Reservoir Computing Model for Human Hand Locomotion Signal Classification

Human-movement recognition is a novel challenge in soft robotics. In recent years, there have been several attempts to develop wearable devices for supporting human-robot interfaces. Many algorithms and programming languages are available integrate device with robot. One such promising algorithm reservoir computing (RC), which includes of group recurrently randomly connected nodes. The RC model can easily process multidimensional signal data handle nonlineardata has extensively used robotic control. It reported that the speed up network training solve complex sets. However, main existing limitations hand-locomotion classification considerable run-time delayed response. this study, we figure out best machine learning three-dimensional hand-gesture data. We employ two-part strategy: loopback filter included preprocessing initial dataset support 3-dimensional (3D) signs each hand posture; subsequently, applied an artificial neural (ANN), convolutional (CNN), long short-term memory(LSTM), computing(RC). Each optimized various hyperparameters. Furthermore, compare performance machine-learning classifying hand-signal posture results show nonlinear signals by requires comparatively shorter duration (12 minutes times), optimal accuracy 94.17, precision 94.10, recall 93.99 realized time series