原文传递 Knowledge Acquisition Methods for the IHDS Diagnostic Review Expert System.
题名: Knowledge Acquisition Methods for the IHDS Diagnostic Review Expert System.
作者: HUGHS, W. E.; MCGEE, H. W.; SCHUCKEL, C.; SMITH, M.
关键词: KNOWLEDGE ACQUISITION
摘要: The Federal Highway Administration's Interactive Highway Safety Design Model (IHSDM) is a suite of CADD-compatible programs that highway designers can use to evaluate the safety effects of various design alternatives. The IHSDM will include a Policy Review Module, a Design Consistency Module, a Driver/Vehicle Module, a Traffic Analysis Module, and an Accident Analysis Module. IHSDM development efforts have thus far been concentrated on two-lane rural roads. Current plans for the Accident Analysis Module consist of models to estimate the number and severity of accidents on specified roadway segments, a benefit/cost analysis model to evaluate alternate roadside designs, and an expert system that can evaluate a design alternative and identify geometric deficiencies that may impact safety. In this report, the term Diagnostic Review Component (DRC) is used to define this expert system. The key to the DRC will be the knowledge base that will be accessed to 'Sag' potential design-related safety deficiencies. In addition, the DRC will ultimately have the capability to provide users with suggested improvements that could mitigate the safety problems. This report presents the results of a feasibility study of alternative methods for developing the knowledge base for the DRC. Alternative sources of knowledge that were reviewed included the following: Interviews with State highway design personnel and review of State highway safety improvement programs; Interviews with experts in safety, geometric design, accident investigation traffic engineering, and human factors; In-depth investigations of accident sites; and review of highway safety research and safety-audit literature. The advantages, disadvantages, and development implications for the expert system's knowledge base are discussed for each method Findings from the feasibility investigations of each method are documented. Finally, recommendations for subsequent research, including a detailed experimental plan for the development of a prototype DRC, are presented.
报告类型: 科技报告
检索历史
应用推荐