Simple EMG-driven musculoskeletal model enables consistent control performance during path tracing tasks.

Area of Science:

• Biomedical Engineering
• Neuroprosthetics
• Rehabilitation Robotics

Background:

• Neurally-controlled powered upper limb prostheses require consistent performance for user acceptance.
• Existing controllers often use complex models, limiting practical real-time application.

Purpose of the Study:

• To evaluate the performance consistency of a simplified electromyography (EMG)-driven controller based on a two degree-of-freedom musculoskeletal hand model.
• To assess the controller's practicality for real-time prosthesis control.

Main Methods:

• Computed parameters for four virtual muscles using numerical optimization from able-bodied subject kinematic and EMG data.
• Had the subject trace virtual hand paths of varying complexity, repeating straight-path tasks on a second day.
• Evaluated tracing accuracy, task duration, straightness, and smoothness, with a short EMG re-normalization period.

Main Results:

• Consistent tracing accuracy was observed across tasks of different complexity and between test days.
• Task duration, straightness, and smoothness remained consistent between test days.
• Consistent performance was achieved with a brief (~15 seconds) EMG re-calibration period.

Conclusions:

• A musculoskeletal model-based controller demonstrates consistent and effective performance, suggesting practicality for upper limb prosthesis control.
• Further research is needed to validate effectiveness in real-world tasks and applicability to individuals with neuromuscular impairments.