Adaptive Filter Feedback Suppression in Hearing Aids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hearing aid devices experience feedback issues due to the close proximity of microphone and output transducers, leading to unwanted howling sounds, which existing methods struggle to effectively suppress, especially when the feedback path's gain is greater than 1 and phase shift is optimal for oscillation.
Innovation Solution
A method that involves determining a first feedback transfer function, forming a weighted mean value function from amplitude values, and using an adaptive filter to estimate a second feedback transfer function, with coefficients updated based on impulse response parameters to control adaptation speed, ensuring quick adaptation where energy is high and stability where energy is low, thereby effectively suppressing feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a fixed step size is used in the adaptive filter, then the filter either adapts quickly (large step size) or maps the input function better with small changes (small step size), but it cannot simultaneously achieve both fast adaptation and high precision mapping
Solution Approach 1:
The patent applies dynamics by making the step size variable rather than fixed. The step size is adapted dynamically based on the energy distribution across different time delays in the feedback impulse response. This allows the adaptive filter to use larger step sizes for high-energy components (fast adaptation) and smaller step sizes for low-energy components (precise mapping), resolving the contradiction between speed and precision.
Solution Approach 2:
The patent applies local quality by assigning different step sizes to different coefficients of the adaptive filter based on their corresponding time delays. Each coefficient is updated with a step size tailored to its specific energy level in the feedback path, rather than using a uniform step size. This localized adaptation enables fast convergence for dominant feedback paths while maintaining precision for weaker paths.
2Adaptability or versatility
If the adaptive filter updates coefficients rapidly to follow changes in feedback path, then it can adapt quickly to new conditions, but it may produce artifacts due to incorrect adaptation when excitation energy is low
Solution Approach 1:
The patent changes the parameter of step size based on the energy level at different time delays. By modifying the step size parameter dynamically according to the feedback impulse response energy distribution, the system achieves high adaptability for strong feedback components while maintaining reliability for weak components through reduced step sizes that prevent incorrect adaptation.
3Device complexity
If uniform weighting is applied to all coefficients in the adaptive filter, then the update process is simple, but coefficients associated with larger time delays are updated too aggressively causing instability
Solution Approach 1:
The patent applies local quality by assigning different weighting factors to different coefficients based on their time delay positions. Coefficients associated with larger time delays receive smaller weights, preventing aggressive updates and maintaining stability, while coefficients for shorter delays can be updated more aggressively. This localized weighting strategy improves stability without significantly increasing computational complexity.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The invention relates to a method and a device for reducing feedback in a hearing aid. The method comprises the step of acquiring a first feedback transfer function at a first time point along a feedback path from a signal processing unit via an electro-acoustic transducer, an acoustic signal path from the electro-acoustic transducer to an acousto-electric transducer, and via the acousto-electric transducer back to the signal processing unit. In a further step, a weighted average value function is determined as a function of the amplitude values of the first feedback transfer function. A second feedback transfer function is estimated using an adaptive filter, wherein the coefficients of the adaptive filter are determined as a function of the weighted average value function.The adaptive filter is applied to a signal derived from an acoustic input signal of the acousto-electric transducer.