Smart Broadcast Management for Hearing Aids
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical devices, such as hearing aids and cochlear implants, face challenges in filtering and presenting relevant audio information from wireless broadcasts, leading to user confusion and potential safety issues due to irrelevant announcements.
Innovation Solution
A system with voice activity detection and keyword detection circuitry that segments audio streams into speech-including and speech-excluding segments, evaluates these segments for keywords, and selectively communicates relevant information to the user, allowing for tailored notification and storage of important announcements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all broadcast audio information is presented to the user, then complete information is provided, but user confusion and safety issues occur due to irrelevant announcements
Solution Approach 1:
The broadcast audio stream is divided into multiple segments through voice activity detection, which identifies speech-containing portions and non-speech portions separately. This segmentation allows the system to process and filter different types of audio content independently, presenting only relevant speech segments to the user while eliminating irrelevant background noise and non-speech announcements.
Solution Approach 2:
The system extracts specific keywords from the segmented audio data using keyword detection circuitry. By identifying and extracting only those audio segments containing predetermined keywords of interest to the user, the system filters out irrelevant information while preserving important announcements, thus preventing user confusion without losing critical information.
2Ease of operation
If keyword detection is implemented to filter announcements, then relevant information is provided, but device complexity increases
Solution Approach 1:
The broadcast receiver is integrated with voice activity detection circuitry and keyword detection circuitry to create a multi-functional device. This universal device can simultaneously receive broadcasts, detect voice activity, segment audio streams, identify keywords, and present filtered information through multiple output modalities (audio, visual, haptic), thereby managing complexity through consolidation rather than adding separate systems.
Solution Approach 2:
The voice activity detection circuitry acts as an intermediary between the raw broadcast audio and the keyword detection circuitry. It pre-processes the audio stream by identifying speech segments, which reduces the data volume and complexity for subsequent keyword analysis. This intermediary layer simplifies the overall processing pipeline by filtering out non-speech content before keyword detection occurs.
3Ease of operation
If voice activity detection and keyword detection are used, then irrelevant announcements are filtered, but processing time increases
Solution Approach 1:
Voice activity detection is performed as a preliminary action before keyword detection. By first identifying speech-containing segments and filtering out non-speech portions, the system reduces the total data volume that requires subsequent keyword analysis. This preliminary filtering accelerates the overall processing time compared to analyzing the complete unfiltered audio stream for keywords.
Solution Approach 2:
The audio stream is segmented into speech and non-speech portions, allowing parallel or sequential processing of only relevant segments. This segmentation enables the system to focus computational resources on speech-containing portions only, reducing total processing time while maintaining high information accuracy in the filtered output.
Data Source
AI summary
An apparatus includes voice activity detection (VAD) circuitry configured to analyze one or more audio broadcast streams and to identify first segments of the one or more broadcast streams in which the audio data includes speech data. The apparatus further includes derivation circuitry configured to receive the first segments and, for each first segment, to derive one or more words from the speech data of the first segment. The apparatus further includes keyword detection circuitry configured to, for each first segment, receive the one or more words and to generate keyword information indicative of whether at least one word of the one or more words is among a set of stored keywords. The apparatus further includes decision circuitry configured to receive the first segments, the one or more words of each of the first segments, and the keyword information for each of the first segments and, for each first segment, to select, based at least in part on the keyword information, among a plurality of options regarding communication of information indicative of the first segment to a recipient.