Computerized Crash Reconstruction of Real World Crashes Using Optimization Methodology

Vikas Hasija (GESAC, Inc.), Erik G. Takhounts and Stephen A. Ridella (NHTSA)

CHALLENGE – Computerized automobile crash reconstruction is carried out to investigate sequences and study occupant kinematics. It involves several unknown parameters and therefore cannot be solved solely by using traditional parametric methodology. This study focuses on using an optimization scheme to find an ideal solution for reconstructing crashes within a given range of unknown parameters.

SOLUTION – Real world crashes were selected from the Crash Injury Research (CIREN) database for the study and Human-Vehicle-Environment (HVE) software was used to generate the crash pulse where Event Data Recorder (EDR) data was missing. The problem was set up in MADYMO. During the set-up, the unknown parameters were identified and treated as design variables; modeFRONTIER was then used for optimization. The objective function and constraints were defined to minimize the differences between real world data and model predictions in terms of injuries and occupant-vehicle contacts.

BENEFITS – The kinematics predicted by the optimal solution provides insight on occupant movements inside the vehicle in crash situations. By reconstructing a crash, the researchers were able to correctly predict what injuries occupants sustained and ascertain which safety features require improvement or need to be implemented. A best-scenario study using modeFRONTIER was carried out to modify vehicle design with the aim of mitigating brain injuries.