Toward Unified Control of a Powered Prosthetic Leg: A Simulation Study

Area of Science:

• Biomedical Engineering
• Robotics
• Control Systems

Background:

• Current powered prostheses often require distinct controllers for different gait phases, complicating operation.
• Hybrid Zero Dynamics (HZD) offers stability but can be sensitive to mismatches with human biomechanics.

Purpose of the Study:

• To develop a unified control strategy for a powered knee-ankle prosthesis that covers the entire gait cycle.
• To improve prosthesis adaptability and robustness against human-machine controller discrepancies.

Main Methods:

• Utilized a reduced-order Discrete Fourier Transformation (DFT) to create unified virtual constraints.
• Converted hybrid-invariant Bézier polynomials into DFT virtual constraints for seamless gait control.
• Simulated the controller on an amputee biped model, comparing it to the original HZD design.

Main Results:

• The unified controller successfully approximated HZD performance in ideal conditions.
• The unified controller demonstrated superior performance over HZD when hybrid invariance was compromised.
• Both feedback linearizing and joint impedance implementations of the unified constraints yielded comparable simulation results.

Conclusions:

• A novel, unified control strategy using DFT virtual constraints can effectively manage the entire gait cycle for powered knee-ankle prostheses.
• This approach offers improved robustness and adaptability compared to traditional HZD methods, particularly in scenarios with human-machine controller mismatches.
• The unified controller presents a promising advancement for more intuitive and reliable prosthetic limb control.