Activity recognition of the torso based on surface electromyography for exoskeleton control

• We present an activity mode recognition approach to identify motions of the human torso.
• Approach uses decision tree classification in order to leverage its computational efficiency.
• The recognizer uses surface electromyography as the input and CART (classification and regression tree) as the classifier.
• Results indicate that the recognizer can extract the user's intent with minimal delay.
• We achieved a low recognition error rate and a user-unperceived latency using sliding overlapped window.

This paper presents an activity mode recognition approach to identify the motions of the human torso. The intent recognizer is based on decision tree classification in order to leverage its computational efficiency. The recognizer uses surface electromyography as the input and CART (classification and regression tree) as the classifier. The experimental results indicate that the recognizer can extract the user's intent within 215 ms, which is below the threshold a user will perceive. The approach achieves a low recognition error rate and a user-unperceived latency by using sliding overlapped analysis window. The intent recognizer is envisioned to a part a high-level supervisory controller for a powered backbone exoskeleton.