BioRID, a Crash Test Dummy for Rear Impact: A Review of Development, Validation and Evaluation

Neck injuries in rear-end car collisions is a problem on the increase. It causes severe pain and suffering to the subjected individuals and huge societal costs. The lack of a biofidelic crash dummy for rear-end impact testing has made it difficult to assess the function of the protective systems in cars, mainly the seats and head restraints. A Biofidelic Rear Impact Dummy (BioRID) was developed in several steps and validated against volunteer and cadaver test results and compared to the Hybrid III. The new dummy was partly based on the Hybrid III dummy. It had a new articulated spine with curvature and range of motion resembling that of a human being. A mathematical model (Madymo) of the head and neck was also developed and used during the development of the dummy. The BioRID was found to have more humanlike kinematics than the Hybrid III. This was found for the angular, vertical and horizontal displacement of the upper torso. It was also found for the head relative to upper torso horizontal and angular displacement. For future rear impact testing, the BioRID is an important step towards a more biofidelic dummy. INTRODUCTION Low velocity impacts causing soft tissue neck injuries are frequent and increasing both in total number and in relative frequency van Kampen (1993; Ono and Kanno, 1993; von Kock et al., 1994; Morris et al., 1996; Ono and Kaneoka, 1997; Krafft, 1998; Watanabe et al., 2000). For rear end impacts these injuries occur at low velocity changes, typically between 10-25 km/h (Hell et al., 1998; Eischberger et al., 1996). To be able to assess the protection performance and develop safety system that protects against soft tissue neck injuries, human-like crash test dummies are needed. The most commonly used dummy in crash tests is the Hybrid III. Hybrid III was developed as a biofidelic tool for high-velocity frontal-impact testing. The thoracic and lumbar spine of the Hybrid III consists of three rigid elements joined together and the neck consists of rubber block joined together by aluminium blocks. The Hybrid III is stiff and unlikely to interact with the seat back in the same complicated way as a human. ForetBurno et al. (1991) concluded that the human head can move relative to the torso with low stresses in the neck, but this is not the case for the Hybrid III. The Hybrid III has also a different contour of the back and mass distribution of the torso than a human. For rear-end collision testing the Hybrid III can supplemented by a RIDneck (Svensson and Lövsund, 1992) or by a TRID-neck (Thunnissen et al., 1996) to improve the head angular response in rear-end collision testing. However limitations still remain with the design. The Hybrid III thoracic spine that does not replicate the straightening of the spine and vertical acceleration of the head and T1 that occur in volunteer tests (Scott et al., 1993; Ono and Kaneoka, 1997; Davidsson, 1999) and limitations of the RID neck compared to human subject test has been reported by Geigl et al. (1995). There is thus a need for a complete new dummy with an articulated spine that is able to reproduce the kinematics of an occupant. The aim of this study was to develop, validate, and evaluate a new crash test dummy for rear end low velocity rear impact testing. The development of the dummy focused on the head, neck and torso kinematics. METHODS AND MATERIALS To develop a new dummy, called the BioRID, a consortium was formed. It consisted of industry partners with Autoliv AB Research, Volvo Car Corporation and Saab Automobile AB and Chalmers University of Technology. The total cost for the project was $ 1.3 million (AU), which was equally shared between the industry and the Swedish government. The project ran for 4 year and ended in 1999. When designing the BioRID following strategic decisions were made. The dummy should represent a 50 percentile male and its motion should be restricted to the sagittal plane, the dummy spine should consist of the same number of vertebrae as a human spine, the range of motion of the spine should be biofidelic and contain both a lordosis of the neck and the kyphosis of the thoracic spine, the dummy should have a human like mass distribution of the torso, the models should be simple in terms of number of components and the design of the components. For the developing of the neck of the BioRID first a mathematical model was developed to evaluate a mechanically useful combination of components for the mechanical neck model. Neither the documented model in the literature nor the commercially available mathematical models contained basic information on what components a biofidelic mechanical neck model should consist of. Therefor the need for the development of a mathematical model was identified. The mathematical neck model was implemented in MADYMO 2D (TNO, 1997), which is capable of simulating planar motion. Motion in the mathematical neck model was restricted to the sagittal plane by revolute joints connecting the vertebrae. The mathematical neck was developed using the complete dummy model presented by Jakobsson et al. (1994). Jacobsson’s model was chosen due to its basic structure and delimitations similar to those applying for the current neck model design. All elements of the mathematical neck model were transferred to physical parts in the BioRID. Validation Occupant dynamics can be obtained from volunteer and post-mortem human-subject (PMHS) tests. Both of these test types have their advantages and disadvantages. For the PMHS tests, loads can be applied that represent load levels where injuries occur. However, the lack of muscle tone, internal pressure and other changes due to time after death limit the interpretation of the results. Volunteer tests have to be carried out at load levels that will not cause injuries. In volunteer studies, muscles are active even though their influence is difficult to interpret in terms of loads. When using occupant dynamics for validation it is essential that the test conditions can be reproduced and that detailed dynamic data is available for as many parameters as possible. For these reasons two set of PMHS data (Eichberger et al., 2000: Viano et al., 2000) was chosen as validation data for the dummy. Detailed volunteer test data was used for validation of the dummy (Davidsson et al., 1998; Linder et al., 1998; Davidsson et al., 1999a, Linder et al., 1999) The mathematical model was validated against volunteer data (Davidsson, 1999). Running identical tests with the BioRID and the Hybrid III carried out the comparison between the dummies (Davisson et al., 1998, Linder et al., 1998, Davidsson et. al, 1999, Linder et al., 1999, Linder et al., 2000). Repeatability was tested by performing three identical tests with each dummy, and reproducibility was tested by running three identical dummies in the same sled test set-up as used for the validation (Davidsson et al., 1998). RESULTS The mathematical neck model (Linder, 2000) consisted of 7 identical vertebrae and the T1 vertebra. The motion of T1 was prescribed using displacement data from volunteer tests (Davidsson, 1999).