Bioinspired Wearable Robotic Control

We directly measure an individuals muscle mechanics in various walking tasks and feed back that information into the muscle-based exosuit assistance.

We applied active impedance control to emulate passive-elastic exoskeletons.  Active control of impedance (stiffness) mitigated the effect of deformation in the mechanical exoskeleton structure and allowed for rapid sweep stiffness parameters.

Assistance profile shape is parameterized on time (as a percentage of an average stride; e.g. onset time) and magnitude (e.g. peak applied force). Typically, these are parameters are derived, as a starting point, from joint moment and power.

Using a simple neuromuscular model, we estimated the contribution to biological ankle torque from reflexive force-feedback (i.e. Golgi organ).  Exoskeleton assistance magnitude and profile shape changed as the person's estimated contributions changed.

The exoskeleton assistance profile was derived from the wearer's muscle activation profile and scaled to account for walking speed.