HUMAN BODY MODELING FOR APPLIED TRAFFIC SAFETY

Övrigt konferensbidrag, 2011

Traffic injuries are an important public health issue. To prevent, diagnose and treat injuries it is vital to understand the mechanics of injuries. Here, mathematical models of the human present a valuable complement to other models, such as animal models and crash dummies. Today, Human Body Models (HBM) are recognized as important tools within traffic safety research. To successfully apply an HBM to improve and evaluate real life safety systems, it has to: (1) be numerically robust in a wide range of crash loading conditions, (2) be computationally efficient to enable analyses with full car models, (3) represent the human population with respect to age, gender and anthropometry, (4) maintain its posture in a gravitational field for pre-crash events, (5) predict the onset of tissue injury and organ failure, and (6) simulate muscle tension due to bracing and muscle reflexes. Therefore, work is ongoing to model the active muscle response and improve the injury predictability of currently available FE HBM. The commercially available HBM Total HUman Model for Safety [1], called THUMS, was used with the explicit capabilities in the FE code LS-DYNA [2]. It is a model of a 50th percentile adult male vehicle occupant and contains approximately 150,000 elements. To study thoracic injuries, the responses of the THUMS were compared to several cadaver experiments. Then, a sensitivity study was performed to evaluate the influence of belt interaction and tissue parameters on the predicted thoracic response. Lastly, several candidates to predict rib cage fractures were compared in loading conditions relevant to frontal car crashes. The central nervous system controls the muscle contraction and was modeled using feedback proportional, integral, and derivative (PID) control. The reference signal is a joint angle defining a body position. The neural delays, due to the time needed for the nerve signals to travel back and forth to the central nervous system, and muscle activation dynamics are included. Firstly, this was applied to evaluate the response of the elbow joint compared to volunteer experiments [3], and secondly, to compare passenger kinematics in autonomous braking events. It was seen that by changing the controller gains, the model can can capture differences in the muscle response when the human is relaxed compared to tensed, which is important to study the difference between occupants who are or who are not aware of an oncoming accident.