SEMG Detection And Processing Based On FPGA

Surface electromyography, SEMG is a one-dimensional time series signal, whichreflects the bioelectrical changes of neuromuscular system. It can be detected bybiological electrode over skin. Motor unit action potential is cased by correspondingexcited motor unit, which will propagate along the muscle fiber and through volumeconductor filter composed of skin, fat and muscle. Then SEMG is result of comprehensivesuperposition of the motor unit action potential trains both in time and space. Since thenoninvasive characteristic of SEMG, it has a vide range of application, such as the basicresearch of neuromuscular, clinical diagnosis, rehabilitation project, human computerinterface and so on. But raw EMG offers us valuable information in a particular uselessform, and it is easily interfered by various factors. So it needs to be processed andquantified.Firstly, we analyzed characteristic of EMG and various interference factors, then thepaper introduced and compared the traditional methods that used to EMG signalprocessing. We designed SEMG signal detection and acquisition circuit that includesfront-end detection and amplified circuit, band pass filter, north filter and analog to digitalconversion circuit which controlled by FPGA on basis of SEMG theory above. At last, thedigital was send to FPGA for three level discrete wavelet transform. This paper emphasison implementation of discrete wavelet transform based on FPGA, and studied FPGAdistribute algorithm and Mallat fast algorithm.The digital acquired from circuit was processed by MATLAB and VerilogHDLprogram, respectively. The comparing result presents that DWT can be implementedconveniently by FPGA, meanwhile, it has characteristic of fast and reliable.