Cervical spine injuries in rear impact collisions are very common and result in enormous societal cost, with estimates on the order of $8.0 billion annually (NHTSA, 2004). In order to reduce the possibility of cervical spine injuries, car manufacturers have developed and improved seats with restraint systems by using anthropomorphic test devices (ATDs) as safety testing tools. Rear impact ATDs should accurately reproduce the kinematic and kinetic responses, known as biofidelity of the head/neck, spine and pelvis of humans. Despite the fact that rear impact events have been studied extensively, there is no consensus as to the exact mechanism of injury or most relevant injury criterion. Therefore, the objective of this study is to evaluate the biofidelity of rear impact ATDs by comparing their responses with post mortem human subjects (PMHS) in realistic testing conditions and to investigate correlation between cervical spine injuries and head-neck kinematics/kinetics of PMHS in rear impact conditions.
An experimental seat was designed to measure external loads on the head restraint (4 load cells), seat back (6 load cells), and seat pan (4 load cells) such that occupant dynamic interaction with the seat could be evaluated. A method for obtaining accurate 6 degrees of freedom (DOF) head kinematics in rear impact conditions is proposed and validated. A new instrumentation and dissection technique was proposed and validated in which instrumentation (three accelerometers, three angular rate sensors) capable of measuring the detailed intervertebral kinematics are installed on the anterior aspect of each vertebral body with no muscular damage. A total of seventeen sled tests using eight male PMHS (175± 6.9 cm of stature and 78.4 ± 7.7 kg of weight) and a total of ten sled tests using rear impact ATDs (BioRID II, RID3D, and Hybrid III) have been conducted in both low (8.5g and 17km/h) and moderate (10.5g and 24km/h) speed rear impacts.
Biofidelity of the rear impact ATDs was assessed using the NHTSA Biofidelity Ranking system, which has been improved by compensating phase differences between the PMHS and the ATDs. Overall, both the BioRID II and RID3D were more biofidelic than the Hybrid III in both low and moderate speeds. Injury analyses were performed to find the best predictable physical parameters. Based on the results from the injury analysis, the intervertebral rotation exhibited best statistical significance, correlation with the injuries (0.68 of Pseudo R² and 0.80 of Nagelkerke R²), goodness of fit of the regression model (-11.09 of log-likelihood with 0.00 of p-value), and predictive ability of the injuries (0.95 of Goodman-Kruskal Gamma). The current neck injury criteria were assessed and IV-NIC was selected as a best predictor of the injuries. The NDC for the rotation (NDCr) also showed good correlation with the intervertebral rotation. Two possible physical parameters for the potential neck injury criteria, OC-to-T1 linear velocity in the shear direction and OC-to-T1 angular velocity, were selected based on the correlation analyses.