Keyword-Based Audio Source Localization With Acoustic Beamforming
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Devices struggle to distinguish between human speech and background noise, leading to difficulty in understanding commands due to environmental interference.
Innovation Solution
Implementing multiple listening zones with beamformers to detect keywords, determine the direction and distance of the speaker, and form active acoustic beams for focused speech recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the device listens for speech from all directions (keyword listening mode), then it can detect keywords spoken by any user, but it cannot distinguish between the target speaker and background noise or other speakers
Solution Approach 1:
The device divides the 360-degree listening space into multiple discrete listening zones using beamformers. Each beamformer points in a specific direction and creates a focused listening zone. This segmentation allows the system to later select which zone contains the keyword speaker, resolving the contradiction between omnidirectional detection and precise localization.
Solution Approach 2:
The system dynamically transitions between keyword listening mode (omnidirectional) and directed speech listening mode (focused). During keyword detection, all beamformers are active to maximize detection capability. Once a keyword is detected in a specific zone, the system switches to directed mode, activating only the beamformer pointing at that zone, thus maintaining versatility while achieving precision.
2Adaptability or versatility
If the device uses multiple beamformers pointing in various directions, then it can cover all directions for keyword detection, but it increases device complexity
Solution Approach 1:
Each beamformer serves multiple functions: during keyword listening mode, all beamformers collectively provide 360-degree coverage for keyword detection. During directed speech listening mode, the selected beamformer provides focused listening for command recognition. This multi-functionality reduces the need for separate systems for different modes, managing complexity while maintaining versatility.
3Measurement precision
If the device switches to directed speech listening mode with active acoustic beams, then it improves speech recognition accuracy, but it limits the ability to hear speech from other directions
Solution Approach 1:
The system dynamically adjusts its listening configuration based on operational mode. In directed speech listening mode, it focuses on a specific direction for high-precision command recognition. When switching back to keyword listening mode, it restores omnidirectional coverage. This dynamic reconfiguration resolves the contradiction by making the limitation temporary and context-appropriate.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the ability to accurately recognize spoken commands by isolating the source of speech and reducing background noise interference.
Implementation Method 1
form one or more active acoustic beams directed toward the person speaking
Data Source
AI summary
Systems, apparatuses, and methods are described for determining a direction associated with a detected spoken keyword, forming an acoustic beam in the determined direction, and listening for subsequent speech using the acoustic beam in the determined direction.