A Transformer-based Multi-task Learning Framework for Myoelectric Pattern Recognition Supporting Muscle Force Estimation

• Published in: IEEE transactions on neural systems and rehabilitation engineering Vol. 31; p. 1
• Main Authors: Li, Xinhui, Zhang, Xu, Zhang, Liwei, Chen, Xiang, Zhou, Ping
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.01.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Classification; Control systems; Electrodes; Electromyography; Estimation; Force; Gesture recognition; Interfaces; Learning; multi-task learning; Muscle contraction; muscle force estimation; Muscles; Myoelectric control; Myoelectric pattern recognition; Myoelectricity; Pattern recognition; Prostheses; Robust control; Task analysis; transformer model; Transformers; varying muscle strengths

Simultaneous implementation of myoelectric pattern recognition and muscle force estimation is highly demanded in building natural gestural interfaces but a challenging task due to the gesture classification accuracy degradation under varying muscle strengths. To address this problem, a novel method...