Safety Assessment Model for Multilane Freeways using Loop Detector Traffic Flow Data

摘要

With the rapid economic development around the world, a number of four-lane or six-lane freeways were widened to satisfy the increasing traffic demand. The primary objective of this study is to develop a safety assessment model for multilane freeways using high-resolution traffic flow data collected from loop detectors. Data were collected from the 1-880 freeway in the state of California, United States. A binary logit model was used to develop the relationship between crash likelihood and real-time traffic flow variables. The model estimation results indicate that the real-time traffic flow variables significantly affect crash risks on multilane freeways. Model evaluation results show that the model has the potential to be used for safety assessments of multilane freeways. The overall prediction accuracy of the model is 65.6%, indicating a good prediction performance for evaluating the safety performance of multilane freeways.