Optimization of the driver distraction parameter calibration in PTV VISSIM microsimulation platform using visual scanning patterns

Microscopic traffic simulation models are essential tools for evaluating and optimizing various traffic management and control systems. These tools model driving behavior and include parameters that require calibration before utilizing a simulation platform. However, the critical factor of driver distraction, which significantly affects hazard perception, has been somewhat overlooked in prior microsimulation studies like the PTV VISSIM platform. The VISSIM driver distraction parameter comprises three key components: the probability of distraction, distraction duration distribution, and the distribution of the drivers' lateral deviation. The study aimed to calibrate each of these three factors using vehicle performance data.  In this study, the distraction source was the use of cellphones for reading or writing text messages, tested on 30 drivers through a driving simulator. Messages were sent to them, and drivers had to respond to messages while maintaining awareness of road conditions. The results indicate that distraction duration follows a lognormal distribution with an average duration of 1.24 seconds and a standard deviation of 0.87. Additionally, the drivers' lateral deviation exhibits a Dagum distribution with an average deviation of 2.2 degrees and a standard deviation of 2.74. This research enhances the accuracy of microsimulation models by integrating visual scanning patterns, which refer to the driver’s eye movements and visual attention shifts while operating a vehicle, and carefully calibrating the driver distraction parameter using statistical normal distribution. It provides a thorough understanding of driver behavior and significantly improves the models' ability to predict accidents.