Real-World Rollovers Reconstructed from Interviews and Measurements

Three vehicle rollovers have been reconstructed based on driver interviews and approximate measurements of crash artifacts. The purpose of this work is to provide insight into sensor signals prior to the onset of rollover. The three events include: a car-to-car T-bone leading to a roll of 90 degrees; an impact-induced skid leading to a roll with 6 quarter turns; and a single vehicle partly on-road event with three complete rolls. Sensor signals are derived using the PC-Crash tool, and characteristic features are identified. Insights derived from these events can aid in development of robust rollover detection algorithms and calibrations. Furthermore, by studying how pre-roll maneuvers may presage a rollover event, key requirements for system-of-systems safety architectures can be derived.