Application of Empirical Mode Decomposition to Reduce Signal Noise

Signals can be affected by noise, which is a type of unwanted signal. Noise can make it difficult to extract useful information from a signal and can result in incorrect output. Filtering out noise before using it to analyze a signal is a common pre-processing step. In this paper, a new method is proposed to reduce signal noise based on empirical decomposition method (EMD). EMD decomposes the signal into a set of oscillating modes called intrinsic mode functions (IMFs), and noise reduction involves reconstructing this signal using IMFs by adding them together according to a mathematical formula to finally arrive at a signal with very low noise. Then the result is verified using a standard based on determining the error value (SNR) between the original signal and the filtered signals, after which a spectral analysis of the resulting signals is performed and compared with the spectrum of the original signal. The method was tested on simulated data of a signal (Bumps), where the results show that EMD technology may prove to be an effective technique for reducing signal noise. The motivation for this research work was derived from the quest to develop EMD-based filters for signal filtering and feature extraction with better results as compared to other filters.