A need for empirical validation of the 1985 highway capacity manual (HCM) arises from its development with limited field data and its adoption of radically new procedures. Chapter 9 of the manual, signalised intersections, contains some of the more extensive changes and is the subject of field validation. Emphasis is placed on comparisons of field saturation flows, delays, and progression factors with those estimated by the manual. Data are obtained from ten intersection approaches for twenty-five 15-minute analysis periods. Saturation flows in the field are significantly higher than those in the manual. A value of nearly 2,000 vehicles per hour of green per lane (vphgpl) is consistently observed at field sites where the HCM estimates values of 1,800 passenger cars per hour of green per lane. The difference in saturation flows significantly affects delay estimates: the delays predicted using the value of 2,000 more closely match field delays than do the estimates using 1,800. Our conclusion is that an ideal saturation flow equal to 2,000 (rather than the HCM value of 1,800) is recommended for the analysis of high-design intersections similar with those analysed in this project. The progression factors estimated from our data are statistically significant and differ significantly from HCM values. Our platoon factors are lower than the HCM, indicating that greater reductions in delay are necessary than are currently provided in the progression factor adjustments. In addition to being lower than current HCM values, our progression factors are much less sensitive to the platoon ratio (a relative measure of the percent arrivals on green). It appears that our field data for actuated controllers is almost totally inconsistent with the current HCM.
Abstract