Time-Frequency Analysis of Non-Stationary Biological Signals with Sparse Linear Regression Based Fourier Linear Combiner.

摘要: It is often difficult to analyze biological signals because of their nonlinear and non-stationary characteristics. This necessitates the usage time-frequency decomposition methods for analyzing subtle changes in these that are connected an underlying phenomena. paper presents a new approach time-varying characteristics such by employing simple truncated Fourier series model, namely band-limited multiple linear combiner (BMFLC). In contrast earlier designs, we first identified sparsity imposed on signal model order reformulate sparse regression model. The coefficients proposed then estimated convex optimization algorithm. performance method was analyzed with benchmark test signals. An energy ratio metric employed quantify spectral results show Sparse-BMFLC has high mean (0.9976) outperforms existing as short-time transfrom (STFT), continuous Wavelet transform (CWT) BMFLC Kalman Smoother. Furthermore, provides overall 6.22% reconstruction error.