Knowledge-Based System to Support Site Investigation for Safety Improvement

Specialists in transportation and other areas of engineering are trying to improve road safety. The most difficult part is determining the road features, driver behaviors, and vehicular failures that unduly increase the risk of crash. To help INDOT safety investigation teams, a previous research project developed a prototype Road Safety Investigation Tool (RSIT) for two-way stop-controlled intersections. The research summarized here has evaluated the prototype RSIT and has developed a new version RSIT2 that has incorporated the findings of the evaluation. The evaluation of the RSIT prototype confirmed usefulness of the proposed methodology. The tool has been applied to two two-way stop-controlled intersections by a team of non-experts. The two sites have been previously investigated by experts using the standard approach. The safety investigation outcomes obtained by both the teams indicate strong similarities The proposed RSIT method helped non-experts identify additional road hazards confirmed later by the experts' team. RSIT has been found decreasing the required investigation team size and time as well as being useful as a training tool. RSIT presents the logical connections between driver behaviors, road elements and environment conditions. RSIT2 is the next version of the RSIT method. The knowledge base developed for RSIT 2 includes all the major types of roads in Indiana: a) all-way stop-controlled intersection, b) road segment, c) railroad crossing, and d) signalized intersection, and e) improved two-way stop-controlled intersections. The RSIT 2 method uses the rule-based structure converted to a dynamic checklist that adjusts in real-time to the investigation progress. A friendly graphic-user interface has been developed together with a knowledge base editor to allow updating the knowledge base by the user. Finally, a novel technique of ranking safety countermeasures was developed and applied.