A Development Technique for Expert Systems of Road Safety Management in Settlements
The paper considers several aspects of accidents prediction at conflict objects, such as cross-roads, sidewalks and crossings. The impact of road conditions and other independent factors on traffic safety is analyzed, and methods for assessment of dangerous road segments are investigated. Specifically, the basic classification principles of road safety management methods are proposed, and specific features of expert systems development depending on properties of different classes are elaborated. The method of conflict situations for traffic accidents (TA) prediction is proved to be effective, and its justification is provided. A class of potential dangers governed by regulatory and legislative acts is outlined. For this class, a method for developing a decision-making automated system for safety management is proposed and discussed. A unique feature of the proposed method is that decision-making is performed by an expert system completely, and no experts are required for developing and operating the system. In the paper, basic principles of a knowledge base (KB) development and features of algorithms are proposed. Structured management decisions are made on the basis of the proposed algorithms using KB. The effectiveness of the proposed method is validated by the results of monitoring.
Key words: traffic accident, predicting methods, method of conflict situations, factors of influence, knowledge base, expert systems, decision-making systems
Full text at PDF
, 755Kb. Language: Russian.