Analytical Study On ECG Signal Based On Wavelet Transform And Its FPGA Implementation

Cardiopathy disease is a Common Chronic Diseases with high mortality, and it hasbecome an international public health issue. ECG is now the necessary method for thediagnosis in the cardiac disease in clinic, so the research on the automatic ECG analysissystem has very high value on clinic. ECG is weak and it’s extremely vulnerable toenvironmental impacts, which make it is more hard to detection and diagnose. And theprecise automatic analysis and research play a key role on the clinicopathologic analysis ofcardiovascular disease.The research content of the ECG signal automatic analysis and diagnosis isvery widespread, but this paper only deals with the ECG preprocessing and R wavesdetection. We make the utmost of FPGA device which has some feature of real-time, highreliability and parallel processing. And this article proposed the implementation of a waveletprocessor for ECG preprocessing and R waves detection on a field programmable gate array(FPGA) device based on software.The ECG signal pretreatment algorithm: This paper puts forward a kind of ECG signalpreprocessing hardware algorithm based on db4lifting wavelet structure and its easyimplementation on hardware. This paper uses the soft-threshold to the wavelet liftingcoefficients using the fixed threshold estimated in the sub-bands that include noises. Thenthe algorithm is designed by using Altera SOPC Builder for Matlab/Simulink environment,the various modules are integrated on Cyclone II EP2C35F672C8chip. Experimental showsthat this paper puts forward an efficient structure to remove efficiently baseline wander,muscle power interference and power-line interference from ECG signal. Furthermore,implementation time and power consumption are analyzed, it is shown that the proposeddenoising scheme provides an effective method for the portable, real-timing and low-powerconsumption ECG monitor.The R waves detection algorithm: This paper puts forward the R waves detectionalgorithm based on the Biorthogonal B-spline wavelets. We use the wavelet Lipschitzpolynomial principles to detect the singularities and explain the relation between thezero-crossing point of the modulus maximum pair of its wavelet transfer and the signalsingularity (peak R). This paper also presented some strategies to minishthe probability of dis-detecting and leak-detecting and to improve the detection rate. Thenthe FPGA implementation system is designed based on wavelet transform detectionalgorithm. And the FPGA system includes two modules: wavelet decomposition module anddetection module. The wavelet decomposition module was designed in cascade of pipelinearchitecture to improve operational efficiency, and wavelet transform on each layer can be used as basic computing unit. The detection module uses the output of the waveletdecomposition module to search the modulus maxima in the3rdand4thlayers and detectQRS complex. Finally, the algorithm was carried out on the entire MIT-BIH ArrhythmiaDatabase ECG records (about48records) and the experiment results show that the systemhas the advantages of being highly efficient and practical.