Surface EMG Signal Analysis And Real-time System Construction

Surface EMG (SEMG) is a kind of physiological signals caused by human muscle activity. During the same action, the SEMG at the corresponding muscles shows a repeatable pattern. Thus, characteristics associated with muscle activity can be obtained from SEMG. The aim of this paper is to build a real-time SEMG-based control system. The key problems are how to extract SEMG characteristics, how to obtain the control commands from SEMG and how to build real-time system.Firstly, the experimental raw SEMG was studied to achieve an algorithm for SEMG analysis. Short-time Fourier transform was used to extract the control commands from SEMG of shoulder muscles and calf muscles. Individual differences in handedness, electrocardiogram (ECG) interference and other factors were also considered. The proposed algorithm can reflect the muscle activity from SEMG with good performance of high accuracy and fast processing speed.Secondly, a real-time bio-signal processing platform was constructed. The platform has several functional modules including data acquisition, data analysis and control commands extraction. The platform can be applied to real-time analysis of SEMG and other biological signals.Finally, a real-time SEMG cursor control system is achieved based on the analysis of SEMG and the study of real-time platform. The system is simple and easy to use. The real-time performance is good and the accuracy rate of control commands conversion is high. In addition, the human shoulder and elbow movements were studied to obtain the relationship between SEMG and upper limb movement intention. Real-time movement intention of arm was obtained based on the analysis of SEMG. It is a novel algorithm for upper limb assisted device and has application significance.