EMG-driven models of human-machine interaction in individuals wearing the H2 exoskeleton**This work was supported by the ERC Advanced Grant DEMOVE [267888]

Abstract

EMG-driven modeling has been mostly used offline and on powerful desktop computers, limiting the application of this technique to neurorehabilitation settings. In this paper, we demonstrate the use of EMG-driven modeling in online (i.e. in real-time) running on a fully portable embedded system and interfaced concurrently with a powered lower limb exoskeleton. This work provides evidence of the feasibility of real-time model-based control of complex multi-joint exoskeleton system, thus opening new avenues for personalised robot-aided rehabilitation interventions.