Artificial neural network–based modelling of pacemaker and its pace tracking using discrete wavelet transform–based finite dimension repetitive controller

Abstract

Efficient control of cardiac pacing is a very important aspect as it provides lifesaving regulated cardiac rhythm in this dynamic hostile environment. The foremost control objective is set to design a highly reliable and advanced control strategy to ensure the utmost accuracy in the control effort. A modified artificial neural network (ANN)–based modelling and pace tracking using finite dimension repetitive controller (FDRC) design based on internal model principle (IMP) has been presented here. This controller will not only provide accurate tracking but also minimize the control action time due to less amount of data handling through the deployment of discrete wavelet transform (DWT) in the loop of repetitive controller (RC). Finally, a case study has been propounded considering ANN model using available data sets and software to validate the control strategy and justify the control objective for optimizing the pace tracking in a pacemaker. Result of the experiment showed good accuracy as well as very low error in terms of mean-squared error (MSE), integral absolute error (IAE), integral time absolute error (ITAE) and integral time square error (ITSE). Along with that, it is observed that DWT not only benefits the handling of very less memory but also acts as an additional filter while reconstructing the signal, which serves as an added advantage of this model.