Abstract:
This paper investigates the influence of various roadside geometric variables and traffic volume on the running times of vehicles on urban streets. Geometric, traffic volume, and travel time data were collected from three urban street networks in Fairfax County, Virginia to support modeling efforts. Travel time data were collected on a total of 36 miles of street networks for different times of the day using GPS equipped probe vehicles. Free flow speed data were also collected on the same study segments under very low traffic volume conditions using radar gun measurements. Other data gathered included link length, traffic volume, turning movements, median type, lane width, access density, and green interval length to cycle length ratio for upstream intersections. Analysis of the data was conducted using multiple linear regression technique to determine the effect of each variable on the field measured running times of the study vehicles. Running time was found to be sensitive to free flow speed, spacing between the signals, traffic volume, median type, and turning movement percentage at downstream intersection, median types, and access density per mile of a segment. Median type was used as a surrogate factor to study the influence of vehicles making left turns at mid segment on running speed of through vehicles. Similarly, access density was used to study the influence of vehicles making right turns at mid-block on running speed of through vehicles. A comparison of default running time values from HCM and the showed HCM overestimates running time especially for shorter segment lengths.
