Simulating Human Upper and Lower Limb Balance Recovery Responses Using Nonlinear Model Predictive Control.

Area of Science:

• Biomechanics
• Robotics
• Human Movement Simulation

Background:

• Predictive dynamic simulations of human movement are crucial for designing assistive devices like robotic exoskeletons.
• Existing simulations often overlook upper body contributions to whole-body tasks such as balance recovery, focusing instead on lower body joints.

Purpose of the Study:

• To investigate how actuated upper limbs and varying performance (optimality) criteria influence simulated reactive balance recovery using a novel nonlinear model predictive control (NMPC) approach.
• To explore the emergence of human postural control strategies from individual optimality criteria in simulations.

Main Methods:

• Developed a sagittal biomechanical model of a standing young adult with nonlinear muscle torque generators.
• Generated forward dynamic simulations of reactive balance recovery using NMPC following support-surface perturbations.
• Systematically varied performance criteria (6 total), perturbation direction (forward/backward), and arm joint status (free/locked).

Main Results:

• Simulated joint trajectories demonstrated how specific optimality criteria can lead to the emergence of human-like postural control strategies (e.g., hip-ankle strategies).
• Quantitative analysis indicated potential performance improvements when arm joints were free.
• The emergence of arm responses in simulations may depend on the initial guess of the NMPC problem.

Conclusions:

• Nonlinear model predictive control (NMPC) can simulate reactive balance recovery, highlighting the role of upper limbs and optimality criteria.
• Individual performance criteria can shape simulated postural control, offering insights into human movement.
• Further research should explore additional performance criteria and refine NMPC as a model for the nervous system in movement control.