Modeling the Impact of Automation on Driver Distraction: A Queuing Theory Approach

Abstract

Driver distraction remains a critical safety challenge, and automation is often seen as a potential solution. This paper presents the application of queuing theory to model the effect of automating parts of the driving task and driving monitoring on driver performance. By framing driving and non-driving related tasks as competing demands in a queuing system, we systematically explored how automation prompting drivers to return to the driving task affects performance. Using the simmer package in R, we simulated three driving conditions: manual driving, automation-assisted driving (where the automation completes the lane keeping task), and automation with driving monitoring (where the automation completes the lane keeping task and prompts the driver to stop NDRT after 10 seconds). Each task was assigned a priority that determined what order it should be completed in and if it could interrupt the current task being performed. Performance was evaluated through task completion rates and the Performance Operating Characteristic curve. Results showed that automation improves dual-task performance when paired with driver monitoring. This work demonstrates the utility of discrete event simulation for rapidly evaluating automation interventions and highlights the need for future research using empirically grounded parameters to refine and validate these models.