Comparison of vehicle travel times and measurement techniques along the I-35 corridor in San Antonio, Texas

Abstract

The floating car method of travel time data collection has been used since the 1920's to evaluate transportation issues such as congestion management and freeway operation. However, this technique can be time consuming, labor intensive, and expensive if a large sample size is required. The objective of this research was to evaluate two methods to estimate freeway travel times using loop detector data and to compare these results to the floating car technique. Floating car travel time data were collected for three days and compared to travel times derived from loop detector data along the 1-35 corridor in San Antonio, Texas. The loop detector data were made available through TransGuide's Internet site. TransGuide is the traffic management center of San Antonio whose main tasks are to provide incident detection, traffic control changes, support transit dispatch operations and system reliability on the local freeway network by using technologies such as surveillance cameras and variable message signs. Both corridor and link travel times were studied, as well as possible factors that might affect travel times such as path, time of day, length of link, and average number of lanes per link. These travel times were analyzed using simple and multiple statistical linear regression tests in addition to utilizing other statistical tests. The results indicated that the two extrapolation methods studied produced comparable results in free-flow conditions to the floating car technique, and both methods are recommended to calculate travel times. However, due to the small sample taken during peak traffic conditions, it was not determined completely how comparable the loop detector results were to the floating vehicle results at the time period transportation officials are concerned with the most. Furthermore, in actual conditions, one would not have prior extrapolation methods, as was the case in this study, due to the dependence on the bloating vehicle data for providing this important information. Thus, additional research recommended in this area includes the development of a mechanism to estimate the location of vehicles along roadway that could be supplemented by other data collection techniques, a more extensive data collection process during the peak period, and an investigation of the variability of travel times determined using alternative techniques like computerized license matching and GPS. Finally, the relation of link length to travel time was examined in detail, and it was found that travel time variation becomes more prevalent as the lengths of links become longer especially when longer than 0.5 mile (0.8 km). The measured speed differences between extrapolation methods and floating car were not significant, and no general trend was discovered other than that the difference was minimal.