A Brief Overview on Machine Learning in Rehabilitation of the Human Arm via an Exoskeleton Robot

Abstract

The research in assistance and rehabilitation robotics is a growing and promising domain that has emerged due to various medical needs such as neuromuscular disorder, musculoskeletal fatigue and amputated limbs. Such therapy robots, called exoskeletons, can be used to support motor functionality and rehabilitation. Furthermore, the use of Machine Learning (ML) showed challenging results and achieved high performance accuracy in many fields like robotics. The ML has been widely employed in the classification of biomedical signals for the rehabilitation of the human arm by means of a wearable exoskeleton robot. In this paper, we present an overview on some works achieved on the upper-limb exoskeletons and the classification of biomedical signals using ML approaches for the rehabilitation of the human arm. First, we briefly present wearable exoskeleton robots used for the rehabilitation of upper limbs. Furthermore, we present the different biomedical signals used for such objective. In addition, we describe some ML techniques used for the classification of these biosignals. Some existing limitations and future directions are also presented.