Hearing Device Classifier Using Movable and Fixed Clusters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hearing aids face challenges in adapting to unforeseen listening situations due to differences in input vector distributions between users and manufacturers, leading to unstable behavior in mixed situations and the inability to handle new situations effectively, with existing solutions either being complex or failing to address user input inconsistencies.
Innovation Solution
A method using movable and fixed clusters in a multidimensional space, where the movable cluster shifts based on input signals to adjust device settings, with adjustment variables determined through neighborhood-based regression or recursive updating, allowing for smooth adaptations without frequent user intervention and maintaining basic device behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the hearing device uses fixed clusters for device settings, then the basic device behavior is preserved, but the device cannot adapt to unforeseen listening situations
Solution Approach 1:
The patent implements both fixed clusters (for stability) and movable clusters (for adaptability). The movable clusters dynamically adjust their positions in the input space based on incoming signals, allowing the device to adapt to new listening situations while the fixed clusters maintain the basic reliable behavior. This dynamic structure resolves the contradiction between adaptability and stability.
Solution Approach 2:
The patent segments the cluster structure into two distinct types: fixed clusters that remain stationary and provide stable baseline behavior, and movable clusters that can shift positions to adapt to new situations. This segmentation allows the system to simultaneously maintain reliability through fixed clusters and achieve adaptability through movable clusters.
2Adaptability or versatility
If the movable cluster shifts frequently to adapt to new input signals, then the adaptability improves, but the computational effort increases
Solution Approach 1:
The patent employs periodic updating of movable cluster positions rather than continuous adjustment. The clusters are shifted at specific intervals or triggered by certain conditions, which reduces the overall computational burden while still maintaining good adaptability to changing listening situations.
Solution Approach 2:
Instead of continuously optimizing all cluster parameters, the patent applies partial adjustments to movable clusters only when necessary. The clusters are shifted partially towards new input vectors rather than completely repositioning, which reduces computational effort while maintaining sufficient adaptability.
3Speed
If the device adapts quickly to new situations, then the responsiveness improves, but the instability in mixed situations increases
Solution Approach 1:
The movable clusters act as intermediaries between the fixed clusters and new input signals. When a new situation is detected, the movable cluster gradually shifts towards the new input vector, mediating the transition and preventing abrupt changes in parameter values. This intermediate adjustment mechanism provides both responsiveness and stability.
Solution Approach 2:
The patent prepares the movable clusters in advance to absorb and smooth out transitions when new situations occur. The gradual shifting mechanism cushions against abrupt parameter changes that would otherwise cause instability, allowing the device to respond quickly while maintaining stable parameter values.
Data Source
AI summary
The method involves determining a feature vector (e) from an input signal (10) of an apparatus. A movable cluster and a fixed cluster are provided in a multi-dimensional space, where the fixed cluster is provided at a fixed cluster position (12) in the multi-dimensional space. The movable cluster is displaced in a direction of the feature vector to movable cluster positions (13, 14). Two modifiable labels (15, 16) are respectively assigned to the movable cluster and the fixed cluster. The apparatus is set based on the fixed and movable cluster positions and the two modifiable labels. The input signal is an audio signal. Independent claims are also included for the following: (1) a classifier for automatically adjustable apparatus (2) a hearing device.