Decomposing Non-Stationary Signals With Time-Varying Wave-Shape Functions

Modern time series are usually composed of multiple oscillatory components, with time-varying frequency and amplitude contaminated by noise. The signal processing mission is further challenged if each component has an pattern, or the wave-shape function, far from a sinusoidal pattern even changing to time. In practice, components exist, it desirable robustly decompose into for various purposes, extract desired dynamics information. Such challenges have raised significant amount interest in past decade, but satisfactory solution still lacking. We propose novel nonlinear regression scheme its constituting frequency, function. coined algorithm shape-adaptive mode decomposition (SAMD). addition simulated signals, we apply SAMD two physiological impedance pneumography electroencephalography. Comparison existing solutions, including linear regression, recursive diffeomorphism-based multiresolution decomposition, shows that our proposal can provide accurate meaningful computational efficiency.