Abstract
Accurate occupant injury prediction is important for Advance Automatic Collision Notification in accidents. In previous report, it was found that vehicle body deformation information was needed to improve injury prediction. This paper studied vehicle body deformation factor for occupant injury prediction in side impact crashes using an accident database in the U.S.(NASS-CDS) and constructed injury prediction model using those factors. As a result, roof deformation factor is found to be important for improving injury prediction model. The analytical method using photos of deformed vehicles was considered to measure the roof deformation in NASS-CDS accident cases. Roof deformations with 30 cm and above have a higher odds ratio of 2.944 compared to those below 30 cm. Therefore, this factor is considered to be included in injury prediction model with sensitivity improved to approximately 88% from the conventional model.
