Muscle Activation Strategies in Human Body Models for the Development of Integrated Safety

Other conference contribution, 2015

Human Body Models (HBMs) have been used in crash safety research for some time, and are now emerging as tools for the development of restraints systems. One important challenge in the development of advanced restraint systems is to integrate sensory information about the pre-crash phase (time to collision, impact speed and direction, occupant position) to alter restraint activation parameters. Restraint activation can begin even before the beginning of an impact, providing additional time to reposition or restrain the occupant. However, any such pre-crash intervention would invoke a muscle response that needs to be taken into account in HBMs used in simulation of integrated restraints. The objective of this paper is to provide an update on state-of-the-art modeling techniques for active musculature in HBMs. Examples of applications are presented, to illustrate future challenges in modeling of car occupants muscle responses to restraint activation. The most common approach for modeling active muscle force in HBMs is to use Hill-type models, in which the force produced is a function of muscle length, shortening velocity, and activation level. Active musculature was first implemented in cervical spine models. These models were applied to study occupant kinematic responses and injury outcome in rear-end, lateral, and frontal impacts; it was found that active musculature is essential for studying the response of the cervical spine. One approach utilized to represent muscle activity in HBMs is to use experimentally recorded muscle activities or activity levels acquired through inverse optimization in open-loop. More recently, in order to represent car occupant muscle responses in pre-crash situations, closed-loop control has been implemented for multibody and finite element HBMs, allowing the models to maintain their posture and simulate reflexive responses. Studies with these models showed that in addition to feedback control, anticipatory postural responses needs to be included to represent driver actions such as voluntary braking. Current HBMs have the capacity to model (utilizing closed-loop control) active muscle responses of car occupants in longitudinal pre-crash events. However, models have only been validated for limited sets of data since as high quality volunteer data, although it exists, is scarce. Omni-directional muscle responses have been implemented to some extent, but biofidelity of the simulated muscle activation schemes has not been assessed. Additional experimental volunteer muscle activity measurements (with normalized electromyogram recordings) in complex 3D-loading scenarios are needed for validation and to investigate how muscle recruitment depends on occupant awareness and varies between individuals. Further model development and validation of muscle activations schemes are necessary, for instance startle responses, and individual muscle control. This could improve assessment of restraint performance in complex accident scenarios, such as multiple impacts, far-side impacts and roll-over situations.