Distributed Wake Word Detection for Smart Home Voice Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing smart home systems for voice command control have limited operational ranges and rely heavily on a single microphone location, providing a restricted voice command experience in broader living environments.
Innovation Solution
An AI-based system that integrates a voice command processing server node with machine learning modules, utilizing a neural network to process audio signals, detect wake words, and generate predictive models for controlling connected devices, supported by a blockchain network for secure data management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a single microphone location is used in existing BMS systems, then the system structure is simple, but the operational range is limited and voice command experience is restricted
Solution Approach 1:
The patent divides the audio capture function into multiple independent microphone units distributed throughout the smart home environment. Each microphone acts as an autonomous node that can independently process and transmit audio data, enabling the system to cover a much larger operational area while maintaining manageable complexity through modular architecture
Solution Approach 2:
The patent transitions from a single-point microphone location to a distributed spatial network of microphones. By adding the spatial dimension and distributing microphones across multiple locations, the system expands its operational range from a localized area to an entire smart home environment while using networked architecture to manage the increased complexity
2Reliability
If voice commands are processed with limited audio capture points, then the system is easier to implement, but misinterpretations and malfunctions increase
Solution Approach 1:
The patent merges audio data from multiple distributed microphones into a unified processing stream. By combining signals from various locations and using ensemble processing, the system improves voice command accuracy and reduces misinterpretations while managing complexity through centralized processing of aggregated data
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors audio quality, wake word detection accuracy, and command recognition performance. This feedback enables dynamic adjustment of processing parameters and microphones selection, improving reliability while using adaptive algorithms to manage processing complexity
3Measurement precision
If traditional wake word detection is used, then the processing is simpler, but responsiveness and accuracy are reduced
Solution Approach 1:
The patent performs preliminary processing of audio signals including normalization and spectrogram conversion before wake word detection. By pre-processing the audio data and extracting relevant features in advance, the system improves detection accuracy while reducing the computational complexity of the final classification stage
Solution Approach 2:
The patent replaces traditional mechanical or rule-based wake word detection with machine learning-based neural networks. This substitution enables more accurate and context-aware detection by learning from training data, while the use of optimized neural network architectures and feature extraction techniques manages the increased computational complexity
Data Source
AI summary
A system for an automated voice command processing within a smart home including a processor of a voice command processing server node configured to host a machine learning (ML) module and connected to at least one audio capture entity node and to at least one target node over a wireless network connection and a memory on which are stored machine-readable instructions that when executed by the processor, cause the processor to: acquire raw audio data comprising an audio signal from the at least one audio capture entity node; normalize the audio signal for volume consistency; convert the normalized audio signal into a spectrogram; extract a set of classifying features from the spectrogram; provide the set of classifying features to the ML module configured to generate a predictive model based on a neural network for producing at least one wake word parameter; detect a wake word based on the at least one wake word parameter; and switch the voice command processing server node to an active listening mode for processing subsequent user audio commands through the at least one audio capture entity node.