Noise Spatial Correlation Matrix Weighting for Voice Direction Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice recognition technologies face challenges in accurately estimating the direction of a voice command in noisy environments due to the dependence on inappropriate noise segments, which can significantly decrease estimation accuracy.
Innovation Solution
An information processing apparatus and method that calculates a first weight based on the degree to which an acoustic signal is obtained from stationary noise and applies this weight to a noise spatial correlation matrix, improving the accuracy of noise segmentation and subsequent voice direction estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a noise segment is employed to calculate the noise spatial correlation matrix, then the direction of arrival estimation can be improved, but the estimation accuracy significantly decreases when the employed noise segment is inappropriate
Solution Approach 1:
The patent applies preliminary action by evaluating the appropriateness of the noise segment before using it to calculate the noise spatial correlation matrix. The evaluation unit assesses whether the acoustic signal in the noise segment is stationary noise, and only segments that pass this evaluation are used. This preliminary check prevents inappropriate segments from degrading the estimation accuracy, thus resolving the contradiction between improving direction of arrival estimation and maintaining reliability.
2Adaptability or versatility
If various kinds of noise exist in the real environment, then the voice recognition capability is affected, but the direction of speaker estimation accuracy is lowered
Solution Approach 1:
The patent applies the taking out principle by extracting and removing the noise component from the acoustic signal processing. The system separately calculates the noise spatial correlation matrix from identified noise segments and uses this to eliminate noise effects from the direction of arrival estimation. By extracting the noise characteristic and removing its influence, the system maintains speaker direction estimation accuracy even in noisy environments, resolving the contradiction between adaptability to noisy environments and measurement precision.
Data Source
AI summary
An information processing apparatus including an acquisition section that acquires acquire an acoustic signal indicating a sound collection result of a sound collection device group and a control section that calculates a first weight in accordance with a degree to which the acoustic signal acquired by the acquisition section is a signal obtained by observing stationary noise and to apply the first weight to a noise spatial correlation matrix that is a spatial correlation matrix obtained from a noise signal.