Gestures recognition of sEMG signal based on Random Forest

The surface electromyographic (sEMG) signal is a bioelectric signal generated by muscle activity collected from the human epidermis, which contains rich information about muscle activity, and the EMG signal caused by different movements is different, which can reflect the different movement states of human limbs. Therefore, by collecting, processing and feature extraction of sEMG signals, it can realize the recognition of human hand movements, and can be used as the control source of the prosthesis to realize the control of the prosthesis. In this paper, we propose a Random Forest algorithm-based model for sEMG signal processing and human hand movement recognition, and design an EMG signal conditioning circuit for the acquisition and conditioning of sEMG signals of arm muscles, and store the data using a NI data acquisition card. At the same time, the random forest model is trained using the public dataset to realize the classification of four kinds of actions: Fist clenching, Hand opening, Wrist internal rotation and Wrist external rotation, which provides a new method and idea for hand gesture action recognition and also facilitates further research work in the follow-up.