LSTM Classification of Functional Grasps Using sEMG Data from Low-Cost Wearable Sensor

Modelling human grasping and transferring this data to an anthropomorphic robotic hand to endow it with human like grasping capabilities is a complex task. In this paper the use of surface electromyography (sEMG) for classification of functional grasps associated with everyday life is carried out using a low-cost wearable sensor in conjunction with state-of-the-art recurrent neural networks. The results produced through these experiments demonstrate the potential for sEMG to be used as an effective medium for transferring human demonstration to a robotic system.