A procedure is reported to enable an engineer to place all types of intersection control on a comparable basis so that a control can be selected which promises the best possible operation under given conditions. Data were collected at seventy-one intersections in the Chicago, new York, San Francisco, and Toronto metropolitan areas to derive the prediction equations. The statistical curve-fitting technique of regression analysis was used. The operational criteria were predicted on the basis of such intersection conditions as volume, approaching vehicle speeds, sight distance restrictions, and street widths. The effect of the geographic area, distance to the nearest signal and functional classification of the intersecting streets were included. Studies were made of yield, two-way stop, and four-way stop controlled intersections only. Controls are recommended for consideration based on street priority and continuity of flow conditions as summarized in tabular form. When more than one control qualifies for further consideration after meeting system requirements, the final selection is proposed to be made on the basis of an index of individual intersection operations. This would be derived by predicting delay and accident experience under each of the controls being considered. The predicted values would be estimated through the use of nomographs based on empirical equations derived from research data. The index would be calculated by transforming the delay and accident values to a scale of comparison and applying a weighting factor to account for the relative importance of the criteria. The control having the best index of operation would then be selected. The methods need refinement for general application and should be extended to cover uncontrolled and signalised conditions.
Abstract