Guidelines for Evaluating Safety Using Traffic Encounters: Proactive Crash Estimation on Roadways with Conventional and Autonomous Vehicle Scenarios

Abstract

With the expected arrival of autonomous vehicles, and the ever-increasing levels of automation in today’s human driven vehicles, road safety is changing at a rapid pace. This project aimed to address the need for an efficient and rapid method of safety evaluation and countermeasure identification via traffic encounters, specifically traffic conflicts that are considered useful surrogates of crashes. Recent research-delivered methods for estimating crash frequencies based on these events were observed in the field. In this project we developed a method for observing traffic encounters with two LiDAR-based traffic monitoring units, called TScan, which were recently developed in JTRP-funded projects SPR-3831 and SPR-4102. The TScan units were deployed in the field for several hours to collect data at selected intersections. These large data sets were used to improve object detection and tracking algorithms in order to better assist in detecting traffic encounters and conflicts. Consequently, the software of the TScan trailer-based units was improved and the results generated with the upgraded system include a list of potential encounters for further analysis. We developed an engineering application for analyzing the trajectories of vehicles involved in the pre-selected encounters to identify final traffic encounters and conflicts. Another module of the engineering application visualized the traffic encounters and conflicts to inspect the spatial patterns of these events and to estimate the number of crashes for the observation period. Furthermore, a significant modeling effort resulted in a method of producing factors that expand the conflict-based crash estimates in short observation periods to an entire year. This report provides guidelines for traffic encounters and conflicts, the user manuals for setting up and operating the TScan research unit. and manuals for the engineering applications mentioned above.