CC BY
8
4. Xu Xianhong. Research and design of 12-Lead Synchronization ECG Signal Detection and Analysis System, International journal of medical, pharmaceutical science and engineering, 2011. 5. P. 257-264.
5. Ozhikenov K. A., Altay E. A. Raspredelennaya kardiosistema. Sbornik tezisov XII mezhdunarodnogo kongressa «Kardiostim» Sankt-Peterburg, fevral 18-20, 2016. P. 222.
6. Kening Wang, Weizhao Zhang. Design of ECG Signal Acquisition System Based on DSP. International Workshop on Information and Electronics Engineering, 2012. 5. P. 3763-3767.
7. Chepnyh I. V. Modelipovanie elektpotehnicheskih ustroystv in MATLAB. SimPowerSystems and Simulink. 1-e izdanie, 2007. P. 288.
8. Avdieva D. K., Balohonova M. V., Demyanov S. V. Modelirovanie vliyaniya filtrov na signal EKG v sisteme Matlab. Sovremennyie problemyi nauki i obrazovaniya. 2012, 3. P. 105-112.
9. Varakin L. E. Sistemy svyazi to shumopodobnymi signalami. M.: Padio and communication, 2015. P. 284.
10. Drozdov D. V. Vliyanie filtratsii na diagnosticheskie svoystva biosignalov // Metodicheskie aspektyi: materialyi konferentsii. Moskva: izdatelstvo Altomedika, 2011. P. 75-78.
11. Mahesh S. H., Agarbala R. A. Design and implementation of digital FIR equiripple notch filter on ECG signal for removal of power line interference // Wseas transactions on signal processing, 2008. № 4. P. 221-230.
12. Durgesh Kumar Ojha, Monica Subashini. Analysis of electrocardiograph (ECG) signal for the detection of abnormalities using Matlab // International journal of medical, pharmaceutical science and engineering, 2014. № 2. P. 54-57.
Improving the accuracy of the identification of cardio using a filter bank Shayakhmetkyzy D.1, Koishybayev D.2 Повышение точности идентификации кардиосигнала с помощью банка
фильтров Шаяхметкызы Д.1, Койшыбаев Д. Н.2
'Шаяхметкызы Динара / Shayakhmetkyzy Dinara — магистрант; 2Койшыбаев Даулен Нурланулы /Koishybayev Daulen — магистрант, кафедра сенсорики, факультет лазерной и световой инженерии, Санкт-Петербургский национальный исследовательский университет информационных технологий, механики и оптики, г. Санкт-Петербург
Abstract: the paper deals with the identification and treatment of cardio using a filter bank. With the filter bank disclosed method suppression harmonic interference cardio. For interference suppression and identification of the main cardiographic complex filter bank is designed. Work performed by the most famous multi-tool MATLAB.
Аннотация: в работе рассматривается идентификация и обработка кардиосигнала с помощью банка фильтров. С помощью банка фильтров раскрыт метод подавления помех гармонике кардиосигналов. Для помехоподавления и идентификации основных кардиографических комплексов разработан банк фильтров. Работа выполнена с помощью наиболее известного многофункционального инструмента MATLAB.
Keywords: identification, treatment, cardiology, harmonic, filter bank.
Ключевые слова: идентификация, обработка, кардиосигнал, гармоника, банк фильтров.
In recent decades, actively developing pre-treatment technology of biological signals, enabling the identification of the functional state of the organism (FSO) person at an early stage of the disease. This may be the initial stage of myocardial infarction, atrial fibrillation, tachycardia. In such cases, especially important methods to identify even minor deviations from normal cardio [1, p. 105-112].
One of the major problems for cardio is a filtration system using filter bank. The use of filtering algorithms provides a clear-cut teeth and to identify the most important areas of the cardiac cycle [2, p. 69-77]. In solving this problem is the difficulty in choosing a filtration method to remove certain types of artifacts, as well as [3, p. 75-78] criteria optimization algorithms used. To identify undistorted electrocardiosignal filters with high precision on the level, it is necessary to develop a bank of filters to suppress interference from power lines.
