A Simplified Approach to Automatic Stance Control of Unilateral Knee Exoskeletons.

Area of Science:

• Robotics
• Biomechanics
• Rehabilitation Engineering

Background:

• Practical lower limb exoskeletons require safe and effective control systems.
• Simplifying sensor sets is crucial for reducing complexity without sacrificing performance.
• Existing control strategies often require extensive sensor integration.

Purpose of the Study:

• To develop and evaluate a stance control algorithm for unilateral knee exoskeletons using limited sensor data.
• To ensure safety and effectiveness through automatic joint locking during the stance phase of gait.
• To validate the algorithm's performance across different walking speeds and user variations.

Main Methods:

• A novel stance control algorithm utilizing only motor current data and inertial measurement unit (IMU) kinematic information.
• Implementation on a unilateral knee exoskeleton controlled by a Finite State Machine for automatic joint locking.
• Real-time evaluation with six healthy subjects on an instrumented treadmill at 0.60, 0.80, and 1.00 m/s.

Main Results:

• The algorithm demonstrated accurate state transitions without false positives or negatives across all subjects.
• Successful automatic joint locking prior to heel strike with a mean phase error of 4.81% ± 2.70%.
• Timely unlocking before toe-off within the terminal stance phase, with a mean phase error of 5.28% ± 2.52%.

Conclusions:

• The developed stance control algorithm is effective and reliable for unilateral knee exoskeletons in healthy individuals.
• The simplified sensor approach enhances the practicality of lower limb exoskeletons for real-world applications.
• Preliminary results support further investigation in larger populations, including individuals with mobility impairments.