In-Vehicle Voice Control Using Monosyllabic Phoneme Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice control systems in vehicles struggle with reliable recognition of user-spoken terms from large databases, especially when users use free-form phrasing or when databases change, requiring significant storage, recompilation, and internet connectivity, and limiting flexible input sequences.
Innovation Solution
A motor vehicle system using a phoneme generation module with a statistical language model for monosyllabic phonemes and a phoneme-to-grapheme module via a neural network to convert speech inputs into vehicle commands, reducing the search space and enabling flexible input sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If precompiled language models are used to enable recognition of terms from large databases, then recognition capability is improved, but storage requirements increase significantly
Solution Approach 1:
The patent segments the large language model into a hierarchical structure with a phoneme dictionary at the base level and multiple specialized databases (navigation, media, messaging) built upon it. This segmentation allows the system to store only essential phoneme representations rather than complete language models for each domain, significantly reducing storage requirements while maintaining recognition capability across multiple domains.
Solution Approach 2:
The patent creates a universal phoneme dictionary that serves as a reusable template for building multiple domain-specific databases. Instead of storing separate complete language models for navigation, media, and messaging, the system copies and adapts the phoneme dictionary structure for each domain, eliminating redundant storage and enabling efficient recognition across diverse term sets.
2Reliability
If precompiled language models are used to recognize terms from large databases, then recognition capability is improved, but adaptability to database changes deteriorates
Solution Approach 1:
The patent implements a dynamic database structure where the phoneme dictionary and domain-specific databases can be easily updated. When source databases change (e.g., new addresses, media titles, or messaging terms are added), the system can dynamically regenerate the affected phoneme representations and update the corresponding domain databases without requiring complete recompilation of the entire language model, thus maintaining adaptability.
3Measurement precision
If online connection is used to access updated language models, then recognition accuracy is improved, but system reliability deteriorates due to dependency on internet connectivity
Solution Approach 1:
The patent performs preliminary compilation of the phoneme dictionary and domain-specific databases during system initialization or offline updates. Once compiled, these data structures are stored locally in the vehicle's system memory, enabling the speech recognition system to operate independently of internet connectivity. The preliminary preparation of phoneme representations ensures high recognition accuracy while eliminating dependency on continuous online access.
4Reliability
If rigid term sequences from database are required for recognition, then recognition reliability is improved, but ease of operation deteriorates
Solution Approach 1:
The patent transitions from requiring exact word-matching in a fixed sequence to analyzing phoneme sequences in a hierarchical structure. By organizing the recognition system around phoneme-level representations and using hierarchical matching across multiple databases, the system can recognize terms regardless of their position in the input sequence or slight variations in pronunciation, thereby improving ease of operation while maintaining reliability.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
The invention relates to a motor vehicle, wherein: the motor vehicle comprises a navigation system and an operating system which is connected to the navigation system for data transmission via a bus system; the motor vehicle has a microphone; the motor vehicle comprises a phoneme generation module for generating phonemes from an acoustic voice signal or the output signal of the microphone; the phonemes are part of a predefined selection of exclusively monosyllabic phonemes; and the motor vehicle comprises a phoneme-to-grapheme module for generating inputs to operate the motor vehicle depending on monosyllabic phonemes generated by the phoneme generation module.