Subject-Exoskeleton Contact Model Calibration Leads to Accurate Interaction Force Predictions.

Area of Science:

• Biomechanics
• Robotics
• Human-Computer Interaction

Background:

• Understanding human-exoskeleton interaction forces is vital for optimizing exoskeleton design.
• Computational modeling offers a method for in silico prediction of these forces.

Purpose of the Study:

• To develop and validate an optimal control framework for predicting human-exoskeleton collaborative movement and interaction forces.
• To assess the framework's accuracy in reproducing experimental kinematics and forces during sit-to-stand tasks.

Main Methods:

• An optimal control framework with three phases was developed.
• Phase A and B involved calibrating foot-ground and subject-exoskeleton contact models using experimental sit-to-stand trials.
• Phase C predicted collaborative movements and interaction forces for six sit-to-stand trials.

Main Results:

• Calibrated contact models accurately reproduced experimental kinematics (RMSD ≤ 1.1° for coordinates, ≤ 6.8°/s for velocities) and forces (RMSD ≤ 22.9 N for ground reaction, ≤ 5.4 N for interaction).
• The framework successfully predicted collaborative movements (RMSD ≤ 3.5° for coordinates, ≤ 15.0°/s for velocities) and interaction forces (RMSD ≤ 13.1 N) in prediction trials.

Conclusions:

• The developed optimal control framework provides reliable in silico predictions of human-exoskeleton interaction forces.
• This framework can be a valuable tool in the design phase of exoskeletons to explore optimal movements and interaction dynamics.