A Cyber Expert System for Auto-Tuning Powered Prosthesis Impedance Control Parameters.

Area of Science:

• Biomedical Engineering
• Rehabilitation Engineering
• Robotics

Background:

• Manual tuning of powered lower limb prosthesis control parameters by human experts is time-consuming and resource-intensive.
• Automated tuning methods are crucial for increasing the clinical practicality of advanced prosthetic devices.

Purpose of the Study:

• To develop and evaluate a novel cyber expert system (CES) for automated tuning of powered knee prosthesis control parameters.
• To compare the efficiency and effectiveness of CES tuning against traditional human expert tuning.

Main Methods:

• A cyber expert system (CES) was developed, encoding human expert tuning decisions into computer rules.
• The CES was used to auto-tune control parameters for a powered knee (passive ankle) prosthesis.
• Tuning performance was evaluated on two able-bodied subjects and two transfemoral amputees.

Main Results:

• CES tuning required less time and no human intervention compared to human expert tuning.
• Both CES and human expert tuning resulted in normative prosthetic knee kinematics.
• Gait symmetry and step width showed improvement or slight improvement within subjects after both tuning methods.

Conclusions:

• The cyber expert system (CES) shows promise for enhancing the practical value of powered prosthetic legs through automated control parameter tuning.
• CES tuning objectives may not fully encompass all human expert goals, as evidenced by differences in trunk sway.
• Further research is necessary to refine tuning objectives for CES to optimize amputee walking function and reduce trunk motion.