Speech Recognition Error Handling via Alternative Name Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech recognition systems for car navigation often experience erroneous recognition when users attempt to input facility names not registered in the database, leading to user frustration and decreased usability.
Innovation Solution
A speech recognition apparatus with a comparison section that identifies repeated erroneous inputs and prompts users to try alternative names by displaying a list of nearby facilities or connecting to an external server for byname searches, allowing users to input registered names.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the speech recognition apparatus uses a fixed database for speech recognition, then the recognition accuracy for registered facilities is improved, but the system fails to recognize unregistered facility names, leading to repeated erroneous recognition
Solution Approach 1:
The system performs preliminary comparison of speech data patterns before final recognition. When a facility name is not found in the database, the system proactively analyzes the speech pattern against known patterns to identify potential matches, rather than simply rejecting unrecognized names. This preliminary analysis action enables the system to handle unregistered facilities while maintaining accuracy for registered ones.
2Ease of operation
If the speech recognition apparatus continues to process repeated erroneous inputs, then the user has multiple opportunities to correct their input, but the user experiences frustration and may give up using the system
Solution Approach 1:
The system implements feedback by comparing repeated speech inputs and providing intelligent responses. When the same facility name is recognized as erroneous multiple times, the system feeds back this pattern information and offers alternative suggestions or corrections, rather than simply repeating the same erroneous recognition. This feedback mechanism maintains user confidence by showing the system is learning and adapting to user input patterns.
Solution Approach 2:
The system takes preliminary anti-action by detecting repeated erroneous recognition patterns and preemptively intervening before the user becomes frustrated. When the system identifies that the same facility name has been unsuccessfully recognized multiple times, it proactively provides alternative names or corrections, preventing the harmful effect of repeated failures and user frustration before they occur.
3Measurement precision
If the speech recognition apparatus uses strict matching against the database, then the recognition precision for registered names is improved, but the system cannot accommodate variations or alternative names of facilities
Solution Approach 1:
The system applies parameter changes by adjusting the recognition criteria based on the situation. For registered facilities, it uses strict matching to ensure precision. For unregistered or repeatedly erroneous inputs, it transitions to a more flexible parameter set that allows pattern matching and alternative name recognition. This dynamic parameter adjustment enables the system to maintain exact name recognition while accommodating variations and alternative names.
Data Source
AI summary
A speech recognition apparatus is disclosed. The apparatus converts a speech signal into a digitalized speech data, and performs speech recognition based on the speech data. The apparatus makes a comparison between the speech data inputted the last time and the speech data inputted the time before the last time in response to a user's indication that the speech recognition results in erroneous recognition multiple times in a row. When the speech data inputted the last time is determined to substantially match the speech data inputted the time before the last time, the apparatus outputs a guidance prompting the user to utter an input target by calling it by another name.