Speech Processing System Wakeword Timing Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech-processing systems face challenges in accurately determining the intended speech-processing system for utterances containing multiple wakewords, leading to incorrect processing and response generation, as they often rely solely on the first detected wakeword without considering the user's intended intent.
Innovation Solution
The system determines the appropriate speech-processing system by analyzing the timing and context of multiple wakewords in an utterance, using models like classifiers and neural networks to select the system associated with the later-uttered wakeword, ensuring that the user's intended command is processed by the correct system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the system relies solely on the first detected wakeword to determine the speech-processing system, then the processing speed is fast, but the accuracy of identifying the intended system is low
Solution Approach 1:
The system dynamically determines which speech-processing system to use based on the temporal sequence and contextual analysis of multiple wakewords detected in the audio stream. Instead of statically assigning to the first wakeword, the system adapts its decision based on the pattern of wakeword appearances and their timing relationships, resolving the contradiction between speed and accuracy through dynamic decision-making.
Solution Approach 2:
The system performs preliminary detection and analysis of multiple wakewords before finalizing the speech-processing system selection. By pre-identifying all wakewords in the audio stream and analyzing their temporal patterns, the system prepares the information needed for accurate system determination without delaying the overall processing timeline.
2Reliability
If the system analyzes multiple wakewords and their timing context to determine the intended system, then the accuracy of processing the correct command is improved, but the processing time increases
Solution Approach 1:
The system uses feedback from analyzing the temporal patterns and contextual relationships between multiple wakewords to refine its determination of the intended speech-processing system. By continuously monitoring the audio stream and adjusting its decision based on the sequence and timing of wakeword detections, the system achieves reliable command processing while optimizing processing time through iterative analysis.
Data Source
AI summary
A system may include first and second speech-processing systems with corresponding first and second wakewords. An utterance may contain two or more wakewords. The system determines which speech-processing system to use to perform further audio processing and to determine a response to the utterance.