Subject-Independent Estimation of Continuous Movements Using CNN-LSTM for a Home-Based Upper Limb Rehabilitation System

• Published in: IEEE robotics and automation letters Vol. 8; no. 10; pp. 6403 - 6410
• Main Authors: Li, He, Guo, Shuxiang, Bu, Dongdong, Wang, Hanze, Kawanishi, Masahiko
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.10.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Artificial neural networks; Back propagation networks; Brushless motors; Convolutional neural networks; deep learning methods; Elbow; Electromyography; Estimation; Exoskeletons; Intention recognition; model learning for control; motion control; Motion simulation; Neural networks; Real-time systems; Rehabilitation; rehabilitation robotics; Time lag; Training
• ISSN: 2377-3766; 2377-3766
• DOI: 10.1109/LRA.2023.3303701

Exoskeleton-assisted home-based rehabilitation plays a vital role in the upper limb rehabilitation of stroke patients in early stage. The surface electromyography (sEMG)-based control can facilitate friendly interactions between individuals and rehabilitation exoskeletons. The exoskeleton can also meet the requirements of home-based rehabilitation, including affordability, portability, safety, and active participation. Although various systems have been proposed to enhance upper limb training, few studies have addressed the inter-subject variability of sEMG signals, which limits the generalization capability of the intention estimation model. In this letter, a subject-independent continuous motion estimation method combining convolutional neural networks (CNN) and long and short-term memory (LSTM) is proposed and applied to a home-based bilateral training system. The sEMG-driven CNN-LSTM model builds the relationship between sEMG signals and continuous movements. To verify the effectiveness of the CNN-LSTM model in achieving subject-independent estimation, the offline estimation under the backpropagation neural network, CNN, and CNN-LSTM are compared. Moreover, the online intention estimation and the real-time control are performed, and the estimation angle error and time delay are controlled at approximately 10° and 300 ms, proving the feasibility of the subject-independent estimation method and its availability in the upper-limb rehabilitation system.