Blind Noise Power Estimation via Eigenvalue Decomposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise estimation methods in multi-signal environments are ineffective when signals contaminate the data record, leading to inaccurate noise power estimation and challenges in adaptive modulation systems, especially in scenarios with high signal-to-noise ratios and continuous information-bearing signals.
Innovation Solution
A blind noise estimation system that forms a temporal covariance matrix of consecutive data samples, using eigenvalue decomposition to separate noise power from signal power, allowing for adaptive modulation and robust signal detection without requiring array processing or specialized resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional outlier rejection schemes are used for noise estimation, then noise power estimation is accurate in high SNR scenarios, but the method fails when data records are contaminated with signal components
Solution Approach 1:
The data record is segmented into multiple blocks, and the covariance matrix is divided into signal subspace and noise subspace components. This segmentation allows separate estimation of signal and noise powers, enabling accurate noise estimation even when the entire record contains signal contamination.
Solution Approach 2:
The signal components are extracted and removed from the data record before performing noise estimation. By taking out the signal subspace spanned by the estimated signal eigenvectors, the remaining noise subspace can be used for accurate noise power estimation without signal contamination.
2Measurement precision
If array processing techniques are used to aid noise estimation, then measurement accuracy improves, but device complexity and resource requirements increase
Solution Approach 1:
The single sensor performs noise estimation using its own received data without requiring additional sensors or array processing resources. The method uses the covariance structure of the received signal to self-determine the signal subspace and extract noise information, making the system self-sufficient.
Solution Approach 2:
The method changes the parameter being estimated from the traditional single noise power value to a structured covariance matrix with distinct signal and noise subspaces. This parameter transformation enables separation of signal and noise components using eigenvalue decomposition, achieving array-processing-level accuracy with scalar processing.
3Measurement precision
If specialized resources are reserved for noise estimation, then estimation accuracy improves, but system productivity and resource utilization decrease
Solution Approach 1:
The received data record serves multiple functions: it is used both for signal detection and for noise estimation. The same data blocks that are processed for signal identification are also utilized to compute the covariance matrix and estimate noise power, eliminating the need for separate dedicated estimation resources.
Solution Approach 2:
Instead of using the entire data record for signal processing, a portion (N consecutive samples) is used for covariance matrix computation and noise estimation. This partial use of data for dual purposes (signal and noise analysis) maintains throughput while enabling accurate noise estimation.
Data Source
AI summary
A system estimates noise power in a scalar, multi-signal communications channel. A data sampler collects N data samples from communications signals received from the communications channel. A module forms a covariance matrix of the N data samples based on a model order estimate. A module also computes the eigenvalue decomposition of the covariance matrix and ranks resultant eigenvalues from the minimum to the maximum for determining the noise power.