligence and geographical information
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00714948
  • Record Type: Publication
  • ISBN: 0309061636
  • Files: TRIS, TRB
  • Created Date: Dec 20 1995 12:00AM