The objective of this study is to evaluate route diversion as a strategy for reducing crash risk on freeways using Microsimulation. Traffic simulation environment provides the 'loop detector data' which in turn are the inputs to the models used for real-time crash risk estimation. It has been found that rear-end crashes are more accurately described as occurring within one of two distinct traffic regimes. Hence, the crash risk estimates for rear-end crashes belonging to Regime 1 and Regime 2 are output posterior probabilities from two different models which are not directly comparable. A method was proposed to transform the output from the two models into a single measure of rear-end crash risk, which can be used to assess the crash risk during simulation environment even when traffic conditions change from Regime 1 to 2 or vice versa. Using the information obtained from simulation package the crash risk estimates for the base case (No route diversion) and cases with route diversion(s) were compared. Route diversion was found to decrease the overall rear-end and lane-change crash risk on the freeway sections with free-flow conditions to low levels of congestion. However, a crash migration phenomenon was observed at higher levels of congestion.
Available at: http://works.bepress.com/apande/20/