Mixed-Signal Estimation Using Gaussian Mixtures for Signal-Noise Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating wireless signals from multiple sources with known directional vectors fail to accurately determine the temporal and spectral supports of the signals, especially when the signals contain noise, as they rely on linear or pseudo-linear processing which cannot distinguish between signal and noise components.
Innovation Solution
A method involving discrete Fourier transforms and conditional expectation estimators, using Gaussian mixture models and decision processing with linear filters to refine signal estimates, allowing for non-linear processing and improved signal delimitation in time-frequency analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If linear or pseudo-linear processing (ML or Capon filtering) is used, then the signal-to-noise ratio is improved, but the temporal and spectral supports of the signal cannot be correctly delimited
Solution Approach 1:
The patent changes the processing approach from linear to non-linear by using conditional expectation estimators with a priori probability density functions. This parameter change in the estimation method enables simultaneous improvement of signal-to-noise ratio and accurate delimitation of temporal and spectral supports, resolving the contradiction between reliability and measurement precision.
2Device complexity
If no a priori knowledge is used, then the processing is simpler, but the signal components cannot be finely determined
Solution Approach 1:
The patent applies preliminary action by incorporating a priori probability density functions that encode prior knowledge about signal characteristics (such as Gaussian mixture models). This preliminary preparation of probabilistic information enables fine determination of signal components without excessive processing complexity, as the a priori knowledge guides the estimation process efficiently.
3Productivity
If linear filtering is applied, then the output signal is always produced, but noise components cannot be distinguished from signal components
Solution Approach 1:
The patent replaces the mechanical linear filtering system with a probabilistic non-linear estimation system using conditional expectation. This substitution introduces a decision-making mechanism that can distinguish between signal and noise components by comparing estimated values against thresholds derived from a priori probability distributions, thereby achieving both continuous output and accurate signal-noise distinction.
Data Source
AI summary
This method for the non-linear estimation of no more than two mixed signals from separate sources, the time/frequency representation of which shows an unknown non-zero proportion of zero components, using an array made up of P>2 antennas, when the directional vectors U and V of the sources emitting these signals are additionally known or estimated, includes the following steps:a) Calculating the successive discrete Fourier transforms of the signal received by the antennas and sampled to obtain a time-frequency P-vector grid of the signal; each element of the grid being referred to as a box and containing a complex vector X forming a measurement;b) For each box, calculating the conditional expectation estimator of the signal, or of the signals, from the measurement X and an a priori probability density for the signals that is a Gaussian mixture.