The aim of this work is to study the interference suppression algorithm and identification of the main components - the teeth, segments and ECG intervals [4, p. 283 s.].
To date, we developed a variety of algorithms for solving the problem [5-6-7]. But most of them have a definite drawback - when passing through the useful signal of the filter unit loses its shape, the noise increases, which leads to a shift of the main components of the ECG and reduces the accuracy of the identification of cardio. In order to eliminate this drawback is proposed bank cascaded filters.
In this paper the filter bank to suppress electrocardiosignal interference, which consists of a low-frequency, high frequency and notch filter. For interference suppression and identification of the main cardiographic complex filter bank is designed [8, p. 221-230]. Filter bank consists of a low-frequency, high frequency and notch filter. Amplitude and phase characteristics (APC) were studied during the filter synthesis. The output filter APC without pulsation.
Filters need to be synthesized for identification and treatment of cardio [10, p. 656 s. il.].
Filters need to be synthesized for identification and treatment of cardio. All filters consist of RLC circuit, and the transfer function are described. For the synthesis of filters define the transfer function of the filter, the filter order, sample rate, bandwidth. All parameters folder are shown in Table 1.
Table 1. Filter Settings
№ Filters Frequency boundaries Fl; fC; fH Quality factor Q
1. FLF 5,92; 15.92; 25.92 0.796
2. FHF 0,004; 0,008; 0,02 0.5
3. NF 48,6; 50; 51,1 20
The transfer function of the filter is parameterized low frequency parameters ro and Q. Frequency rejection and quality factor Q. Quality factor - determines the width of the notch.
1. Calculation of the transfer function low frequency filter has the following form:
W (s) =_TO_=_100!_=_10000_
S2 TO + roo! S2 S + 1002 ^2 + 125 628S + 10000
Q o 0,796
where, ro0 - frequency rejection (rad / s), ro0 - frequency rejection (rad/s), and Q-factor signal.
fc = f fL = V25,92• 5,92 =,/153,45 « 19,92^
Q = —f— = 1592 = 0,796
Q fH - fL 20 ,
where, fH - top notch frequency, fL - Lower frequency rejection; fC - the center frequency. If the result of the research center frequency is 15.92 Hz, the remaining frequency fL, fH are within 5.92 and 25.92 Hz.
2. Calculation of the transfer function high frequency filter has the following form:
W (s)=_s2_=_s!_=_s!_
S 2 A W s2 + 005 S + 0.052 s 2 + °.ls + 00025
Q o 0.5
where, ro0 - frequency rejection (rad / s), and Q -factor signal.
fa = ff = V0,02 • 0,004 « 0,008 Q = —f— = 0008 = 0,5.
* fH - fL 0,016
If the result of the research center frequency is equal to 0,008 Hz, the remaining frequency fL, fH are within 0.004 Hz and 0.02 Hz. As we know, the interference from power lines - a narrow-band noise with center frequency of 50 Hz.
W (5) =
2 , 2 s +a0
2 2
s + — s +
Q
where, ro0 - frequency rejection (rad/s) and Q-factor signal
Q r r ; fc yl fH ' fL
J H ~ JL
where, fH - top notch frequency, fL - Lower frequency rejection; fC - the center frequency. If the center frequency is 50Hz, the remaining frequency fH, fC are within 48.6 and 51.1 Hz.
Consequently for the transfer function (1) and the Q signal (2) calculate the parameters of the frequency f = 50 Hz - the frequency of power lines.
co = 2f = 314rad / 5.
fc =JTL =V51.1-48.6 =72499 « 50r\ fc 50
Q =
Jh - fL 2.5
= 20.
Q parameters can say, the more the value, the faster increase of the resonance frequency of the transmission ratio and deletions.
