Research on f Wave Extraction Method of Body Surface Atrial Fibrillation Signals Based on Adaptive Filtering

Abstract

Atrial fibrillation (AFib) is one of the most common cardiovascular diseases. However, at present, the treatment means of AFib are not perfect. During AFib, its P wave is disturbed and morphologically irregular waves, f waves that are different from F waves in atrial flutter, could be seen in the baseline. Many studies have shown that f waves are of great significance for the study of AFib, so it is essential to find a way to extract f waves accurately, but current f wave extraction methods all have their own limitations. In this thesis, several sets of AFib signal data from real patients were measured by using our self-developed body surface potential mapping (BSPM) system. An adaptive filtering method based on template elimination theory, suitable for single lead f wave extraction was proposed for solving the current problems. It could track the morphological changes of AFib signals, thereby reducing the influence of interferences and the difference of the signal itself on f wave. After a series of signal preprocessing steps, the signal segments were divided into different situations. The f waves were extracted by the traditional template elimination method, and the method proposed in this work separately. The results showed that the proposed method has obvious advantages in terms of accuracy and robustness compared with the traditional method. In addition, this thesis also put forward the optimization algorithms for the traditional signal extraction method, and carried out the relevant experiments. The results revealed that the proposed optimization algorithms effectively improved the accuracy of the original method and it has certain enlightenment for the research in the future.