The present discloses an IIR adaptive filtering method. The method can extract a desired low-frequency signal from a source signal by high-frequency noise cancellation, for example, the low-frequency signal can be an ECG signal. There are two main features in the method. First, it uses a guaranteed stability criterion to bring the output data of the filter into convergence. Second, it decreases the interferences under the finite precision digital IIR adaptive filter by normalizing the input parameter of the filter. In summary, the method can be easily implemented by a finite precision processor, such that user can use minimal cost to design and implement the digital IIR adaptive filter.本發明揭露一種針對有限精準度之適應性線性時變濾波方法,其適用於去除來源訊號中之高頻雜訊訊號以擷取低頻之主訊號,例如心電圖訊號。主要特點係利用一穩定度判別方式以收斂濾波器之輸出結果,以及利用正規化濾波器參數將精準度之干擾降至最低。此方法可以實現於一有限精準度處理器中而不產生發散之濾波結果,讓使用者可以用以較低之成本以完成濾波器。10‧‧‧來源訊號20‧‧‧適應性濾波器30‧‧‧時間延遲裝置40‧‧‧誤差訊號50‧‧‧輸出訊號60‧‧‧延遲訊號
