A Low Cost Surface EMG Sensor Network for Hand Motion Recognition

Surface EMG is widely used for hand motion recognition. This paper has developed a low cost surface EMG sensor network which consists of four surface EMG sensors and a computer software. The design of the wireless surface EMG sensor and the computer software are described in detail. Four time-domain feature are extracted from the raw EMG signals. And the extracted EMG features are used to train the BPNN in MATLAB. The trained BPNN is used to realize the online motion recognition. Experiments of six target hand motion recognition are conducted to verify the designed system. The results show that the average recognition accuracy of using one feature, two features, three features and four features are 91.13%, 94.83%, 95.56%, 96.09% respectively.