Accounting for human-related unobserved heterogeneity in the safety performance of connected vehicles: An incorporation of Bayesian hierarchical negative binomial into simulated work zone warning application

Connected Vehicle (CV) notifications enhance drivers' situational awareness, where, depending on traffic flow characteristics, CV driversmight require sufficient distance and time to adapt their driving behavior accordingly. Hereof, in addition to CVs Market Penetration Rate (MPR) and traffic flow characteristics, humans are still in the safety loop. Hence, there might be a human-related unobserved heterogeneity affecting the safety performance of CVs. A conflation of microsimulation and advanced statistical modeling was followed to account for this unobserved heterogeneity by investigating the random effect of varying delivery distances, specifically provided for the Work Zone Warning (WZW) application. CVs behaviors were calibrated based on the results of driving simulator experiments. A typical WZ was simulated under VISSIM using three levels of low, moderate, and high traffic volumes. Five levels of CV-MPR received the WZW on four different settings of delivery distances, including 1500, 3000, 4500, and 6000 ft, upstream of the Transition Area. Accordingly, 60 scenarios were simulated, and the number of critical Time-to-Collision (TTC) was considered as the response count variable per scenario. Four hierarchical Negative Binomial regressions under Bayesian inference were conducted on 60 observations over four WZ areas, where the delivery distance was treated as a random factor. Analysis of the Intraclass Correlation Coefficient showed that the number of observed critical TTCs is nested within the delivery distances. It was ascertained, at any CV-MPR, when the WZW delivery distance increases, the number of critical TTCs decreases that indicates the same safety performance at higher CV-MPR can be achieved even at lower MPR if the WZW distancing increases. Besides, the optimal Road Side Unit (RSU) placement for WZW was found to be 4500 ft upstream of the Transition Area regardless of traffic volume and CV-MPR. Although this study was restricted to the CV-WZW application, the introduced delivery distance for CV notifications should be regarded in future studies as a contributing factor to the safety assessment of other CV applications. (C) 2021 International Association of Traffic and Safety Sciences. Production and hosting by Elsevier Ltd.