Sound Direction Estimation Using Dual Correlation Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sound direction estimation devices face challenges in accurately estimating the direction of a target sound due to varying noise levels and frequency characteristics, as they rely on limited noise correlation matrices or sound feature values, which are not sufficient to exclude noise influence in real environments.
Innovation Solution
A sound direction estimation device that calculates correlation matrices for both input sound signals and noise signals, using a spatial spectrum based on these matrices to estimate the direction of a sound source, with features like delayed noise signals and eigenvector calculation through generalized eigenvalue decomposition, allowing for improved noise separation and adaptation to target sound variations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If noise correlation matrices or sound feature values are acquired in advance for each type of sound source, then the device complexity is reduced, but the measurement precision of sound direction estimation deteriorates due to varying noise levels and frequency characteristics in real environments
Solution Approach 1:
The patent dynamically adapts the noise correlation matrix to current environmental conditions by calculating it from recent sound signals rather than using fixed pre-acquired values. This allows the system to track changes in noise levels and frequency characteristics over time, maintaining estimation accuracy without requiring complex pre-processing for every possible noise condition
Solution Approach 2:
The patent changes the parameters of the noise correlation matrix based on current sound signal characteristics. By recalculating the noise correlation matrix from recent sound signals and adapting it to current noise conditions, the system maintains accuracy across varying environments without increasing overall device complexity
2Ease of operation
If a limited number of noise correlation matrices are used, then the ease of operation is improved, but the reliability of sound direction estimation deteriorates when noise characteristics vary from pre-acquired values
Solution Approach 1:
The patent performs preliminary calculation of the noise correlation matrix from recent sound signals before using it for direction estimation. This preliminary action adapts the noise model to current conditions, ensuring reliability when noise characteristics differ from any pre-acquired values, while maintaining ease of operation through automated calculation
3Measurement precision
If the time window for calculating correlation matrices is extended, then the measurement precision improves through better noise separation, but the loss of time increases due to longer processing requirements
Solution Approach 1:
The patent uses a partial time window approach, calculating the noise correlation matrix from a limited number of recent sound signals rather than extending over a long period. This provides sufficient noise separation for accurate estimation while avoiding excessive processing time, achieving the optimal balance between precision and speed
Data Source
AI summary
A sound direction estimation device includes a first correlation matrix calculation unit configured to calculate a correlation matrix of a plurality of channels of input sound signals, a second correlation matrix calculation unit configured to calculate a correlation matrix of noise signals based on the plurality of channels of sound signals, and a sound source localization unit configured to calculate a spatial spectrum based on the correlation matrix calculated by the first correlation matrix calculation unit and the correlation matrix calculated by the second correlation matrix calculation unit and to estimate a direction of a sound source associated with the plurality of channels of sound signals using the calculated spatial spectrum.