A physiologically inspired hybrid CPG/Reflex controller for cycling simulations that generalizes to walking

Area of Science:

• Biomechanics
• Computational Neuroscience
• Robotics

Background:

• Motor control simulations often use task-specific controllers, limiting generalizability.
• Existing models include feedback-based, reflex-based, and Central Pattern Generator (CPG) feedforward architectures.
• Hybrid models integrating feedback and feedforward mechanisms offer a promising alternative.

Purpose of the Study:

• To propose and evaluate a generalizable hybrid controller for predictive simulations of cyclical lower limb movements.
• To test the controller's ability to simulate physiological cycling patterns.
• To assess the controller's adaptability to different tasks, such as walking.

Main Methods:

• Developed a hybrid controller combining a feedforward component (Unit Burst Generation model) with reflex-based feedback pathways.
• Simulated stationary cycling at various speeds and seat heights.
• Modified the controller by adding a balance control component to simulate walking.

Main Results:

• The hybrid controller successfully simulated physiological stationary cycling patterns.
• The controller demonstrated generalization to walking simulations with the addition of a balance component.
• The proposed controller provides a physiologically inspired and generalizable model for lower limb motor control.

Conclusions:

• A hybrid controller can effectively simulate diverse cyclical lower limb tasks.
• This generalizable approach advances the understanding of motor control principles.
• The model offers a foundation for future research in neuroprosthetics and robotics.