Finite Element Musculoskeletal Model with Feedback Control to Simulate Spinal Postural Responses

Other conference contribution, 2014

Today, most Finite Element (FE) Human Body Models (HBMs) are intended for crash simulations and not for pre-crash events, due to the lack of active muscles. To study combined pre- and in crash events, muscle activity is essential. Therefore, this work presents a method to implement postural muscle responses in an FE HBM. The Total HUman Model for Safety (THUMS®) AM50 version 3.0 (Toyota Central Labs Inc, Nagakute, Japan) was chosen and a model of active musculature was added (Östh et al. 2012). The trunk, neck, upper and lower extremities were represented by 394 Hill-type line elements. Muscle activation levels were generated by seven proportional, integrative, and derivative feedback controllers for the controlled angles of the spine and upper extremities, Figure 1. For each controller, the deviation from the initial angle was used to generate correcting moment requests to the flexors and extensor muscles in the respective body region. Neural delay was implemented by a time offset for the controlled angle. The request was scaled with the maximum strength of the muscles and then passed through a muscle activation dynamics model. The model response was compared to an experimental volunteer study that measured muscle activity, kinematics, and boundary conditions for drivers and passengers, riding on rural roads in a passenger car, subjected to autonomous and driver braking. The experimental braking pulse was applied to the model seated in an FE model of the front seat and restrained with seat belts. The results show that postural feedback control can be utilized to model driver and passenger responses to autonomous braking interventions in the sagittal plane. However, the model overestimated head rotation for driver braking events. Volunteer muscle activity occurred prior to deceleration onset, which cannot be captured by the feedback control model. Therefore, a hypothesized anticipatory postural response was implemented by modifying the reference value of the feedback controllers based on the volunteer data. The result was earlier onset of muscle activity and a kinematic response that was within one standard deviation of the corresponding test data from volunteers performing maximum braking.