Beamformer Stabilization via Interference Delay Regularization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing beamformers, such as MVDR and LCMV, tend to distort signals from directions other than the principal direction of arrival due to fluctuations in interfering sources and signal spectrum changes, leading to undesirable amplification of low-powered signals and instability in the beampattern.
Innovation Solution
The method involves determining an interference delay pattern for interfering signals, generating regularization signals with matching delay patterns, and using these signals to adapt beamformer coefficients, ensuring stable attenuation of interfering signals and minimizing distortion by injecting artificial noise that resembles the interfering signals and varies smoothly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If adaptive beamformers are used to minimize output signal energy, then interfering audio signals are suppressed, but the beamformer distorts output signals and varies over time causing audible distortion
Solution Approach 1:
The patent applies preliminary action by regularizing the covariance matrix before inversion to prevent ill-conditioning. Spectral regularization techniques are applied in advance to ensure the matrix is well-conditioned, which prevents subsequent beamformer distortion and instability. This preliminary regularization step ensures stable beamformer coefficients are calculated from the start.
2Adaptability or versatility
If beamformer coefficients are adapted in real-time to changing conditions, then the beamformer tracks interfering sources, but sudden changes cause audible distortion in the output
Solution Approach 1:
The patent applies dynamics by implementing smooth transitions in beamformer coefficients over time. The regularization approach ensures that as the beamformer adapts to changing interference conditions, the coefficient variations are controlled and gradual rather than sudden. This dynamic adjustment prevents audible distortion while maintaining adaptability to moving interfering sources.
3Productivity
If the covariance matrix is inverted without regularization, then computation is faster, but the matrix may be ill-conditioned leading to unstable beamformer behavior
Solution Approach 1:
The patent applies parameter changes by modifying the covariance matrix through spectral regularization before inversion. By adjusting the eigenvalues or adding regularization terms, the matrix becomes well-conditioned while maintaining computational efficiency. This parameter transformation ensures stable beamformer behavior without significantly increasing computation time.
Data Source
AI summary
Method, device and computer program product for processing signals at the device. Signals are received, over a range of angles, at a plurality of sensors of the device, the received signals including an interfering signal received from an interfering source location. An interference delay pattern between receipt of signals at the sensors corresponding to receipt of a signal from the interfering source location is determined. A plurality of regularization signals having a delay pattern matching the determined interference delay pattern are generated. The generated regularization signals are used to determine beamformer coefficients to be applied by a beamformer, and the beamformer applies the determined beamformer coefficients to the signals received by the plurality of sensors, thereby generating a beamformer output.