Adaptive Beamforming for Noise Attenuation in Low SNR Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Beamforming performance is degraded in low Signal-to-Noise Ratio (SNR) environments, particularly due to reverberation and inaccurate covariance information, affecting both signal independent and dependent beamforming techniques.
Innovation Solution
An electronic device adaptively performs signal dependent beamforming by determining first and second parameters based on SNR values, using covariance information and phase differences to attenuate noise signals effectively in different frequency bands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If signal independent beamforming is performed to compensate for phase delay based on distance and direction information, then beamforming can be performed without being affected by SNR, but beamforming performance is degraded when direction information changes or reverberation occurs indoors
Solution Approach 1:
The patent applies dynamics by transitioning from static signal-independent beamforming to dynamic signal-dependent beamforming that adapts to changing acoustic environments. The system continuously estimates covariance matrices and updates beamforming weights based on current SNR conditions, allowing the beamformer to dynamically adjust to direction changes and reverberation characteristics.
Solution Approach 2:
The patent changes key parameters including SNR thresholds, covariance matrix estimates, and beamforming weights based on environmental conditions. By monitoring SNR levels and adjusting beamforming parameters accordingly, the system optimizes performance across varying acoustic scenarios including reverberant indoor environments and moving sound sources.
2Adaptability or versatility
If signal dependent beamforming is performed using covariance information to compensate for phase delay, then beamforming is not affected by direction information or reverberation, but accuracy of covariance information is reduced in low SNR environments
Solution Approach 1:
The patent applies preliminary action by performing covariance matrix estimation and SNR calculation before final beamforming weight computation. The system pre-processes input signals to estimate acoustic environment characteristics, then uses these estimates to guide subsequent beamforming operations, ensuring accurate covariance information even in challenging acoustic conditions.
Solution Approach 2:
The patent implements feedback mechanisms where beamforming output is continuously monitored and fed back to update covariance matrix estimates. This iterative process refines the accuracy of covariance information by using actual observed signal characteristics, allowing the system to adapt to low SNR conditions and improve measurement precision over time.
3Object-generated harmful factors
If beamforming is performed to remove noise signals from input signals, then speech signal is strengthened, but call quality and speech recognition rate are reduced when noise signal is present with speech signal
Solution Approach 1:
The patent applies local quality by implementing frequency-dependent beamforming where different frequency bands are processed with different weights and parameters. The system identifies speech-dominated frequency regions and applies aggressive noise attenuation there, while preserving frequencies where speech components are weak, thereby maintaining call quality and speech recognition rate while effectively removing noise.
Solution Approach 2:
The patent replaces traditional mechanical or fixed beamforming approaches with adaptive signal processing that uses statistical properties of the input signals. By substituting fixed phase compensation with dynamic covariance-based beamforming, the system achieves superior noise attenuation while preserving speech quality through intelligent signal separation rather than simple filtering.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device is disclosed. In addition, various embodiments recognized through the specification are possible. The electronic device includes a plurality of input devices for receiving a plurality of input signals, each of which includes a speech signal and a noise signal and a processor electrically connected with the input devices. The processor is configured to determine a signal to ratio (SNR) value for the plurality of input signals for each frequency band, determine a first parameter indicating a phase shift to frequency in the plurality of input signals in a first frequency band where the SNR value is greater than or equal to a specified threshold, determine a second parameter indicating a phase shift to frequency in the plurality of input signals in a second frequency band where the SNR value is less than the threshold, and perform beamforming for the plurality of input signals based on the first parameter and the second parameter.