Effects of Training on Learning and Use of an Adaptive Cruise Control System 2020-01-1033

This study examined the effects of formalized training on driver behavior and understanding of an adaptive cruise control (ACC) system with drivers experienced with ACC. Sixteen participants drove an ACC-equipped vehicle while following a lead vehicle around a test track. Participants completed three laps, each involving different lead vehicle behaviors, such as making a lane change or stopping at a red light, that test the limitations and capabilities of ACC (i.e., boundary conditions) of the subject ACC system. Immediately before driving, half of the participants watched a training video describing how the ACC system would respond to these lead vehicle behaviors. Participants’ knowledge of the ACC system limitations was assessed by a pre- and post-test questionnaire, and participants’ interactions with the ACC system - including braking behavior, other pedal movements, and actuation of ACC via steering wheel controls - were recorded by video cameras. We did not observe differences in participants’ understanding of ACC between groups as measured by the post-test questionnaire. However, compared to the participants who did not receive training, those who did receive training unexpectedly showed significantly more use of the brake during boundary conditions, versus allowing the ACC to slow or stop the vehicle for them. Conversely, we found that while participants who did not receive training tended not to actuate the brake pedal during boundary conditions; they demonstrated more hesitations between brake and accelerator pedals. We did not observe significant differences between the two groups in time spent using ACC, though participants who did not receive training tended to re-engage ACC more often following automatic deactivation when the system reached its limits than those who received training. Results are discussed in the context of driver experience and trust in automation.