Paul St-Aubin is senior product manager for Transoft Solutions (ITS) and a post-doctoral fellow at the University of Waterloo. He has over a decade of experience practicing road safety consulting and has been at the forefront of research and development of proactive road safety monitoring and diagnostics, and collision modeling and prediction technologies. His areas of expertise lie in safe road design, collision-course modelling, driver behaviour, traffic data collection technologies, computer vision, and ITS architecture. Dr. St-Aubin has taught traffic engineering courses at McGill University and the University of Montreal, and continues to teach proactive road safety monitoring technologies to this day.
Title of Abstract
Measuring Road Safety Improvements Using Video-Based Conflict Analysis in Prince George, BC
Remedial measures were implemented at the signalized intersection of Highway 97 and 22nd Ave, in Prince George, BC. These measures consisted of converting a permissive left-turn phase to a protected-only eastbound left-turn phase, installing pedestrian countdown timers, and extending the pedestrian phases. Prince George wanted to evaluate the effectiveness of these countermeasures immediately, without having to wait for crash data. New video-based surrogate safety techniques were used to provide a better understanding of pedestrian crossing behavior, vehicle conflict patterns, turning behaviours, speed profiles, as well as road user volumes before vs. after the remedial measures were put in place.
The goal of this study was to assess changes in the level of safety for drivers and pedestrians at the intersection before vs. after the remedial measures were completed.
1. The City installed one camera on the northeast corner of the intersection to capture traffic movements for a total of 72 hours during the before period: Sept 25 to Sept 27, 2018.
2. Remedial measures were completed January 2019.
3. The City installed one camera on the northeast corner of the intersection to capture traffic movements for a total of 67 hours during the after period: Sept 10 to Sept 17, 2019.
4. Brisk Synergies calibrated and processed the video from both collection periods to provide several standard surrogate safety outputs, which were accessible on an online Safety Dashboard.
5. True North Safety reviewed the outputs and prepared a report summarizing the findings.
-Conflicts between eastbound left-turn and westbound through traffic were almost eliminated.
-In the before period, approximately 6% of the eastbound left turning vehicles were involved in conflicts with PET less than 3 seconds. This proportion decreased to 0.04% in the after period.
-The conflicts with a PET of less than 3 seconds at the north crosswalk were eliminated by the changes made to the intersection traffic signal plan.
-The speed at the intersection decreased in the after period by approximately 11%.
Overall, the implemented remedial measures were very effective at minimizing conflict risk at the intersection. The City plans to compare the before/after collisions statistics once they are available to determine if the reduction in conflicts resulted in actual collision reductions.
Traditional methods of before and after studies use collisions as the primary safety performance measure. However, collisions are relatively rare events and a waiting period of three to five years is typically needed to obtain a reliable sample size. With recent developments in video analytics, surrogate safety measures obtained from video cameras can be used to evaluate the effectiveness of a countermeasure without having to wait years for collision data. Prince George was able to show significant improvements to safety, measured by reductions in high-risk interactions and speed reductions.