Improvement of Fetal Electrocardiogram Extraction by Application of Fuzzy Adaptive Resonance Theory to Adaptive Neural Fuzzy System

non-invasive fetal electrocardiogram (FECG) signal extraction from signals recorded at abdominal area of mother is a challenging problem for the biomedical and signal processing communities. In this paper, we improve the FECG extraction approaches which consist to find the relation-ships between the cardiac potentials generated at the heart level of mother and the potentials recorded on the abdominal area. We used an efficient signal processing method combining a hybrid learning algorithm based on the fuzzy adaptive resonance theory and the hybrid soft computing technique called Adaptive Neuro Fuzzy Inference System (ANFIS) trained with modified Particle Swarm Optimization (PSO) endowed with an initialization strategy to adjust the antecedent parameters of fuzzy rules. We implemented our algorithm and other algorithm on simulated signals, and we found that the proposed ANFIS with hybrid learning algorithm achieves superior performance in learning accuracy and allowed yielding best processing results to extract the FECG signal.