Accident data analyses conducted at the Institute for Traffic Accident Research and Data Analysis (ITARDA) in Japan reported that over 60% of drivers who faced unavoidable crash situations made evasive maneuvers on braking and steering in 2007. In such emergency cases, drivers also might brace their body with their muscle activity to prepare the upcoming impacts. Their muscle activity would not only generate muscular forces but also change muscular stiffness and mechanical properties of their articulated joints. Therefore, occupant behaviors during impacts could be different from those observed in dummy tests and cadaver tests.

In this study, we developed an active human finite element (FE) model with 3D geometry of muscles. The muscle was modeled as a hybrid model by combination of bar elements with active muscle properties and solid elements with passive muscle properties. The bar elements were modeled with a Hill type muscle model to generate muscular force according to inputted activation levels. The solid elements were modeled with a rubber-like material model to simulate 3D geometry of individual muscles and non-linear passive properties. This combined muscle model was validated against human volunteer test data and reproduced increase of muscular stiffness with increase of muscle activation level as observed in the tests.

A volunteer test with one healthy male subject was conducted to obtain physiological information in a bracing situation with braking under his informed consent based on the Helsinki Declaration. In this test, the subject was asked to push his right foot on a brake pedal and his hands on a steering with his maximal voluntary force in the test apparatus fixed on the laboratory. Besides three reaction forces of a brake pedal, a steering, and rigid flat seats, the posture, pressure distribution on the seats, and 24 surface EMG (Electromyography) signals during his braking motion were measured in this test. His maximal braking force was reached to 750N and was well matched to previously reported values for emergency braking situation.

We performed simulations using the active human model to reproduce the bracing condition. In the simulations, the activation levels of 24 muscles obtained from the EMG data were directly inputted to the corresponding muscles of the active human model and those of the other muscles were estimated to reproduce the reaction forces. After reconstructing the reaction forces for the braced volunteer, we performed frontal impact simulations to compare occupant behavior and injury outcome in an active human body with those in a cadaveric human body. The simulation results showed significant differences between both human bodies. Different from the cadaveric human body, the active human body could have less injury risks in the thorax and more in extremities. These injury outcomes correspond to those previously reported in comparison between real-world accidents and laboratory cadaver sled tests. Although the active human model has some limitations on accuracy of estimation of muscular activation levels due to lack of consideration for muscular reflex and posture stabilization, it could have possibility to evaluate injury outcome in real-world accidents.