PRISM: Parameter Recovery Identification from Synthetic Modeling.

Area of Science:

• Biomechanics
• Robotics
• Motor Control

Background:

• Human motor learning dynamics are complex and challenging to model precisely.
• Existing methods in robotics and system identification have limitations in capturing human dynamics.
• Persistent excitation is crucial for accurate modeling but underexplored in human motor control.

Purpose of the Study:

• To develop and test a forward-inverse method for estimating human motor control parameters.
• To identify optimal experimental conditions for accurate parameter estimation.
• To establish a framework for individualized motor control modeling.

Main Methods:

• Simulated planar arm movements using a synthetic model.
• Applied a forward-inverse method to estimate joint stiffness, damping, and feedback gains.
• Utilized regression analysis to identify crucial task and disturbance conditions for parameter estimation.

Main Results:

• The forward-inverse method accurately estimated motor control parameters.
• Specific perturbation conditions were identified as most informative for gain parameter estimation.
• A set of 40 movements under optimized perturbation conditions yielded high parameter estimation accuracy ($R^{2}=0.97$).

Conclusions:

• The developed technique effectively estimates human arm control properties using simulated or digital twin methods.
• This approach provides a framework for optimizing experimental designs in motor learning research.
• The findings pave the way for more personalized approaches to understanding and enhancing motor control.