Direction of Arrival Estimation Using Replicated Phase Difference Matrix
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional multichannel sound source Direction of Arrival (DOA) estimation algorithms suffer from spatial aliasing issues, limiting their effectiveness to frequencies below the minimum aliasing frequency, which restricts the use of higher frequency bands and leads to inaccurate localization of sound sources.
Innovation Solution
The method involves replicating phase difference values in the phase difference matrix to cover all potential sinusoidal periods, allowing for the use of higher frequency bands and incorporating post-processing techniques to identify the most prominent peak values in histograms for more accurate DOA estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the distance between microphones is increased to improve localization accuracy, then the minimum aliasing frequency decreases, but spatial aliasing problems occur at higher frequencies
Solution Approach 1:
The frequency spectrum is segmented into multiple bands, with different processing strategies applied to each band. Lower frequency bands (below aliasing frequency) use conventional phase difference methods, while higher frequency bands use alternative localization approaches, allowing the system to utilize the full frequency spectrum without spatial aliasing contamination.
Solution Approach 2:
The system dynamically changes processing parameters based on frequency band. For frequencies below the minimum aliasing frequency, phase difference-based TDOA estimation is used. For frequencies above the aliasing frequency, the system switches to alternative methods such as energy-based or waveform similarity-based localization, adapting the measurement approach to avoid spatial aliasing effects.
2Reliability
If only frequency bands below the minimum aliasing frequency are used for localization, then spatial aliasing is avoided, but the quantity of usable frequency data is reduced
Solution Approach 1:
The frequency spectrum is divided into usable bands below the aliasing frequency and higher frequency bands above it. Both segments are processed separately with appropriate methods, and their results are combined to provide localization information from the entire frequency spectrum, maximizing the quantity of usable data.
Solution Approach 2:
The system employs multiple localization methods that serve different frequency ranges. The phase difference method handles lower frequencies, while alternative methods handle higher frequencies, creating a universal localization system that works across the entire audio spectrum rather than being limited to a single frequency range.
3Ease of manufacture
If conventional phase difference estimation is used for narrowband localization, then the method is simple to implement, but it fails when phase differences exceed the wrapped range
Solution Approach 1:
The system changes the estimation parameter based on frequency band. For narrowband signals below the aliasing frequency, phase difference is used. For broadband or higher frequency signals, the system transitions to time domain correlation or energy-based parameters that do not suffer from phase wrapping, maintaining both simplicity and accuracy across different conditions.
Data Source
AI summary
A device for estimating Direction of Arrival (DOA) of sound from Q≥1 sound sources is provided. The device is configured to obtain a phase difference matrix, which includes measured phase difference values, each of the measured phase difference values being a measured value of a phase difference between two microphone units for a frequency bin in a range of frequencies of the sound. The device is further configured to generate a replicated phase difference matrix by replicating the measured phase difference values to other potential sinusoidal periods, calculate a DOA value for each phase difference value in the replicated phase difference matrix, and determine, as Q DOA results, the Q most prominent peak values in a histogram generated based on the calculated DOA values.