Vehicle Speech Recognition Segmentation for Navigation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice recognition systems in vehicles face challenges with non-standard and specialized instructions, leading to lower success rates and user frustration due to the unpredictability of commands, particularly in navigation tasks.
Innovation Solution
A method and system that involves receiving speech at a vehicle microphone, determining if it includes navigation instructions, and sending it to a remote facility for corrective action when local recognition fails, allowing remote personnel to provide assistance and improve speech recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the voice recognition system handles a large number of non-standard or specialized navigation instructions locally, then the system can respond faster, but the speech recognition success rate decreases due to the varied and unpredictable commands
Solution Approach 1:
The system segments speech recognition tasks by type: standard commands are processed locally by the vehicle's speech recognition system for fast response, while non-standard or uncertain commands (particularly navigation instructions with low confidence scores) are segmented out and forwarded to a remote facility for specialized processing. This segmentation allows the local system to maintain high success rates for common tasks while the remote system handles the varied and unpredictable navigation commands.
Solution Approach 2:
A remote facility acts as an intermediary between the vehicle's local speech recognition system and the ultimate command execution. The remote facility receives uncertain or non-standard navigation instructions, processes them with specialized capabilities, and returns results to the vehicle system. This intermediary approach enables the local system to maintain fast response times for standard commands while leveraging remote expertise for difficult cases.
2Reliability
If the system forwards all uncertain speech to a remote facility, then the speech recognition accuracy improves, but the response time increases and user frustration increases
Solution Approach 1:
Instead of forwarding all speech commands to the remote facility, the system applies partial action by selectively forwarding only those commands that fall below a confidence threshold or are identified as non-standard navigation instructions. Common, well-understood commands are processed completely locally without remote intervention, maintaining fast response times while still improving accuracy for uncertain cases through targeted remote processing.
3Adaptability or versatility
If the system uses a large vocabulary for navigation instructions, then the system can handle more commands, but the complexity of the speech recognition system increases
Solution Approach 1:
The system segments the vocabulary and processing responsibilities: the local vehicle system maintains a core vocabulary for standard commands, while the remote facility handles the extended vocabulary for non-standard navigation instructions. This segmentation allows the local system to remain relatively simple while still achieving high versatility through the combined local-remote capability.
Data Source
AI summary
A method for recognizing speech in a vehicle includes receiving speech at a microphone installed to a vehicle, and determining whether the speech includes a navigation instruction. If the speech includes a navigation instruction, the speech may be sent to a remote facility. After sending the speech to the remote facility, a local speech recognition result is provided in the vehicle to the user. The speech sent to the remote facility may be used to provide corrective action. A system for recognizing speech in a vehicle may include a microphone, and may be configured to determine a local speech recognition result from the speech command and determine when the speech command includes a navigation instruction. The system may further include a remote server in communication with the vehicle that receives a sample of the speech command from the speech recognition system when the speech command includes a navigation instruction.