Evaluation of ITS Alternatives Targeting Crash Mitigation on Freeways Using Micro-simulation

This study provides an analysis of the methodology of using Crash Prediction models in simulation to evaluate different scenarios related to safety in dynamic traffic conditions. The analysis is carried out on a network calibrated according to the needs of the crash prediction algorithm. Multiple scenarios for ITS alternatives viz. ramp metering, variable speed limits, and their different combinations, are tested and a methodology developed for quantifying their safety benefits over multiple locations at different time intervals. The Crash Prediction models used have been developed based on I-4 loop data on a 36 mile (58 km) stretch in Orlando, Florida. A 9 mile (14.5 km) section has been calibrated using PARAMICS micro-simulation. Along with validation of volumes special emphasis has been placed on speeds and queues at the locations of interest in the network (Loop Detector stations). Also, since the crash prediction algorithms use data aggregated over 5 minute intervals, the simulation has been calibrated for Traffic Counts and Speeds at 5 minute intervals. Loop data was used to arrive at 5 minute means and variances of speeds at each of the locations in the field. The speeds in simulation were then compared to the actual scenarios to determine the best cases. The calibrated network was then used to test the safety benefits from using the different ITS strategies mentioned above. This meant testing the strategies implemented in a coordinated manner. An experimental design was used to arrive at the best possible scenario on a whole section of the freeway. The main conclusions of the study include: 1) The Crash Prediction Model used on the network calibrated at 5 minute intervals provided values closer to the actual Crash risk indices at these locations. As such any conclusions from the testing on simulation could be more effectively traced to field implementatio, 2) Safety Benefits at each location could be quantified and compared across locations and these measures could be used to determine an overall Safety Benefit Index between cases, and 3) The key effect of the ITS measures especially Ramp Metering and VSL was reducing the Variances in speed leading to lowered risk of crash.