Voice Interaction Control System with Dynamic Model Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice interaction systems face delays in response due to communication rate limitations between terminal devices and servers, leading to unnatural conversations and insufficient stress relief for users.
Innovation Solution
A control system that initially uses a tentatively selected voice recognition model for voice interaction, determines an optimal model based on recognition results, and switches to it for improved accuracy and reduced unnaturalness, considering past usage and similarity levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If voice recognition and response generation are performed on the server side to increase accuracy, then voice recognition accuracy is improved, but response time increases due to communication rate restrictions
Solution Approach 1:
The system performs voice recognition using a locally stored model before receiving the server's recognition result. This preliminary action allows the system to generate a response without waiting for the server, significantly reducing response time while still allowing for later correction if the server's optimal model provides better accuracy
Solution Approach 2:
The system dynamically switches between local and server-side voice recognition models based on conversation context and server feedback. The calculation unit determines whether to use the local model or switch to a server-provided model, creating a flexible hybrid architecture that balances speed and accuracy
2Loss of time
If simple responses are used during delay periods to gain time, then response time is reduced, but conversation naturalness deteriorates
Solution Approach 1:
The system performs preliminary voice recognition and response generation using the local model before the delay period ends. This allows the system to provide a meaningful, natural response rather than simple filler phrases, while still reducing the overall delay by having the response ready in advance
Solution Approach 2:
The system uses feedback from the server's voice recognition results to improve subsequent responses. By incorporating server feedback about optimal models and recognition accuracy, the system can maintain natural conversation flow while progressively improving response quality
Data Source
AI summary
The control system includes a calculation unit configured to control a voice interaction system including voice recognition models, in which the calculation unit instructs, when a conversation with a target person is started, the voice interaction system to first perform voice recognition and response generation by one voice recognition model tentatively selected from among the voice recognition models, determines a voice recognition model estimated to be optimal among the voice recognition models held in the voice interaction system based on results of the voice recognition of a speech made by the target person in a voice recognition server, and instructs, when the voice recognition model estimated to be optimal is different from the one voice recognition model tentatively selected, the voice interaction system to switch the voice recognition model to the one estimated to be optimal and to perform voice recognition and response generation.