Adverse Behavioral Adaptation to Adaptive Forward Collision Warning Systems: An Investigation of Primary and Secondary Task Performance.

Advanced driver assistance systems can effectively support drivers but can also induce unwanted effects in behavior. The present study investigates this adverse behavioral adaptation in adaptive Forward Collision Warning (FCW) systems. Other than conventional FCW systems that provide warnings based on static Time-To-Collision (TTC) thresholds, adaptive FCW systems consider the driver's need for support by adjusting warning thresholds according to distraction. A neglected question is how drivers adapt their behavior when they use adaptive FCW systems under realistic conditions, i.e., when warnings occur infrequently but system functionality is anticipated. Forty-eight participants drove with two different FCW systems (adaptive vs. non-adaptive) while working on a secondary in-vehicle task in a driving simulator. During the main part of the experiment, no brake events occurred and hence FCW functioning was largely anticipated. Additionally, visual system feedback about the driver's distraction state was manipulated between groups. Participants had significantly shorter minimal time-headways and TTCs when driving with the adaptive relative to the non-adaptive system. Participants with system feedback about distraction state spent generally more time with engaging in the secondary task. These results indicate behavioral adaptation which, however, is restricted to the task that is specifically supported by the system, namely longitudinal control.

Adaptive forward collision warnings: The impact of imperfect technology on behavioral adaptation, warning effectiveness and acceptance.

Adaptive ADAS that adjust warnings according to the driver´s current need for support offer a great potential to increase safety. However, it is crucial to understand how drivers deal with dynamically adapting technologies particularly in situations in which driver state monitoring fails and the system shows unexpected behavior. To better understand the consequences of unreliable adaptive ADAS on safety and to assess how failures of an adaptive FCW influence driving behavior, we conducted a driving simulator study with N = 48 participants. Participants experienced critical brake events in situations with and without a distracting secondary task. An adaptive FCW provided visual warnings to undistracted drivers but highly supportive visuo-haptic warnings (brake jerks or vibration) to distracted drivers. In 20% of brake events, however, the system unexpectedly provided incorrectly adapted warnings in which the combination of warning type and distraction was reversed. This adaptive FCW was compared to a non-adaptive standard FCW that provided visual warnings only. We found that incorrect warnings impaired driver reactions and safety in distracted drivers, and these adverse behavioral effects had two sources: (1) Violations of the drivers´ expectancies about the warning, and hence, behavioral adaptation. (2) The absence of the compensatory effect of the highly supportive warning in case of distraction. In contrast, correctly adapted warnings reduced decrements in brake reaction times and fully offset safety deficits associated with driver distraction. Crucially, however, an effectiveness evaluation of the adaptive system's potential to support drivers when correct warnings were elicited failed to demonstrate a benefit of the adaptive FCW over the non-adaptive FCW. Our results thus emphasize that a high reliability is crucial for adaptive ADAS to improve safety and to prevent adverse effects due to behavioral adaptation.