Elsevier

Ad Hoc Networks

Volume 12, January 2014, Pages 100-114
Ad Hoc Networks

Low-dimensional signal-strength fingerprint-based positioning in wireless LANs

https://doi.org/10.1016/j.adhoc.2011.12.006Get rights and content

Abstract

Accurate location awareness is of paramount importance in most ubiquitous and pervasive computing applications. Numerous solutions for indoor localization based on IEEE802.11, bluetooth, ultrasonic and vision technologies have been proposed. This paper introduces a suite of novel indoor positioning techniques utilizing signal-strength (SS) fingerprints collected from access points (APs). Our first approach employs a statistical representation of the received SS measurements by means of a multivariate Gaussian model by considering a discretized grid-like form of the indoor environment and by computing probability distribution signatures at each cell of the grid. At run time, the system compares the signature at the unknown position with the signature of each cell by using the Kullback–Leibler Divergence (KLD) between their corresponding probability densities. Our second approach applies compressive sensing (CS) to perform sparsity-based accurate indoor localization, while reducing significantly the amount of information transmitted from a wireless device, possessing limited power, storage, and processing capabilities, to a central server. The performance evaluation which was conducted at the premises of a research laboratory and an aquarium under real-life conditions, reveals that the proposed statistical fingerprinting and CS-based localization techniques achieve a substantial localization accuracy.

Introduction

Location-sensing has been impelled by the emergence of location-based services in the transportation industry, emergency situations for disaster relief, the entertainment industry, and assistive technologies in the medical community. Location-sensing systems can be classified according to their dependency on and/or use of (a) specialized infrastructure and hardware, (b) signal modalities, (c) training, (d) methodology and/or use of models for estimating distances, orientation, and position, (e) the coordination system (absolute or relative), scale, and location description, (f) localized or remote computation, their mechanisms for device identification, classification, and recognition, and their accuracy and precision requirements. The distance can be estimated using time-of-arrival (e.g., GPS, PinPoint [1]) or signal-strength measurements, if the velocity of the signal and a signal attenuation model for the environment can be accurately estimated, respectively. Positioning systems may employ different modalities, such as, IEEE802.11 (e.g., Radar [2], [3], Ubisense, Ekahau [4]), infrared (e.g., Active Badge [5]), ultrasonic (e.g., Cricket [6], [7], Active Bat), Bluetooth [3], [8], [9], [10], [11], 4G [12], vision (e.g., EasyLiving project), and physical contact with pressure (e.g., Smart Floor), touch sensors or capacitive detectors. They may also combine multiple modalities to improve the localization, such as optical, acoustic and motion attributes (e.g., SurroundSense [13]).

The popularity of IEEE802.11 infrastructures, their low deployment cost, and the advantages of using them for both communication and positioning, make them an attractive choice. Most of the signal-strength based localization systems can be classified into the following two categories, namely signature- or map-based and distance-prediction-based techniques. The first type creates a signal-strength signature or map of the physical space during a training phase and compares it with the signature generated at runtime (at the unknown position) [2], [14], [15]. To build such signatures, signal-strength data is gathered from beacons received from APs. During a training phase, such measurements are collected at various predefined positions (of the map) and signatures are generated that associate the corresponding positions of the physical space with statistical measurements based on signal-strength values acquired at those positions. Such maps can be formed with data from different sources or signal modalities to improve location-sensing [3], [6]. The distance-prediction-based techniques use the signal-strength values and radio-propagation models to predict the distance of a wireless client from an AP (or any landmark) or even between two wireless clients (peers) with estimated position (such as CLS [16]). In situations where a deployment of a wireless infrastructure may not be feasible, positioning mechanisms may exploit cooperation by enabling devices to share positioning estimates [1], [16], [17], [18], [19], [20], [21], [22]. A survey of positioning systems can be found in [23].

In this paper, first we build on our earlier work on CLS [16], [20], which generates statistical-based fingerprints using the received signal-strength (RSSI) measurements from an IEEE802.11 infrastructure. The vast majority of current fingerprint positioning methods does not take into account the interdependencies among the RSSI measurements at a certain position from the various APs. These interdependencies provide important information about the geometry of the environment and can be quantified using the second-order spatial correlations among the measurements. Hence, the employment of multi-dimensional distributions is expected to provide a more accurate representation of the RSSI signatures, leading to improved positioning performance. Simple models whose second-order statistics can be accurately and easily estimated could be used in practice. In particular, a multivariate Gaussian-based approach is employed to take into consideration the statistics of the RSSI measurements not only from each distinct AP but also the interplay (covariance) of measurements collected from pairs of APs. The signature comparison and position estimation is based on the Kullback–Leibler divergence (KLD): the cell corresponding to the minimum KLD is reported as the estimated position. This approach is generalized by applying it iteratively in different spatial scales.

