Multi-source estimation consistency for improved multiple direction-of-arrival estimation

File Description SizeFormat 
Hafezi_HSCMA2017.pdfFile embargoed until 01 January 10000694.35 kBAdobe PDF    Request a copy
Title: Multi-source estimation consistency for improved multiple direction-of-arrival estimation
Author(s): Hafezi, S
Moore, AH
Naylor, PA
Item Type: Conference Paper
Abstract: In Direction-of-Arrival(DOA) estimation for multiple sou rces, removal of noisy data points from a set of local DOA esti- mates increases the resulting estimation accuracy, especi ally when there are many sources and they have small angular separation. In this work, we propose a post-processing tech - nique for the enhancement of DOA extraction from a set of local estimates using the consistency of these estimates wi thin the time frame based on adaptive multi-source assumption. Simulations in a realistic reverberant environment with se n- sor noise and up to 5 sources demonstrate that the proposed technique outperforms the baseline and state-of-the-art a p- proaches. In these tests the proposed technique had the wors t average error of 9 ◦ , robustness of 5 ◦ to widely varying source separation and 3 ◦ to number of sources.
Publication Date: 1-Mar-2017
Date of Acceptance: 23-Jan-2017
URI: http://hdl.handle.net/10044/1/45039
Publisher: IEEE
Journal / Book Title: Joint Workshop on Hands-free Speech Communication and Microphone Arrays
Sponsor/Funder: Commission of the European Communities
Funder's Grant Number: 609465
Conference Name: Joint Workshop on Hands-free Speech Communication and Microphone Arrays
Copyright Statement: This paper is embargoed until publication.
Publication Status: Accepted
Start Date: 2017-03-01
Finish Date: 2017-03-03
Conference Place: San Francisco, USA
Embargo Date: publication subject to indefinite embargo
Appears in Collections:Faculty of Engineering
Electrical and Electronic Engineering



Items in Spiral are protected by copyright, with all rights reserved, unless otherwise indicated.

Creative Commons