2008文献Disambiguation of TDOA Estimation for Multiple Sources in Reverberant Environments.pdfVIP

IEEE TRANSACTIONS ON AUDIO, SPEECH, AND LANGUAGE PROCESSING, VOL. 16, NO. 8, NOVEMBER 2008 1479 Disambiguation of TDOA Estimation for Multiple Sources in Reverberant Environments Jan Scheuing and Bin Yang, Senior Member, IEEE Abstract— This paper presents a novel approach to estimate either under the ideal assumption that each source is exclusively the time difference of arrival (TDOA) for multiple sources in active during some time intervals [6]–[8] or by a blind source reverberant environments. It resolves ambiguities in TDOA esti- separation [9], [10]. Another completely different idea is to mation caused by multipath propagation and multiple sources. By exploiting two TDOA constraints, the raster condition and the scan the volume of interest for possible sources by maximizing, zero cyclic sum condition, we are able to identify and reject the e.g., the steered response power (SRP) [11]. echo path TDOAs and to assign the direct path TDOAs correctly In this paper, we present a novel approach for disambiguation to different sources. For the latter purpose, an efficient algorithm of TDOA estimation in multipath multisource environments for the synthesis of approximately consistent TDOA graphs has (DATEMM) [12], [13]. It is based on a fairly simple observation been developed. A real experiment demonstrates the superior performance of our algorithms. of two TDOA constraints implying information redundancy. By applying these constraints to TDOA estimates derived Index Terms—Disambiguation of TDOA estimation in