Navigation Speech Processing with Region Language Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Navigation systems face challenges in providing accurate and efficient voice processing due to limited mobile device resources, noise, and the need for low latency, especially when handling large vocabularies and diverse user inputs, which affects usability and accuracy.
Innovation Solution
A navigation system that employs a speech processing mechanism using a region language model and sub-region search grammar to recognize user inputs, allowing for flexible parsing and generation of location identifiers from spoken commands, enabling efficient decoding and interpretation of user requests in a single utterance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If voice processing uses a large set of vocabularies to support diverse user inputs, then adaptability improves, but accuracy deteriorates due to processing constraints and noise
Solution Approach 1:
The patent segments the voice processing task into multiple stages: initial vocabulary filtering based on context, followed by targeted speech recognition. This divides the large vocabulary processing into manageable segments, maintaining adaptability while improving accuracy by focusing computational resources on relevant terms only.
Solution Approach 2:
The system performs preliminary actions by pre-processing the spoken input to identify context clues and filter the vocabulary set before full speech recognition. This preliminary filtering action reduces the effective vocabulary size, improving recognition accuracy while maintaining support for diverse inputs.
2Measurement precision
If the system processes speech in multiple attempts with prompts, then accuracy improves, but latency increases
Solution Approach 1:
The patent applies preliminary action by performing context-based vocabulary filtering before speech recognition, rather than waiting for multiple user attempts. This preparatory step enables the system to achieve higher accuracy in a single pass, reducing latency by eliminating the need for repeated prompts and attempts.
3Measurement precision
If the system uses constrained speech formats to improve recognition accuracy, then measurement precision improves, but ease of operation deteriorates
Solution Approach 1:
The patent implements dynamics by making the speech processing approach adaptive rather than fixed. The system dynamically adjusts between constrained and unconstrained processing modes based on the specific input and context, allowing users to speak naturally while maintaining high recognition accuracy through context-aware vocabulary filtering.
4Measurement precision
If the system filters vocabulary based on context to improve accuracy, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent segments the complex vocabulary filtering process into distinct, manageable components: context analysis module, vocabulary filtering module, and speech recognition module. This segmentation reduces device complexity by organizing the processing mechanism into independent, reusable components with clear interfaces.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A method of operation of a navigation system includes: receiving a single utterance of a spoken input; generating a search region from the spoken input with a region language model; and generating a location identifier based on a sub-region search grammar and the search region for displaying on a device.