Driving Monitoring System Considering Safety, Volatility, and Operational Performance: An Application to Highways

Abstract Understanding driving behaviour can support strategies to improve road safety. Anticipating dangerous situations and warning drivers has been a matter of extensive studies. However, there is a lack of research integrating safety, volatility in driving decisions, and vehicle operational mode in driving analysis for different road conditions. The main objective of this paper is to develop a decision support system that evaluates driver performance and provides information on safety, volatility, and operational performance, using specific variables such as time headway, stopping distance, acceleration, vehicular jerk, and vehicle-specific power mode. The proposed model is called Driving-Safety-Volatility-Operational (DSVO) indicator. To validate DSVO, a microscopic traffic model in VISSIM was calibrated using traffic and dynamics (1 Hz resolution) from probe vehicles along two highways with different traffic volumes and toll systems. Each driver’s performance was classified based on the proposed DSVO indicator. The decision on safety was supported based on a Markov Decision Process (MDP). Trajectory files were exported to the Surrogate Safety Assessment Model (SSAM) to correlate driving assessment with traffic conflicts. DSVO proved to be effective when correlated with SSAM conflict analysis mainly with the contribution of volatility failures. Vehicle Specific Power (VSP) methodology was used to calculate the VSP mode, allowing to evaluate the driver’s operating performance. The variables incorporated into the driving indicator are vehicular factors that influence accidents and pollutant emissions. Therefore, there is a potential to include this indicator in an electronic vehicle unit.