Development of a rating system for the conversion of an all-way stop controlled intersection to a roundabout

Modern roundabouts are becoming increasingly popular in the United States and merit consideration for use at intersections facing operational and/or safety problems. In this thesis, a knowledge-based expert system is used to create an automated tool for evaluating an all-way stop-controlled (AWSC) intersection for its potential success if converted to a roundabout. Traffic analysis software is used to conduct a comparative analysis between AWSC, roundabout, and signalized intersections. The software analysis was calibrated for local (Delaware/Maryland) conditions using driver gap acceptance characteristics obtained from data collection at two Maryland roundabouts.; Data reduction yielded a local driver critical gap of 3.9 seconds and a follow-up time of 2.0 seconds; both values are slightly lower than those recommended by the Highway Capacity Manual. The findings of the comparative analysis agree with earlier studies in showing that AWSC intersections always perform "worse" (less capacity, more delay, etc.) than both roundabouts and signalized intersections under the same traffic and geometric conditions. Results from the software analysis were then used in developing the rules of the expert system. The developed system was tested using real AWSC intersection traffic data. It was shown that the system is capable of logically evaluating an AWSC intersection and recommending an intersection type (roundabout or signal) to which to convert. The rating system also provides advice on a variety of issues associated with the new intersection type and references for further guidance.