A study of muscle control with two feedback controls for posture and reaction force for more accurate prediction of occupant kinematics in low-speed frontal impacts

In future automotive crashes that involve advanced safety vehicles or autonomous vehicles, the number of minor or moderate injuries may increase because of vehicle slowing by safety systems such as autonomous emergency brakes. Recent studies suggest that pre-crash muscle activity of occupants could have significant effects on the kinematics of occupants in such situations. In previous studies, we developed a human body finite element (FE) model with whole-body muscles and a muscle controller with posture control to predict relaxed occupant kinematics during deceleration. However, the controller could not predict tensed-occupant kinematics. The objective of this study is to develop a muscle controller for more accurate prediction of relaxedand tensed-occupant kinematics and to validate it in low-speed frontal crash situations. Total HUman Model for Safety (THUMS) version 5, including 262 one-dimensional Hill-type muscle models, is used and a new muscle controller using proportional-integral-derivative (PID) control is developed. The controller has two feedback controls involving three-dimensional angles of 16 joints and reaction forces using a steering wheel and pedals. The control of each joint angle works to return to the initial joint angle in order to maintain overall body posture. The control of each reaction force works to achieve a pre-determined target force. The controller is validated using a series of experimental data from cadaver and volunteer tests reproducing low-speed frontal impacts with peak sled accelerations of 2.5 G and 5.0 G, which are obtained from the literature. Simulation results demonstrate that head excursions predicted without any control and when using only posture control are similar to excursions from cadaver and relaxed volunteer test data, respectively. Head excursions predicted by a total controller with the two feedback controls of posture and force show a tendency similar to that for tensed volunteers. The forces predicted by the total controller are similar to those for the tensed volunteer test data for the pedals but not for the steering wheel. Further studies on optimization methods are needed in order to determine valid PID gain parameters in various dynamic environments. THUMS using the developed controller shows the potential for representation of both relaxedand tensed-occupant kinematics during low-speed impact decelerations.