The general algorithms, assumptions, validations and example application are presented for estimating roundabout safety performance in comparison to signalized and unsignalized intersections using conflict opportunity technology rather than regression-based accident estimates. Comparison of conflict opportunity accident estimates versus onsite accident records for 100 signalized and unsignalized intersections are presented along with single and dual lane roundabouts accident comparisons compared to Maryland DOT data. For both intersections and roundabouts, annual accident prediction accuracies of approximately 80 percent for total accidents and 70 percent for angle and rear-end accidents was achieved compared to the 3-5 year average reported historical crash events. When combined with Highway Capacity Manual (HCM) delay estimates and delay values, the conflict opportunity-based injury and HCM delay values can be combined into a Performance Index for comparing alternate traffic control scenarios. An example safety+delay performance evaluation of a single lane 15,000 ADT roundabout compared to Two and All-way stop and signalized control indicated that All-Way stop had excessive delay, a signal would not be warranted and of the remaining options, a roundabout provided annual safety+delay performance sufficiently comparable to Two-way stop control and was thus the developers preferred design alternate.
Samenvatting