FECG Extraction Using Adaptive LMS

To study characteristics of Fetal Electro Cardio Gram (FECG) such as heart rate, PQRS waveform and its dynamic behavior are vital parameters for determining fetal life, development, its maturity and congenital heart disease. In fetal monitoring, FECG extraction is a critical process for obtaining significant information during pregnancy and labor since 1906. In this research, an adaptive Least Mean Square (LMS) algorithm is evaluated. Optimal convergence can be achieved by using proper values of step size and weight coefficient. For assessment of algorithm, we have considered two key parameters of adaptive filter i.e. the step size and weight coefficient. We have studied the relationship of weight coefficient and step size with convergence time and convergence rate. Earlier study indicates the step size of 0.00007 and weight coefficient of 15 trading off with slow convergence and less stability. In contrast, our results proves to give fast convergence with more stability while optimizing the value for step size and weight coefficient.