Five intersection crashes between two vehicles, one autonomous and one manually
driven, were investigated to identify the most severe crash scenarios for
occupants in autonomous vehicles. The SAFER human body model and models of a
semi-rigid seat and a triple-pretensioned, load-limited belt system were used to
calculate occupant kinematics and injury metrics for the five scenarios. Both
driver and passenger positions and upright and reclined postures were
investigated. The kinematics and injury metrics were compared to those of
traditional full-frontal crashes at 40 km/h and 56 km/h, with the purpose of
complementing current frontal safety assessments with additional scenarios for
assessing automated vehicle safety. The intersection crash with a vehicle
turning left across the path of a vehicle arriving from the opposite direction
(“left turn across path opposite direction” with velocities of 19 km/h and 66
km/h, respectively) was found to be the most severe of the five in terms of
loading on the head, neck, pelvis, and lumbar spine. In fact, in the upright
posture the lumbar spine loading was equal to that in the full-frontal 56 km/h
crash. A second intersection crash, with the vehicle struck perpendicularly by
another vehicle (“straight crossing path” with velocities of 63 km/h and 58
km/h, respectively) was found to be the most severe in terms of loading on the
chest and the risk of slipping out of the belt due to oblique, far-side occupant
kinematics. Notably, the full-frontal 56 km/h crash did not reproduce this
kinematics. In all other respects, all of the intersection crashes were less
severe than the full-frontal 56 km/h crash. The “left turn across path opposite
direction” and “straight crossing path” crashes were identified as low velocity
crash scenarios to be considered in the development of countermeasures and in
future legislation and rating programs together with more stringent requirements
in injury assessments.
