Noise separation in analog integrated circuits using EMD-PCA-ICA

The idea of applying blind source separation (BSS) algorithms has recently been introduced for noise separation in integrated circuits. Until now, the introduced methods were based on multi-channel BSS. But in many real applications only one mixture is available. Therefore, multi-channel BSS methods are not helpful. In this paper, we propose a new approach to separate individual noise and source signals from an observed mixture in analog integrated circuits. This method is based on one-channel BSS and includes empirical mode decomposition (EMD), principle component analysis (PCA) and independent component analysis (ICA). By using this method, we are able to separate noise and estimate desired source signals from only a single observed mixture. Experimental results substantiate the strong potential of the proposed method for noise separation in analog integrated circuits.