Objective: A driving simulator study was conducted to evaluate the longitudinal effects of an intervention and withdrawal of a lane keeping system on driving performance and cognitive workload.
Background: Autonomous vehicle systems are being implemented into the vehicle fleet. However, limited research exists in understanding the carryover effects of long-term exposure.
Methods: Forty-eight participants (30 treatment, 18 control) completed eight drives across three separate days in a driving simulator. The treatment group had an intervention and withdrawal of a lane keeping system. Changes in driving performance (standard deviation of lateral position [SDLP] and mean time to collision [TTC]) and cognitive workload (response time and miss rate to a detection response task) were modeled using mixed effects linear and negative binomial regression.
Results: Drivers exposed to the lane keeping system had an increase in SDLP after the system was withdrawn relative to their baseline. Drivers with lane keeping had decreased mean TTC during and after system withdrawal compared with manual drivers. There was an increase in cognitive workload when the lane keeping system was withdrawn relative to when the system was engaged.
Conclusion: Behavioral adaptations in driving performance and cognitive workload were present during automation and persisted after the automation was withdrawn.
Application: The findings of this research emphasize the importance to consider the effects of skill atrophy and misplaced trust due to semi-autonomous vehicle systems. Designers and policymakers can utilize this for system alerts and training.
Keywords: driver behavior; human-automation interaction; situation awareness; trust in automation; vehicle automation.