Investigating the Impact of the Microscopic Traffic Simulation Model Calibration on the Surrogate Safety Assessment

This study investigated the impact of calibration on safety assessments using conventional microscopic traffic simulation model calibration methodologies focusing on mobility measures. The Broward Boulevard VISSIM model was calibrated based on three different calibration levels: 1) no calibration, 2) calibration with link volumes (i.e., volume-based calibration), and 3) calibration with volume, travel time, and speed (i.e., full calibration), and traffic conflicts were estimated using these three simulation models. The estimated traffic conflict results were analyzed through its correlation strength with actual traffic crashes collected at a given area to investigate whether the calibration was effective in evaluating surrogate safety. In addition, these three simulation models were tested with optimized signal timing plans to check whether or not the calibration levels affect the surrogate safety assessment results (i.e., reduction in traffic conflicts). The correlation analysis results showed that the conflict estimation results from the ‘volume-based calibration’ scenario had stronger correlation with traffic crashes than the results from the ‘full calibration’ scenario, and the results from the ‘no calibration’ scenario had significantly low correlation. In addition, while the safety benefits of the optimized timing plans were not statistically significant in the ‘no calibration’ scenario, it was significant in the calibrated simulation models (i.e., ‘volume-based calibration’ and ‘full calibration’ scenarios). Based on the findings from this study, it is recommended that a microscopic traffic simulation model be calibrated at least with the link volumes for conducting surrogate safety studies.      «

This study investigated the impact of calibration on safety assessments using conventional microscopic traffic simulation model calibration methodologies focusing on mobility measures. The Broward Boulevard VISSIM model was calibrated based on three different calibration levels: 1) no calibration, 2) calibration with link volumes (i.e., volume-based calibration), and 3) calibration with volume, travel time, and speed (i.e., full calibration), and traffic conflicts were estimated using these thre...     »