1536-1233 (c) 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TMC.2015.2439278, IEEE Transactions on Mobile Computing 1 Robust Acoustic Self-Localization of Mobile Devices Diego B. Haddad, Wallace A. Martins, Maur´ ıcio V. M. Costa, Luiz W. P. Biscainho, Leonardo O. Nunes, and Bowon Lee Abstract—Self-localization of smart portable devices serves as foundation for several novel applications. This work proposes a set of algorithms that enable a mobile device to passively determine its position relative to a known reference with centimeter precision, based exclusively on the capture of acoustic signals emitted by controlled sources around it. The proposed techniques tackle typical practical issues such as reverberation, unknown speed of sound, line-of-sight obstruction, clock skew, and the need for asynchronous operation. After their theoretical developments and off-line simulations, the methods are assessed as real-time applications embedded into off-the-shelf mobile devices operating in real scenarios. When line of sight is available, position estimation errors are at most 4 cm using recorded signals. Index Terms—Acoustic sensor localization, least-squares, time of flight, time-difference of flight ✦ 1 I NTRODUCTION U BIQUITOUS smart portable devices equipped with loudspeakers and microphones such as mo- bile phones, tablets, and laptops constitute a potential acoustic network by themselves. Precise estimation of their positions may extend their inherent mobility and network capabilities, allowing the development of applications for indoor navigation in public places [3], [4], location-based services [5], monitoring patients indoors [6], extended gameplay experience [7], or ad- hoc microphone arrays [8], [9], [10], [11]. Acoustic sensor localization (ASL) algorithms aim to solve such estimation problems [12], [13], [14], [15], [16], [17], [18], [19] and may also be applied in other contexts, such as underwater acoustic sensor networks [20]. Originally concerned with the general problem of acoustic node localization [11], [21], most state-of- the-art ASL systems rely on estimates of the time of flight (TOF) between sound emission by the source and sound capture by the microphone (hereafter also called acoustic sensor or, simply, sensor) to feed their corresponding models. Moreover, most of related mathematical models fall into the broad class of least- squares (LS) problems whose aim is to match the unknown variables (e.g. the sensor 3-D locations) to the estimated TOFs [15], [16], borrowed from sound source localization (SSL) methods [22], [23], [24], [25]. In practice, however, several issues prevent one D. B. Haddad is with the Telecommunications Coordination of the Federal Center for Technological Education (CEFET/RJ), Rio de Janeiro, Brazil (e-mail: diegohaddad@gmail.com). W. A. Martins, M. V. M. Costa, L. W. P. Biscainho, and L. O. Nunes are/were with Poli & COPPE, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil (e-mail: wallace.martins@smt.ufrj.br; wagner@smt.ufrj.br; mauri- cio.costa@smt.ufrj.br; leonardo.nunes@smt.ufrj.br). B. Lee is with the Department of Electronic Engineering, Inha University, Incheon, Korea (e-mail: bowon.lee@inha.ac.kr). Some preliminary results of this work appeared in [1], [2]. from measuring TOFs accurately, which is essential to the localization performance: a) acoustic effects such as reverberation, 1 noise and obstacles, typically found indoors, which may corrupt the data required for TOF estimation; b) asynchrony between transmitters and receivers, due to both their independent clocks and inherent sending and recording delays, which turns the time base unreliable; c) a poor estimation of the speed of sound, which is inversely proportional to the TOF. An embedded software to provide a portable device with the capability of passive acoustic self- localization indoors must address all these challenges. This paper proposes an ASL tool tailored to localize mobile devices with built-in acoustic sensors inside a practical closed environment containing a set of con- trolled loudspeakers placed at known positions. Each loudspeaker sends, from time to time, a particular pseudo-noise signal, which is known by the mobile device. The position-estimation task is performed by the device itself, which acts as a passive acoustic sen- sor, without any information exchange with a central node or synchronization between sensor and loud- speakers (except possibly for loudspeaker positions loaded into the device when the localization tool is activated in a specific environment for the first time). The proposed techniques rely on the local estimation of TOFs and, if needed, on the subsequent evaluation of time-differences-of-flight (TDOFs). From the mobile devices’ viewpoint, since they work passively, there is no further energy consumption in addition to the energy spent in position calculations; therefore, due to the use of small-order LS techniques, the proposal can be regarded as energy-efficient. The use of additional hardware (the loudspeakers) 1. Acoustic equivalent of multipath propagation effect, well- known in electromagnetic-based wireless communications.