Synergy-informed bilevel optimization reveals joint-specific benefits for realistic exoskeleton torque prediction

Synergy-informed simulations predict exoskeleton assistive torques more accurately than effort-minimization simulations across four single-joint configurations, avoiding non-physiological muscle shutdown and matching human-in-the-loop values more closely.

• Why it matters: Simulation-driven optimization is constrained by how well muscle coordination is predicted, because strict effort-minimization can shut down muscles under high assistance and inflate peak assistive torques beyond human-in-the-loop reports.
• What they did: A bilevel optimization estimated parameterized assistive torque to minimize metabolic cost, using assisted muscle dynamics from either effort-minimization or synergy-informed simulations constrained by a synergy structure, compared descriptively within each of five unimpaired adults.
• The result: Effort-minimization shut down agonist activations and overestimated peak assistive torque, while synergy constraints prevented this and produced torques closer to experiments, with benefits varying by degree of freedom and subject.

AI summary of the abstract · may contain errors · not medical advice · read the paper.