Research Of Signal Processing Algorithm In ECG Remote Monitoring

Heart disease is a common and chronic disease. The state of this illness is concealed, slow-developing and high morbid risk. So it has been principal disease to threaten people's life. About 70% of the death caused by heart disease is the accident happened outside hospitals. Monitoring the ECG of high risk group can greatly reduce the death rate of cardiopaths. For this reason, to develop ECG remote monitoring and diagnosis system is of great significance.The key point to improve the intellectual ability and availability is to realize quick and exact auto-detection of ECG. So conducting research on this field is very meaningful in both theory and practice.This research is supported by:"Spring Sunshine Program; Ministry of Education". Accomplished works are as follows:First of all, to begin with ECG pretreatment and morphology characteristic detection, various existed algorithms of ECG automatic detection are summarized and analyzed. Each algorithm is discussed in detail and its disadvantage is pointed out. This is the foundation of developing new ECG detection method.Secondly, the theory of artificial neural network (ANN) is combined to optimize traditional linear adaptive whitening filter. An ANN-based nonlinear whitening filter is designed to model the lower frequencies of ECG which is inherently nonlinear and nonstationary. The residual signal which contains mostly higher frequency QRS complex energy is then passed through a linear matched filter to detect the location of the QRS complex. This algorithm can adapt the nonlinearity and is very effective for QRS detection.Thirdly, a template which can be updated for matched filter is designed. A 3-layers BP net is trained by ECG data, and the output of this net can be updated immediately. The test results have good reliability.Finally, in this study, lots of experiments using Matlab and ECG data of MTT-BIH database validated the accuracy and feasibility of the proposed algorithm. Compared with linear adaptive filter, it enhanced the adaptability to nonlinear signal and improved the identified rate of QRS complex in a very noise environment. Compared with multilayer structure nonlinear filter, it increased the rapidity. The pilot study to intelligentize ECG remote monitoring system is conducted in this paper, and it is helpful for the further research in the future.