A Home-Based Upper Limb Rehabilitation System via Cloud-Based Teleoperation and sEMG-Driven Bilateral Control

With the growing number of patients suffering from upper limb hemiplegia, robot-assisted rehabilitation has attracted more and more attention. Compared to traditional face-to-face rehabilitation, telerehabilitation is an effective alternative with therapist-in-the-loop. Meanwhile, bilateral rehabilitation based on surface electromyography (sEMG) enables patients to train by themselves. Although numerous telerehabilitation or bilateral rehabilitation systems have been proposed, limited studies have addressed cloud communication and inter-subject variability. This paper proposes a home-based upper limb rehabilitation (HB-ULR) system utilizing cloud-based teleoperation and sEMG-based subject-independent bilateral control. In the cloud-based telerehabilitation (CBTR) subsystem, experiments with the master side in Beijing City (China) and the slave side deployed in three different cities are conducted through one cloud server. The slave side is controlled by the master side, while the contact force is transmitted back to the master side. In the sEMG-driven subject-independent bilateral rehabilitation (sEMG-SIBR) subsystem, continuous motion can be predicted by a model after transfer learning. The validity of transfer learning in solving inter-subject variability is verified by both offline and real-time experiments, with the prediction error kept within 100. Therefore, the HB-ULR system integrating CBTR and sEMG-SIBR subsystems is built. It supports both routine tele-rehabilitation and daily self-training at home, offering significant potential for enhancing the recovery outcome.