EMD-Based Denoising of UWB-IR Signal

The empirical mode decomposition (EMD) applied to analysis of nonstationary and nonlinear signals has attracted great research interests. In this paper, a novel adaptive approach based on EMD framework is introduced to remove noise from ultrawideband impulse radio (UWB-IR) signals. Specifically, the noisy UWB-IR signal is decomposed into a finite number of oscillatory modes known as intrinsic mode functions (IMFs) and a residue. Due to different correlation characteristics of signal and noise in IMF domain, significant features of signal are improved when noise is suppressed by multiplying adjacent IMF components directly. By comparing values of original IMF with results of correlation, signal points are extracted resulting in removing noise. The proposed denoising approach is complemented by numerical examples to illustrate the efficient elimination of noise on UWB-IR signal and to demonstrate the signal-to-noise ratio (SNR) improvement induced by this approach as a performance criterion.