A Real-Time Lift Detection Strategy for a Hip Exoskeleton

Area of Science:

• Robotics
• Biomechanics
• Human-Computer Interaction

Background:

• Repetitive heavy lifting poses significant risks of back pain and injury to workers.
• Current active exoskeletons often lack automatic lift detection, limiting their factory application.

Purpose of the Study:

• To develop and validate a practical lift detection strategy for active exoskeleton control.
• To enable automatic assistance during lifting tasks, reducing L5/S1 joint strain.

Main Methods:

• A novel lift detection strategy using only integrated exoskeleton sensors was proposed.
• Data from seven healthy subjects performing lifting and locomotion tasks were collected for training and testing.
• The algorithm was evaluated for both subject-dependent and subject-independent detection accuracy and timeliness.

Main Results:

• High accuracy was achieved: 97.97% for subject-dependent and 97.48% for subject-independent detection.
• The detection system demonstrated a low time delay of less than 160 ms.
• Surface electromyography confirmed the exoskeleton's efficacy in assisting load lifting.

Conclusions:

• The proposed sensor-integrated lift detection strategy is effective and practical for exoskeleton control.
• This technology shows promise for reducing musculoskeletal injuries in industrial settings.
• Automatic lift detection enhances the usability and safety of exoskeletons for workers.