Semantic Audio Routing for Concurrent Speech Device Command Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Concurrent use of multiple speech-controlled devices often results in unintended actions due to similar sound commands, leading to unwanted responses from multiple devices intended for a single command.
Innovation Solution
An apparatus and method for routing audio streams using semantic analysis, which includes an audio receiver, classifier, and scheduler to select a Spoken Language Understanding (SLU) engine based on semantically generated result sets, ensuring commands are directed to the appropriate device, allowing for concurrent use of multiple speech-controlled devices while minimizing accidental actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple speech-controlled devices are used concurrently, then device versatility and functionality are improved, but command ambiguity and unintended actions increase
Solution Approach 1:
The patent introduces a semantic audio routing system as an intermediary between the microphone and multiple SLU engines. This intermediary analyzes the semantic meaning of audio commands and routes them to the most appropriate SLU engine, preventing ambiguous commands from being misinterpreted by multiple devices simultaneously.
Solution Approach 2:
The system changes the parameter of command analysis from simple acoustic pattern matching to semantic meaning analysis. By transforming commands into semantic representations and using confidence scores to evaluate meaning, the system can distinguish between similar-sounding commands and route them accurately to the intended device.
2Productivity
If audio is routed to multiple SLU engines simultaneously, then command coverage and device responsiveness are improved, but bandwidth consumption and energy usage increase
Solution Approach 1:
Instead of routing audio to all possible SLU engines (excessive action), the system uses semantic analysis to identify the most relevant SLU engine(s) and routes audio only to those (partial action). This is achieved by generating semantic representations and confidence scores that indicate which SLU engines are most likely to correctly interpret the command.
Solution Approach 2:
The system performs preliminary semantic analysis of the audio command before routing it to SLU engines. By pre-processing the audio to generate semantic representations and confidence scores, the system can make informed routing decisions that avoid unnecessary bandwidth consumption while ensuring accurate command execution.
3Measurement precision
If cloud-based SLU engines are used for all commands, then processing accuracy and language understanding are improved, but latency and internet dependency increase
Solution Approach 1:
The patent implements local SLU engines that can process commands locally on the device without requiring cloud connectivity. These local engines provide rapid response for common commands while maintaining the option to use cloud-based engines for more complex processing when needed, creating a hybrid architecture that balances speed and accuracy.
Data Source
AI summary
An example apparatus for routing audio streams includes an audio receiver to receive audio from a microphone. The apparatus also includes a classifier to semantically generate a result set based on the audio. The apparatus further includes a scheduler to select a spoken language understanding (SLU) engine based on the result set. The apparatus includes a router to route the audio to the selected SLU engine.