Adaptive Filtering for Blind Source Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In environments with multiple simultaneous sources, effective blind source separation from sensor response mixtures becomes difficult as the number of sources increases, especially when the true number of sources is unknown and changing over time.
Innovation Solution
The implementation of an adaptive filtering architecture that leverages non-sensor information to validate source hypotheses, extract estimated representations of source signals, and improve the separation of hidden source signals from sensor response mixtures, referred to as Only Mostly Blind Source Separation (OMBSS). This involves modeling environmental transfer functions as filters and using source hypothesis signals to enhance the separation process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If blind source separation is applied to sensor response mixtures, then source signals can be separated, but the separation effectiveness deteriorates as the number of sources increases
Solution Approach 1:
The patent applies preliminary action by using non-sensor information (such as source location, signal characteristics, or environmental data) to form hypotheses about source signals before performing blind source separation. This pre-processing step creates initial estimates of source signals that guide the subsequent separation process, making it more effective even when the number of sources is large or unknown.
Solution Approach 2:
The patent introduces non-sensor information as an intermediary element that mediates between the sensor responses and the source separation process. This intermediary information provides additional constraints and guidance that improve separation effectiveness without requiring direct modification of the sensor array or signal processing architecture.
2Measurement precision
If blind source separation is applied without prior information, then the method remains simple and general, but the number of sources cannot be determined and separation quality deteriorates
Solution Approach 1:
The patent performs preliminary analysis using non-sensor information to form hypotheses about source signals before the main separation process. This allows the system to leverage available information (such as spatial location, frequency characteristics, or temporal patterns) to improve separation quality without requiring complete prior knowledge of all source parameters.
Solution Approach 2:
The patent changes the parameter set by incorporating non-sensor information parameters (such as source location, signal power, or spectral characteristics) into the separation process. These additional parameters provide constraints that improve separation quality while maintaining the adaptability of the blind source separation framework.
3Measurement precision
If all source components are processed through blind source separation, then complete separation is achieved, but processing time increases
Solution Approach 1:
The patent segments the source separation process into two parts: sources that can be identified using non-sensor information (and thus require minimal processing) and sources that require full blind source separation. This segmentation allows the system to process only the necessary subset of sources through the computationally intensive BSS algorithm, reducing overall processing time while maintaining separation completeness.
Solution Approach 2:
The patent applies partial action by using non-sensor information to partially identify and separate certain source signals before applying blind source separation to the remaining components. This reduces the dimensionality and complexity of the BSS problem, achieving complete separation with less processing time than applying BSS to all sources simultaneously.
Data Source
AI summary
In environments (such as acoustic and bioelectrical environments) characterized by multiple simultaneous sources, effective blind source separation from sensor response mixtures becomes difficult as the number of sources increases-especially when the true number of sources is both unknown and changing over time. However, in some environments, non-sensor information can provide useful hypotheses for some sources. Embodiments of the present invention provide an adaptive filtering architecture for validating such source hypotheses, extracting an estimated representation of source signals corresponding to valid hypotheses, and improving the separation of the remaining “hidden” source signals from the sensor response mixtures.