A task-space oriented approach for the reproduction of sit-to-stand motion using operational space control

To this day, biomechanics actively research how the nervous system selects particular motion strategies in order to achieve a desired motion. The human sit-to-stand transfer, as one of the most important tasks influencing the quality of life, has been studied extensively through experimental studies. Simulative studies, however, are relatively rare, although anticipating the sit-to-stand movement accurately is essential for many applications, e.g. treatment planning or the design of exoskeletons.

Existing research in sit-to-stand motion prediction has mainly focused on developing direct dynamics strategies in joint space. This thesis addresses the application of a control method in operational space through assigning a desired end-effector motion, i.e. the head trajectory, to a biomechanical four-link skeletal model in the sagittal plane in order to predict healthy sit-to-stand motion. The dynamics of the quadruple pendulum is enhanced by a new Gaussian function model to replicate the contact forces between chair and buttocks during sitting.

Since infinitely many possible configurations exist to perform the desired motion, the mechanism is called redundant with regard to its task. The proposed operational space control scheme is adapted from the field of robotics and resolves those redundancies at kinematic level using a general projector matrix in the null space of the task Jacobian. By this means, the task control is complemented by human posture regulation, which, in this case, considers the orientation of the foot related link. Besides controlling human posture, this method provides freedom to the choice of a suitable pseudo-inverse of the Jacobian matrix that affects performance preferences. With the purpose of finding the fundamental objective that underlies the sit-to-stand motion on a neural level, different pseudo-inverses are tested within posture control for two simulation approaches considering the task command variables. The first approach controls the two dimensional head position as well as the orientation of the upper body. The second approach only regards the vertical position and the orientation as the task.

Simulation results demonstrate physiologically accurate performance predictions in
comparison to reference data based on performed sit-to-stand measurements. The
method also emerges as being superior compared to the application of a recently published optimization-based control approach in terms of computation time, efficiency and overall generality.



Citation style:
Could not load citation form.


Use and reproduction:
All rights reserved