The disability associated with limb amputation makes it difficult to perform the simplest everyday activities. Robotic prostheses can be used to address this complication. These prostheses apply machine learning methods to the EMG/ENG signals to understand the amputee's intention. The use of ENG signals compared to EMG signals is very recent, and allows not only the amputee to perform gestures, but also to mitigate the symptoms of the phantom limb and to restore the sense of touch, since the robotic arm can provide tactile feedback to the peripheral nervous system. In this work, a technique to classify ENG signals, recorded from individuals with limb amputation, is described. All the steps that compose the technique are illustrated in detail. In the last part of this article, some innovative deep learning techniques are suggested in order to improve the state-of-the-art.
Toward a brain-controlled prosthetic arm through advanced machine learning methods
Massa S. M.;Manca M. M.
2020-01-01
Abstract
The disability associated with limb amputation makes it difficult to perform the simplest everyday activities. Robotic prostheses can be used to address this complication. These prostheses apply machine learning methods to the EMG/ENG signals to understand the amputee's intention. The use of ENG signals compared to EMG signals is very recent, and allows not only the amputee to perform gestures, but also to mitigate the symptoms of the phantom limb and to restore the sense of touch, since the robotic arm can provide tactile feedback to the peripheral nervous system. In this work, a technique to classify ENG signals, recorded from individuals with limb amputation, is described. All the steps that compose the technique are illustrated in detail. In the last part of this article, some innovative deep learning techniques are suggested in order to improve the state-of-the-art.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.