W ( s) =
2 , 2 s +&2
s2 +3142
s2 +98596
s2 4
Q
s + a
2
2 314 2 s 2 +-s + 3142
s 2 + 15.7s + 98596
20
We determine the sampling frequency.
75[hit/minutes] * 90[countdown/hit]
Fsicountdown/secondsl =-—7-——:-=-= llOicountdown/secondsl
60 [seconds/minutesj
At the rate we can say that the sampling frequency or frequency sampling Fs = 110 Hz. Consider the problem of identifying the frequency signal sinusoidal: 5 - Input net pacemaker,S - measurement noise, 50 Hz.
Fig. 1. Block diagram identifying cardiac harmonics
Here, the blocks W1 (s), W2 (s) and W3 (s) entered the transfer functions of low, high, and notch filter. A feedback combine to improve the accuracy of identification. Modeling filter bank is done by MATLAB program [9, p. 2103- 2108].
Experimental studies were carried out using a set of real signals of records obtained directly in medical institutions, as well as of the available data banks, in particular, from databases hosted on the website of MIT (USA). Figure 10 shows the input cardio derived from MIT database. When you run the simulation we get the data, they look like
15 2
Titne(9ec5)
Fig. 2. Input cardio
1.5 2
Time(secs)
Fig. 3. Noisy cardio
IS 1
Time(secs)
Fig. 4. Waveform generator noise
1.Î 2
Time (secs)
Fig. 5. Filtered cardio
2
Time (sees)
Fig. 6. Definitions of the peaks for the filtered ECG signal
The result of simulation I want to mention, all the peaks P, Q, R and teeth S, T recorded without harmonics and distortion. Create a filter bank corresponds to the identification and ECG signal processing.
R-peak detection, and therefore, QRS complexes in the ECG are information heart rate and conduction velocity. They also have information on the condition of the tissue in the heart, as well as a variety of diseases. They are evidence for diagnostics of heart disease. For this reason, these peaks are drawn considerable attention in the processing of ECG signals.
References
1. Avdeev D. K, Balahonova M. V, Demyanov S. V. Modelling the impact of the filters on the ECG signal in Matlab system // Modern problems of science and education, 2012. № 3. S. 105-112.
2. Petrovskii M., Bodin O. N. Portable ECG sensor "Cardiovid" computer diagnostic system // Modern problems of science and education, 2013. № 4. S. 69-77.
3. DrozdovD. V. Effect of filtration on the diagnostic properties of biosignals // Conference materials / functional diagnostics, 2011. № 3. P. 75-78.
4. Fedotov A. A., Akulaev S. A. Mathematical modeling and analysis of error transducers biomedical signals. M: "FIZMATLIT", 2003. 283 s.
5. Durgesh Kumar Ojha, Monica Subashini. Analysis of electrocardiograph (ECG) signal for the detection of abnormalities using Matlab // international journal of madical, pharmaceutical science and engineering, 2014. № 2. Р. 54-57.
6. Egorov E. N., Koronovskii A. A., Khramov A. E. Active filters. Teaching aids. Saratov: Izd GosUNTs "College", 2002. 14 s.
7. Istomin T. V., Krivonogov L. Y. Questions noise immunity for measurements electrocardiosignal parameters. Information - measuring equipment: Hi. Sat. scientific. tr. Penza: Izd Penz. state. University Press, 2000. Vol. 25.
8. Mahesh S. H., Agarbala R. A. Design and implementation of digital FIR equiripple notch Filter on ECG signal for removal of power line interference // Wseas transactions on signal processing, 2008. № 4. P. 221-230.
9. Yihang Zhang, Guangmin Sun. 12 lead ECG data acquisition system based on ADS 1298// Procedia engineering, 2012. № 29. P. 2103-2108.
10. Kuliash M., Yeldos A. Computer modeling electrocardiogram signals using notch filters // European research, 2015. № 3 (4).
11. Layons R. Tsifrovaya Signal Processing: Second izdanie.Per.s English. M.: OOO "Bean-Press", 2006. 656 s.
