Real-Time CANFIS and ANFIS Based Pacemaker Controller Design and Analysis

الموضوعات :

Asghar Dabiri Aghdam ¹ , Nader Dabanloo ² , Fereidoun Nooshiravan ³ , Keivan Maghooli ⁴

1 - Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
2 - Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
3 - Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran
4 - Department of Biomedical Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran

تاريخ الإرسال : 19 الأحد , رجب, 1443 تاريخ التأكيد : 03 الأربعاء , شوال, 1443 تاريخ الإصدار : 05 الخميس , صفر, 1444

الکلمات المفتاحية: ANFIS, PID, eCG, CANFIS, Ventricular tachycardia, Atrial fibrillation, Heartbeat, Artificial Pacemaker, Bradycardia, Tachycardia,

ملخص المقالة :

This paper describes ANFIS introduced by R. Jang et al. ANFIS actually is an offline method in fuzzy control systems. First, a fuzzy file called FIS (Fuzzy Inference System) is designed that relates the input and output of the system by membership functions that are optimized during the learning process. Input and output learning data are given to the ANFIS (MATLAB command line or ANFSI utility) and the output file is used to test or predict new input data. We can then construct a SIMULINK file to simulate the control system. This simulation is not real-time and if the environmental or input conditions are changed, the output will be altered because the FIS file is fixed and not adapted to input variations. The library of online ANFIS and CANFIS introduced does not have that problem and easily learns the online training data and then can mitigate the output in real-time. To avoid the unsuitable patient data itself as training data, we should use a healthy person ECG (heart rate) data in memory to train our fuzzy system and then switch the input data from healthy data to the patient original heart rate as input data. If the heartbeat falls below (60 bpm that is called Bradycardia) or exceeds (100bpm that is called Tachycardia) from a predetermined value. The online controller will switch the controller to healthy data and will stimulate heart muscles at a right beat rate (70-75 bpm). To distinguish tachycardia from body natural states like running, practicing, walking, sleeping and resting, MEMS accelerometer and in some situations, gyros are used. The Bode diagram stability shows gain and phase margin as follows: GM (dB)= 42.1 and PM (deg) = 100. FIS file is saved after an acceptable rms error (0.38). The simulation results of unity step input response (Rise time, settling time, overshoot) will be demonstrated in chapter 4. The overshoot was less than 2 percent and rise time of 2 seconds with settling time of less than 2 seconds. The parameters have been shown for 60 and 72 and 85 bpm.