Smart Red Clearance Extensions to Reduce Red-Light Running Crashes.

摘要

Red-light running (RLR) is a safety hazard at signalized intersections in Oregon and around the United States. RLR crashes occurring during phase transitions at signalized intersections can be addressed through the use of Red Light Extension (RLE) systems. This research project evaluated the accuracy, efficiency, and operations of three RLE system alternatives in Oregon. A field study was conducted at five intersections in Oregon with high rates of RLR vehicles. To understand driver behavior in response to the onset of the yellow change interval (on major approaches) or the green indication (on minor approaches), 149 h of video data were transcribed. Together with signal timings, geometries, peak volumes, and operating speeds, the video data were used to calibrate a VISSIM hardware-in-the-loop (HIL) simulation of one intersection. The HIL simulation was used to code and evaluate the performances of three alternative RLE systems. Novel code was developed in R to process the VISSIM output into time-space diagrams (TSDs) for further analysis. RLR vehicles triggered the highest rate of RLE events when the downstream detection alternative was used. However, an examination of the TSDs showed that the RLE events triggered by the smart upstream speed-conditional detection alternatives (at 215 ft. & 475 ft.) consistently provided a greater degree of safety for vehicles that received a RLE while no significant increase to intersection delay occurred. Use of the 2070 controller and the NWS VoyageTM software allows the logic to be field implemented in the future.