Simulation-Driven Exoskeleton Control: Predicting Soft Pneumatic Gel Muscle Actuator Assistance to Reduce Metabolic Cost at Different Walking Speeds.

PGM actuators with coupled control mode reduced estimated metabolic cost by 5.3-16.0% across walking speeds.

• Two PGMs were modeled at each user's leg to assist the hip joint.
• Coupled control mode uses the same control parameters for both actuators on a leg.
• Three walking speeds were evaluated in the simulations.
• The range 5.3-16.0% represents variation across speeds and actuator placements under coupled control.

PGM actuators with independent control mode reduced estimated metabolic cost by 10.5-17.5% across walking speeds.

• Independent control mode allows different control parameters for each of the two PGMs on a leg.
• Independent control mode consistently achieved higher metabolic savings than coupled control mode.
• The range 10.5-17.5% represents variation across speeds and actuator placements under independent control.
• Control parameters optimized included stiffness, onset time, and duration.

Medial actuator placement with coupled control mode offered the best trade-off between control simplicity and metabolic savings at slow walking speeds.

• Three actuator placements relative to the hip joint center were evaluated: medial, neutral, and lateral.
• Medial placement with coupled control was identified as potentially most useful for older adults and rehabilitation settings.
• Coupled control mode is simpler than independent control as it uses shared parameters for both actuators.
• This finding specifically applied to slow walking speeds.

Neutral actuator placement tended to outperform other placements across all walking speeds in terms of metabolic savings.

• Three placements (medial, neutral, lateral) relative to the user's hip joint center were compared.
• Neutral placement showed the best metabolic cost reduction performance across the range of tested walking speeds.
• This trend was observed under both coupled and independent control modes.
• The finding was based on musculoskeletal simulation predictions rather than experimental validation.

A bilevel optimization framework was implemented to identify optimal PGM control parameters for hip assistance during walking.

• The framework optimized three control parameters: stiffness, onset time, and duration.
• Two control modes were evaluated: coupled (same parameters for both PGMs) and independent (different parameters for each PGM).
• Musculoskeletal simulations were used to predict metabolic cost outcomes without physical experiments.
• The optimization was applied across three walking speeds and three actuator placements (medial, neutral, lateral).

PGMs were characterized as having a high power-to-weight ratio and compliant structure suitable for integration into smart garments.

• PGM intrinsic properties have been studied over the past decade.
• Their compliant structure enables potentially easy integration into wearable garments.
• These properties make PGM-based assistive devices promising for daily use.
• The study noted that little was previously known about how to leverage PGM dynamics to effectively and optimally assist motion.