Circular Statistics-based low complexity DOA estimation for hearing aid application

TL;DR

This paper introduces a low-complexity, circular statistics-based method for estimating the direction of arrival in hearing aid systems, improving accuracy and speed over existing benchmarks.

Contribution

It presents CIMPL, a novel approach using circular statistics and Kalman filtering for efficient DoA estimation in binaural hearing aids.

Findings

Improves accuracy of DoA estimation for moving sources.

Runs approximately 75 times faster than benchmark methods.

Provides reliable angle estimates with integrated variance as a confidence measure.

Abstract

The proposed Circular statistics-based Inter-Microphone Phase difference estimation Localizer (CIMPL) method is tailored toward binaural hearing aid systems with microphone arrays in each unit. The method utilizes the circular statistics (circular mean and circular variance) of inter-microphone phase difference (IPD) across different microphone pairs. These IPDs are firstly mapped to time delays through a variance-weighted linear fit, then mapped to azimuth direction-of-arrival (DoA) and lastly information of different microphone pairs is combined. The variance is carried through the different transformations and acts as a reliability index of the estimated angle. Both the resulting angle and variance are fed into a wrapped Kalman filter, which provides a smoothed estimate of the DoA. The proposed method improves the accuracy of the tracked angle of a single moving source compared with the benchmark method provided by the LOCATA challenge, and it runs approximately 75 times faster.