A simulation-based approach to evaluate safety impacts of increased traffic signal density.

49223106Final report.Nov. 2000Dec. 2001.PDFTech ReportVTRC 02-R7FHWA/VTRC 02-R7VirginiaVirginia Transportation Research CouncilDrummond, Kenneth PHoel, Lester AMiller, John SVirginia Transportation Research CouncilVirginia. Dept. of TransportationUnited States. Federal Highway AdministrationUS Transportation CollectionOne of the most controversial access management techniques practitioners face is also one of the most common: restricting signal density. Increased signal density can improve access for minor approaches to a corridor; however, it can also increase delays and rear-end crashes for vehicles on the mainline (major) approach. An ability to evaluate the impacts of increased signal spacing is thus critical for decision makers. Because crash data are not always easy to obtain, a logical question arises: Can simulation models be used to evaluate the safety impacts of increased traffic signal density? This report describes a method for using simulation models to evaluate the safety impacts of increased traffic signal density in suburban corridors. Using 10 years of data from two major arterials in Virginia, actual crash rates were compared with operational performance measures simulated by the Synchro/SimTraffic model. As expected, crash rates were positively correlated with stops per vehicle and delay per vehicle and negatively correlated with mainline speed. Three findings were significant. First, the correlation between crash rates and select mainline performance measures (delay, speed, and stops) was relatively strong despite the inherent variability in crash rates: R\uc2\ub2, a measure of explained variance in crash rates, yielded values from 0.54 to 0.89. Second, three distinct regimes relate stops per vehicle to signal density: the installation of the first few signals causes a drastic increase in stops, the addition of the next set of signals causes a moderate increase in stops, and the addition of a third set of signals does not significantly affect the number of stops per vehicle. Third, multiple regime models also relate delay per vehicle to signal density. This study recommends two practical applications. To the extent these mainline performance measures correlate with crashes, simulation modeling may be used to estimate safety impacts of increased signals, which is appealing because simulation packages are becoming easier to apply. Further, three regime models can suggest when, in the timeline of corridor development, the addition of a traffic signal is likely to degrade corridor performance significantly versus when it will have little effect, thereby allowing decision makers to expend political capital when it is most beneficial (e.g., the occasions when there is significant degradation of corridor performance). Most important, the approach herein suggests a long-range corridor-planning tool for evaluating the impacts of different access densities