Using Segment Speed to Estimate 85th Percentile Spot Speed on Nonfreeway Facilities

Abstract When speed limits are set via an engineering speed zone study, the operating speed for a site is needed. This research project investigated whether probe segment speed data (rather than speeds measured in the field) could reasonably be used in a Texas speed zone study. The idea for using probe segment speed data is to develop conversion equations that would convert an average yearly speed measure obtained from probe segment speed data into a speed measure that would represent free-flow conditions for passenger cars. Equations for 85th percentile speed and average speed were created for nonfreeway facilities subdivided into urban and rural contexts, and their performances were evaluated. For nonfreeway facilities, the adjusted R-squared values for the regression equations indicate that the independent variables used in the models can explain between 71% and 89% of the observed speed variability. The root mean square error implies that, on average, the predicted speeds are off by about 4.5 to 5.6 kph (2.8 to 3.5 mph). Given that the speed limit decision process rounds the spot speed to the nearest 8.0 kph (5 mph) increment, the developed equations produce a spot speed that would result in similar or within 8.0 kph (5 mph) speed limit value.