Convex Optimization for Robust Beamforming in Hearing Aids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hearing aid technologies face challenges in improving speech intelligibility in noise environments due to steering vector mismatches and high computational complexity, which affects the robustness and efficiency of adaptive beamforming algorithms.
Innovation Solution
A robust adaptive beamforming algorithm using convex optimization with a logarithmic barrier function is implemented, which simplifies computations and maintains fast convergence, making it suitable for real-time processing in hearing aids despite changing sound fields and steering vectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive beamforming algorithms are used to improve speech intelligibility in noise, then speech intelligibility is improved, but computational complexity increases beyond hearing aid capabilities
Solution Approach 1:
The patent extracts and eliminates the steering vector constraint from the beamforming optimization problem. By formulating the problem without requiring accurate steering vector knowledge, it removes the computational burden of steering vector estimation and mismatch handling, achieving robust beamforming with reduced complexity suitable for hearing aids
Solution Approach 2:
The patent changes the optimization parameter from minimizing output power (traditional MVDR) to maximizing the ratio of output power to noise power. This parameter transformation simplifies the mathematical formulation and enables efficient implementation without requiring precise steering vector information, making it feasible for hearing aid processors
2Object-affected harmful factors
If steering vector estimation methods are used to avoid target cancellation, then target cancellation is reduced, but manufacturing precision and reliability deteriorate due to subject variability
Solution Approach 1:
The patent removes the steering vector parameter from the beamforming formulation entirely. By using a constraint-based approach that protects the target region without requiring steering vector estimation, it eliminates the source of inaccuracies related to subject variability and ensures consistent performance across different users
Solution Approach 2:
The patent implements a protective constraint that cushions the target region against cancellation before any steering vector mismatch can occur. The optimization explicitly constrains the beamformer to maintain adequate gain in the target region, providing robustness against steering vector errors without needing to estimate or compensate for them
3Object-affected harmful factors
If adaptation range is limited to reduce target cancellation, then target cancellation is reduced, but speech intelligibility benefit is reduced
Solution Approach 1:
The patent applies local quality by implementing a directional constraint that provides different treatment to different spatial regions. The target region receives protective constraints to prevent cancellation, while other regions are freely adapted for noise reduction, achieving both protection and performance optimization
Data Source
AI summary
Disclosed herein, among other things, are methods and apparatus for improving speech intelligibility for speech-in-noise in audio processing and hearing assistance devices. The present subject matter includes a method for improving speech intelligibility for speech-in-noise for audio processing and hearing assistance devices. The method includes receiving an audio signal using a microphone array and processing the received signal to improve speech intelligibility in noise. A barrier-type beamforming process is used to improve signal-to-noise ratio at the output of the microphone array. The beamforming process includes convex optimization using a logarithmic barrier function, according to various embodiments.