Running Time Prediction for Signalized Urban Streets

The development of a procedure for predicting the average running time along an urban street segment is described. Running time is shown to be a function of free-flow speed, running speed, residual lost time at the upstream intersection, and the volume of vehicles turning from the traffic stream. The procedure includes models for estimating free-flow speed, running speed, and residual lost time. These models were calibrated with field data from many sites collectively located in several states. The model for estimating free-flow speed includes variables for speed limit, median type, access point density, curb presence, and number of lanes. An increase in access point density was found to correspond to a reduction in free-flow speed. The model for estimating running speed includes a variable for flow rate. The effect of flow rate on speed is similar to that found on highways, such that speed decreases with an increase in flow rate, especially at high flow rates. The procedure is recommended for inclusion i...