Empirical models for estimating automobiles running speeds on urban streets

This research study determines the factors affecting running speed of automobiles on urban streets. The Highway Capacity Manual (HCM), the most widely used tool in transportation and traffic engineering practice in the United States, evaluates the performance of urban streets in terms of average travel speed of through vehicles. The HCM provides default values for running time as a function of street classification, free-flow speed, and segment length. However, the HCM methods does not take into consideration the influence of traffic volume and turning vehicles on running speeds of through vehicles.; This research examines the factors influencing running time or running speed of automobiles on urban streets and develops quantitative models to estimate running time as a function of various geometric variables and traffic flow conditions. The research further determines the influence of different geometric factors on free-flow speed of vehicles on urban streets.; Field data were collected for roadside geometric variables, automobiles free-flow speed, and running time under varying traffic flow conditions. Free-flow speed data of 1742 passenger cars and 1300 miles of travel time data were collected from three different urban street networks in northern Virginia. The analysis of free-flow speed data revealed posted speed to be the most significant factor affecting free-flow speed of vehicles on urban streets. Other significant factors influencing free-flow speed include spacing between the signals and presence and absence of medians.; Three statistical models - linear regression model, regression tree model, and generalized linear models (GLM) - were developed relating running time with roadside geometric and traffic flow variables. The findings from data analysis showed free-flow speed, segment length, and traffic flow rate as the most significant variables affecting running speed of vehicles. Free-flow speed and segment length were found to be positively related with running speed. This means that higher the free-flow speed and larger the spacing between the signals, higher would be the running speeds of vehicles. Traffic flow rate, on the other hand, was determined to have an inverse relationship with running speed. The regression model determined 0.1 percent increase in running time for one percent increase in traffic flow rate. Other factors found to be affecting running speed were percentage of turning vehicles at signalized intersections, median type, such as median separated or two way left turn lane, or no median, and number of driveways and access density per unit length of segment. Median type and access density were used as surrogate factors to study the influence of left and right turning vehicles on mid-block speeds of through vehicles.; The values of running time from the linear regression model were compared with the default values of running time given in HCM. It is determined that HCM overestimates running time especially, for smaller segments. Also, the decline in running time from smaller to larger segments from the model is not as steep as the decline in HCM values. The prediction models developed in this study for free-flow speed and running time of vehicles on urban streets could be effectively used in the absence of field data.