A Probabilistic Spectral Analysis of Multivariate Real-Valued Nonstationary Signals

TL;DR

This paper introduces a probabilistic spectral analysis framework for multivariate nonstationary signals using complex Gaussian models, enabling robust detection and characterization of signal properties in challenging environments.

Contribution

It presents a novel probabilistic spectral representation for nonstationary signals, with statistically consistent estimators and a generalized likelihood ratio test for nonstationarity detection.

Findings

Effective spectral analysis in low-SNR conditions.

Consistent estimators derived within a maximum likelihood framework.

A new test for nonstationarity detection.

Abstract

A class of multivariate spectral representations for real-valued nonstationary random variables is introduced, which is characterised by a general complex Gaussian distribution. In this way, the temporal signal properties -- harmonicity, wide-sense stationarity and cyclostationarity -- are designated respectively by the mean, Hermitian variance and pseudo-variance of the associated time-frequency representation (TFR). For rigour, the estimators of the TFR distribution parameters are derived within a maximum likelihood framework and are shown to be statistically consistent, owing to the statistical identifiability of the proposed distribution parametrization. By virtue of the assumed probabilistic model, a generalised likelihood ratio test (GLRT) for nonstationarity detection is also proposed. Intuitive examples demonstrate the utility of the derived probabilistic framework for spectral analysis in low-SNR environments.