Related Experiment Video
Updated: May 25, 2026
11:16
Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis
Published on: July 22, 2014
Real-time implementation of an intent recognition system for artificial legs
Fan Zhang1, Zhi Dou, Michael Nunnery
1Department of Electrical, Computer, and Biomedical Engineering, University of Rhode Island, Kingston, RI 02881, USA.
Summary
This study developed a real-time intent recognition system for transfemoral amputees using muscle signals and prosthetic sensor data. The system accurately identifies locomotion modes and predicts transitions, enhancing prosthetic leg control.
Area of Science:
- Biomedical Engineering
- Rehabilitation Robotics
- Neuroprosthetics
Background:
- Transfemoral (TF) amputees often face challenges in controlling prosthetic limbs.
- Accurate intent recognition is crucial for seamless and safe prosthetic leg function.
Purpose of the Study:
- To implement and evaluate a real-time intent recognition system for a TF amputee.
- To fuse neuromuscular and mechanical data for enhanced prosthetic control.
Main Methods:
- Surface Electromyographic (EMG) signals from residual thigh muscles were recorded.
- Ground reaction forces/moments from the prosthetic pylon were collected.
- Neuromuscular-mechanical fusion was employed to identify locomotion modes and tasks.
Main Results:
- The system accurately identified three locomotion modes (level-ground walking, stair ascent, stair descent) and tasks (sitting, standing).
- Overall recognition accuracy in static states reached 98.36%.
- Task transitions were predicted 80-323 ms before critical timing for safe prosthesis control.
Conclusions:
- The neuromuscular-mechanical fusion system demonstrates high accuracy in real-time intent recognition for TF amputees.
- Early prediction of task transitions offers significant potential for improving the safety and functionality of powered prosthetic legs.
- This system shows promise for advancing neural control strategies in neuroprosthetics.
