Dual-Path Microphone Array for Noise-Resistant Keyword Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Far field microphone systems struggle to accurately detect keywords in noisy environments due to slow beam steering, often misdirecting towards noise sources like TVs or multiple talkers, leading to repeated keyword activation.
Innovation Solution
A dual signal processing path system is implemented, combining far-field microphone functionality with a less directional microphone functionality, using delay-and-sum beamforming, allpass filters, and shared echo cancellation to enhance noise reduction and adapt beam directionality, allowing for improved noise reduction and keyword detection across 360°.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a beamforming structure is used to increase signal-to-noise ratio, then speech can be picked up at greater distances, but the beam may be steered towards noise sources like television sets instead of the talker
Solution Approach 1:
The system dynamically switches between far-field directional beamforming mode and near-field omnidirectional mode based on acoustic environment detection. The beamforming structure adapts its characteristics in real-time to track talkers while avoiding distraction from noise sources like television sets.
Solution Approach 2:
The system changes the directional characteristics parameter of the microphone array by switching between different signal processing paths. The far-field path uses highly directional beamforming with specific steering parameters, while the near-field path uses omnidirectional processing, allowing adaptation to different acoustic scenarios.
2Speed
If the beamforming structure is too slow to track the talker, then keyword recognition fails and the talker must repeat the keyword, but increasing tracking speed may reduce system stability
Solution Approach 1:
The system performs preliminary detection of acoustic environments and talker positions using the omnidirectional near-field path before committing to directional tracking. This preliminary action allows the beamforming structure to be pre-positioned optimally, enabling faster tracking response while maintaining reliable keyword recognition.
Solution Approach 2:
The system uses feedback from the omnidirectional microphone path to continuously monitor talker position and acoustic conditions. This feedback loop enables the beamforming structure to adjust its steering quickly and accurately, improving tracking speed while maintaining reliability through continuous verification.
3Measurement precision
If a highly directional beam-shaped sensitivity characteristic is used, then signal-to-noise ratio increases, but coverage area is reduced and multiple talkers from different directions cannot be effectively handled
Solution Approach 1:
The system segments the signal processing into two distinct paths: a far-field path with highly directional beamforming for high SNR in distant talker scenarios, and a near-field path with omnidirectional characteristics for multi-directional coverage. Each path handles specific acoustic scenarios, allowing the system to optimize for both SNR and coverage area depending on conditions.
Solution Approach 2:
The microphone array system performs multiple functions by switching between far-field directional beamforming and near-field omnidirectional processing. This multi-functionality allows the same hardware to handle both distant single talkers (high SNR mode) and near-field multiple talkers (omnidirectional mode), increasing overall adaptability.
Data Source
AI summary
Sound capturing which includes applying a far-field microphone functionality to a multiplicity of first microphone signals to provide a first output signal, and applying a less directional microphone functionality to one or more second microphone signals to provide a second output signal.