Vehicle Head Unit Speech Recognition for Multi-Seat Voice Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech recognition systems in vehicle head units struggle to accurately recognize valid commands among multiple voices in a vehicle environment, leading to potential system malfunctions due to recognition of unnecessary voices.
Innovation Solution
A method and apparatus for speech recognition that processes utterances from multiple seats, determines the intention of each utterance, and prioritizes the processing of valid commands from the intended speaker, while ignoring unnecessary voices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If microphones at all seat locations are activated to recognize voices from passengers in other seats, then voice recognition coverage is improved, but system reliability deteriorates due to recognition of unintentional speech and unexpected actions
Solution Approach 1:
The patent segments the voice recognition process by activating only specific microphones corresponding to the detected speaker's seat location. The microphone selection is segmented based on speaker detection results, allowing the system to focus on relevant voice sources while ignoring others, thus maintaining reliability while improving coverage.
Solution Approach 2:
The system performs preliminary speaker detection and seat location identification before activating microphones for voice recognition. This preliminary action determines which microphones should be activated in advance, preventing the system from processing unintentional speech from other seats while maintaining the ability to recognize valid commands from the intended speaker.
2Measurement precision
If only a microphone at the speaker's seat location is activated to better recognize the speaker's voice, then speech recognition accuracy is improved, but voice recognition coverage deteriorates as voices from passengers in other seats are not recognized
Solution Approach 1:
The system dynamically adjusts microphone activation based on real-time speaker detection results. Instead of a fixed microphone configuration, the system changes which microphones are active depending on where the speaker is located in the vehicle, thereby maintaining high recognition accuracy while adapting to different speakers and seat positions.
Solution Approach 2:
The system uses speaker detection and seat location identification as feedback to determine which microphones should be activated. This feedback loop ensures that the microphone configuration always corresponds to the actual speaker's position, maintaining both accuracy and coverage across different seating arrangements.
3Adaptability or versatility
If the system processes all utterances from multiple seats simultaneously, then voice recognition coverage is improved, but system reliability deteriorates due to recognition of unnecessary voices and potential malfunctions
Solution Approach 1:
The patent extracts and processes only the relevant utterance from the intended speaker by using speaker detection and seat location identification to filter out unnecessary voices from other seats. This extraction approach allows the system to maintain multi-seat recognition capability while processing only valid commands, preventing system malfunctions caused by unintended speech.
4Adaptability or versatility
If the system activates all microphones to recognize voices from any seat, then adaptability is improved, but device complexity increases due to managing multiple active microphones and filtering unnecessary voices
Solution Approach 1:
The system segments the microphone array into inactive and active groups based on speaker detection results. By dividing the microphones into segments and activating only those corresponding to the speaker's seat, the system reduces the complexity of managing multiple microphones while maintaining the ability to recognize voices from any seat in the vehicle.
Data Source
AI summary
A method for speech recognition in a vehicle includes processing an utterance of a first passenger, the utterance requesting execution of speech recognition, generating a command list based on the utterance, determining, based on a plurality of utterances being simultaneously received from a plurality of seats, respectively, intention of each of the plurality of utterances, and processing a speech act of a first utterance among the plurality of utterances based on the intention of each utterance.