The reduction of road casualties and injury risk is a major goal of automotive engineering. Manufacturers are confronted with increasing safety regulations, stricter exhaust gas pollution regulations and strong competition. Hence, cost efficiency is a significant concern of automotive industry. Furthermore the development of innovative safety systems such as adaptive restraints and collision avoidance systems calls for new methods for system design and evaluation beyond laboratory crash tests.
Cost efficiency is achieved with numerical simulations using Multibody System (MBS) or Finite Element Method (FEM) techniques partly substituting full vehicle testing. Focus has been on simulations of the crash phase but innovative safety systems call for consideration of the low g precrash phase. Therefore the use of crash test dummies and their numerical representations which are designed for higher loading, is not satisfying. The use of numerical human body models is a promising approach to further improve bio-fidelity. Nevertheless, for pre-crash simulations the influence of muscle activity on the passenger kinematics induced by the vehicle motion is no longer negligible for real life safety.
Hence the OM4IS (“Occupant Model for Integrated Safety”) project was initiated by a large consortium including scientific partners (Virtual Vehicle Research and Test Center, Graz University of Technology, Bundesanstalt für Straßenwesen- BASt) and industry partners (Partnership for Dummy Technology and Biomechanis, Robert Bosch GmbH, Toyoda Gosei Europe, TRW Automotive, DYNAmore GmbH). The challenge is to identify human movement and behavior patterns (position and muscle activity) during pre-crash phase and implement these patterns into a suitable human body model. The present paper describes first results to implement muscle activity into a simplified version of the numerical model Total HUman Model for Safety (THUMS) developed by Toyota Motor Corporation and Toyota Central R&D Labs. This model represents a 50th percentile American Male (AM50) and is implemented into the explicit finite element software LS-Dyna.
As a starting point, the reactive behavior of humans in two distinct load cases, an emergency braking maneuver and a single lane change are investigated. Movement and behavior patterns as well as muscle activity are analyzed by volunteer tests on sled and full vehicle level. An infrared based 3D motion capturing system and an electromyography measurement (EMG) system are used. Methodology and results of this behavior pattern analysis is presented in a separate paper.
A simplified FE model that qualitatively reproduces human motion patterns in the selected load cases is developed. The first version of the model features a simplification of the THUMS model replacing the deformable parts by rigid body parts and using kinematic joints. Major muscle groups are implemented as beam elements which can be controlled using coupling of LS-Dyna software and Matlab/Simulink. The model should be able to reproduce volunteers’ movements for two load cases (acceleration in frontal and lateral direction) and in the second modeling step identified human movement and behavior pattern should be implemented qualitatively which is presented in a separate publication.
At this stage computing time efficiency, numerical stability and implementation in the automotive development process were not of first priority. Furthermore the study concentrates on occupants’ acceleration induced reactions and not on active movements.