Project start: 1.10.2025.

Loss of hand function caused by conditions such as stroke or multiple sclerosis limits activities of daily living. Rehabilitation robotics offers effective therapies, reducing the need for intensive therapist involvement. Electromyography (EMG)-based robotic rehabilitation demonstrates significant advantages over conventional methods and open-loop control devices, while soft robotics—inspired by natural organisms and plants—enables safer human interaction.

The proposed innovative research project, building on research currently in the process of publication titled “Koopman-driven grip strength prediction through EMG sensing” in IEEE TNSRE (Q1-Exc, IF 2023: 4.8), focuses on developing methods for surface electromyography (sEMG) signal decomposition using Koopman operator theory (KOT) and dynamic mode decomposition (DMD). The aim is to identify motor unit (MU) activities and correlate these activities with grip strength during various grasp types, with an emphasis on real-time applications.

Koopman operator theory and data-driven DMD methodology enable the decomposition of complex EMG signal dynamics into key spatiotemporal components, extracting frequency and damping features relevant to neural control. By utilizing these innovative methods, the project aims to overcome the limitations of existing approaches in terms of accuracy and real-time execution capabilities.

The integration of advanced sEMG decomposition methods with the principal investigator’s previous research in soft robotics—including the paper “Control of soft robots with inertial dynamics” published in Science Robotics (Q1-Exc, 2-year IF: 26.1)—opens the possibility for developing adaptive robotic devices. Such devices would significantly enhance rehabilitation efficiency by providing more precise control of execution force and adaptability to patient-specific needs.