Driver Perceived Threat and Behavior in Rear End Collision Avoidance Situations 2015-01-1414
The focus of this paper is the threat assessment of perceived threat by drivers in collision avoidance situations. The understanding of the decision making process with regards to the initiation of a driver intervention is a crucial step to gain insight into driver's steering and braking behavior in case of an imminent threat (rear-end collision). Hence a study with various test subjects and a test vehicle has been conducted. The study has helped to understand how drivers behave in potential rear-end collision situations arising from the traffic situation (e.g. start of a traffic jam). This information is of major importance for designing autonomous collision avoidance systems and an important step towards autonomous driving. Autonomous driving in vehicles require system interventions to be initiated as early and safely as possible in order to avoid the collision and to avoid unstable vehicle dynamics situations. In parallel, collision avoidance maneuvers need to be balanced between early intervention by the system and delaying the system intervention long enough in order to ensure that the driver will not or is no longer able to intervene.
For this purpose the above mentioned study was designed to determine the braking and steering onsets when the subjects start a collision avoidance maneuver in a rear-end collision scenario either by steering or braking. The results show that the driver intervention onsets are much earlier than the physical limits for both steering as well as braking. It can be assumed that the driver would not consider any system intervention beyond these onsets as an override by the system.
Within this paper the results presented include data collected within a study conducted to determine braking and steering onsets applied by drivers in order to avoid rear-end collisions. Based on this information the driver perceived threat in potential rear-end collision scenarios using the in-vehicle steering actuator is assessed. In the end the results will be used as baseline data for the development of an autonomous steering collision avoidance system.