Hearing Aid Feedback Estimation Using Deterministic Sequences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current hearing aid systems face challenges in accurately and efficiently estimating feedback paths, leading to unstable audio due to acoustic feedback, particularly in challenging environments, requiring lengthy measurement times and compromising between convergence rate and precision.
Innovation Solution
A hearing system utilizing a deterministic sequence with perfect or near-perfect autocorrelation as a probe signal for feedback estimation, allowing for rapid and precise determination of the feedback path, enabling optimal gain limitation and improved feedback management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If stochastic probe signals (white noise or colored noise) are used for feedback path estimation, then the measurement can be carried out in various acoustic environments, but the convergence rate is slow and measurement time is lengthy
Solution Approach 1:
The patent changes the fundamental parameter of the probe signal from stochastic (random) to deterministic with perfect autocorrelation properties. This parameter change transforms the signal structure to enable much faster convergence of the adaptive algorithm while maintaining the ability to operate in various acoustic environments. The deterministic sequence with perfect autocorrelation allows the feedback path estimation to converge in milliseconds rather than seconds.
2Measurement precision
If conventional adaptive algorithms (e.g., NLMS) are used with stochastic signals, then feedback path estimation can be performed, but the convergence rate is slow requiring measurement times of the order of 15 seconds
Solution Approach 1:
The patent applies parameter changes by selecting probe signals with specific deterministic autocorrelation properties (perfect or near-perfect autocorrelation). This parameter selection fundamentally improves the convergence rate of the adaptive algorithm. The perfect autocorrelation property ensures that the signal energy is concentrated in a way that maximizes the information content per sample, enabling rapid convergence without sacrificing estimation accuracy.
Solution Approach 2:
The deterministic sequences with perfect autocorrelation often exhibit periodic or structured patterns that are mathematically optimized for system identification. This periodic structure allows the adaptive algorithm to extract feedback path information more efficiently from each sample, dramatically reducing the number of samples needed for accurate estimation compared to stochastic signals.
3Stability of the object's composition
If gain reduction is applied in challenging feedback situations, then system stability is maintained, but it is often unknown how large a gain reduction is necessary leading to larger-than-needed reductions
Solution Approach 1:
The patent implements a feedback mechanism where the estimated feedback path is used to calculate the actual loop gain in real-time. This feedback information allows the system to determine the precise amount of gain reduction needed to maintain stability. Instead of applying conservative large gain reductions, the system uses the measured feedback path characteristics to compute the exact critical gain, enabling minimal necessary gain reduction while maintaining stability.
Data Source
AI summary
A hearing system comprises respective input and output transducers operationally coupled via a forward path comprising a configurable output combination unit having first and second signal inputs and a signal output. The first and second signal inputs are a signal of the forward path, and an output probe signal, respectively, and the output signal is electrically connected to the output transducer and configurable to consist of either of the first or second signal inputs, or a mixture thereof. The hearing system further comprises a configurable probe signal generator for generating the output probe signal, an adaptive feedback estimation unit for generating an estimate of an unintended feedback path, and a control unit for generating a control signal for controlling the output probe signal, which may comprise a perfect or almost perfect sequence and/or an almost perfect sweep sequence.