Sound Source Localization Using Ultrasonic Ranging and Adaptive DOA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In noisy environments, separating a desired speech signal from background noise is challenging due to the nonstationary nature of acoustic noise, which traditional single-microphone or fixed beamforming methods struggle to address, especially when the user's mouth position changes relative to the microphone.
Innovation Solution
A method that estimates the range of a sound-emitting object using ultrasonic signals and selects the appropriate direction-of-arrival estimation operation based on this range to enhance the direction-of-arrival calculation of audio-frequency components in a multichannel signal, improving the accuracy of source localization and noise separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional single-microphone or fixed beamforming methods are used, then device complexity is reduced, but speech separation performance deteriorates in noisy environments with nonstationary noise
Solution Approach 1:
The system dynamically adapts the direction-of-arrival estimation operation based on the estimated range of the sound source. When the sound source is in the near field (within 1.5 meters), a near-field estimation operation is selected; when in the far field, a far-field operation is selected. This dynamic adaptation allows the system to maintain high speech separation performance in noisy environments while managing computational complexity appropriately for different operating conditions.
2Adaptability or versatility
If the user's mouth position changes relative to the microphone, then adaptability is improved, but direction-of-arrival estimation accuracy deteriorates with fixed beamforming methods
Solution Approach 1:
The system uses ultrasonic ranging to dynamically determine the distance between the user's mouth and the microphone array. Based on this range information, the system selects the appropriate direction-of-arrival estimation operation (near-field or far-field), ensuring accurate source localization even when the user's mouth position changes during speech.
Solution Approach 2:
The patent introduces ultrasonic signals as an intermediary to measure the range between the sound source and the microphone array. This range information serves as a mediator that enables the system to select the appropriate DOA estimation method, thereby maintaining accuracy despite changes in mouth position. The ultrasonic ranging acts as a bridge between the physical position change and the adaptive signal processing response.
3Measurement precision
If ultrasonic ranging is used to estimate sound source distance, then source localization accuracy is improved, but device complexity increases
Solution Approach 1:
The system uses the same microphone array to perform both ultrasonic ranging and audio-frequency direction-of-arrival estimation. The microphone array serves multiple functions: detecting ultrasonic signals for range measurement and detecting audio signals for speech processing. This multi-functionality improves source localization accuracy while avoiding the need for separate ranging hardware, thus limiting the increase in device complexity.
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
This approach effectively enhances the separation of speech signals from background noise by accurately localizing the sound source and adapting to changes in the user's mouth position, leading to improved speech clarity and noise reduction in noisy environments.
Implementation Method 1
estimating a range of a sound-emitting object, based on information from a reflection of an ultrasonic signal
Implementation Method 2
estimating a range of a sound-emitting object, based on information from a reflection of an ultrasonic signal
Data Source
AI summary
A method of signal processing includes calculating a range based on information from a reflected ultrasonic signal. Based on the calculated range, one among a plurality of direction-of-arrival (DOA) estimation operations is selected. The method also includes performing the selected operation to calculate an estimated direction of arrival (DOA) of an audio-frequency component of a multichannel signal. Examples of DOA estimation operations include operations based on phase differences between channels of the multichannel signal and operations based on a difference in gain between signals that are based on channels of the multichannel signal.