Design and Stability Analysis of an Adaptive Neuro-Fuzzy Inference System (ANFIS)-Based Pacemaker Controller in MATLAB Simulink.
Journal:
Journal of long-term effects of medical implants
PMID:
38842228
Abstract
We present the design and stability analysis of an adaptive neuro-fuzzy inference system (ANFIS)-based controller of a pacemaker in MATLAB Simulink. ANFIS uses learning and speed properties of fuzzy and neural networks. Based on body states and preprogrammed situations of patients (age and sex, etc.), heart rate and amplitude of pacing pulse are changed. Output signal that is fed backed from heart is compared to the reference fuzzy bases ANFIS signals. After designing ANFIS based controller, the stability of the proposed system has been tested in both the time (step response) and trequency (Bode diagram and Nichols chart) domains. In our previous study, the step response analyzed and compared with other works. For frequency domain, all the possible frequency analysis methods have been tested but because of nonlinear properties of ANFIS, after linearization, just the Bode diagram achieved good results. The step response results in time domain is compared with previous work's results including optimum heart pulse rate for each particular patient. In the frequency domain, the Bode diagram stability analysis showed gain and phase margin as follows: GM (dB) = 42.1 and PM (deg) = 100.