The difficult to predict nature of the RSSI measurements, due to the impact of transient phenomena on the RSSI values, impels for extensive training, which increases the overhead of the fingerprint-based positioning systems: a larger training set and more sophisticated algorithms are often employed to capture the dynamic complex nature of the RSSI measurements. On the other hand, the inherent sparse nature of the localization of a mobile device in the physical space (since it can be placed at a single position of a discretized grid-like form of the environment) motivates the use of the recently introduced theory of compressive sensing (CS) [24], [25] for target localization [26]. CS states that signals which are sparse in a suitable transform basis can be recovered from a highly reduced number of incoherent random projections. Hence, the CS-based approach comes as an evolution to the traditional methods dominated by the well-established Nyquist–Shannon sampling theory, and consequently it could be exploited in the design of efficient localization systems characterized by limited resources.

In a recent work [27], a CS-based indoor localization method was introduced based on RSSI measurements. In particular, the location estimation algorithm is carried out on the mobile device by using the average RSSI values in order to construct the transform basis. The sparsity-based CS localization algorithm proposed in this paper differs from the work in [27] in several aspects. In contrast to [27], where the estimation is performed by the wireless device with the potentially limited resources, in our proposed algorithm the computational burden can be assigned to a central node (fusion center), where increased storage and processing resources are available. Unlike in [27] that uses the average RSSI values, the proposed CS approach is applied directly on the raw RSSI measurements, thus exploiting their time-varying behavior. Then, the estimation of the unknown position is performed by solving a constraint optimization problem for reconstructing a sparse vector with its coordinates being “1” or “0” depending on whether the mobile device is placed or not at the corresponding cell.

This paper makes the following contributions:

  • 1.

    It proposes and evaluates a novel fingerprinting approach that exploits the spatial correlations of signal-strength measurements collected from various wireless APs based on a multivariate Gaussian model.

  • 2.

    It introduces a novel localization approach that applies compressive sensing (CS), which can achieve an increased accuracy in the position estimation, while reducing the communication overhead required for the exchange of measurements, and thus, becoming more appropriate for energy-constrained devices.

  • 3.

    It performs a comparative performance analysis of various signal-strength fingerprinting methods in the premises of a research laboratory and an aquarium under different conditions.

The paper is organized as follows: Section 2 presents recently introduced statistical signal-strength signature techniques, along with the proposed statistical approach based on the use of multivariate Gaussian distributions for modelling the statistics of the RSSI measurements. In Section 3, the main principles of CS are introduced and the proposed CS-based localization method is analyzed in detail. Section 4 presents a comparative performance evaluation of these techniques in the premises of the Telecommunications and Networks Lab (TNL) at ICS-FORTH as well as in the Cretaquarium. Section 5 overviews related positioning systems for mobile computing, while Section 6 summarizes our main results and provides directions for future work.

Section snippets

Statistical fingerprint methods

A wireless device that listens to a channel receives the beacons sent by APs (at that channel) periodically and records their RSSI values. Typically wireless devices that run fingerprint-based positioning systems acquire such measurements at various positions in a given physical space (with a deployment of wireless APs) and generate fingerprints for these positions applying various statistical metrics (e.g., confidence intervals, percentiles, empirical distributions or the parameters of a

Compressive sensing WLAN localization

Let us first describe the main theoretical concepts of CS as applied in the context of positioning. Let xRN denote the signal of interest, that is, a vector of RSSI measurements. The efficiency of a CS method for signal approximation or reconstruction depends highly on the sparsity structure of the signal in a suitable transform domain associated with an appropriate sparsifying basis ΨRN×D. It has been demonstrated [24], [25] that if x is K-sparse in Ψ (meaning that the signal is exactly or

Performance analysis

The following two subsections evaluate the performance of the proposed algorithms in two distinct real-world environments. As described, the training signatures are generated based on the collected signal-strength measurements at each cell of the two grid-based representations of the environments. In both cases, runtime measurements were collected at 35 randomly selected cells. The trainer (user) remained still for approximately 60 s (30 s) to collect beacons at each position during training

Related work

Significant work has been published in the area of location-sensing using RF signals. Radar [2] employs signal-strength maps that integrate signal-strength measurements acquired during the training phase from APs at different positions with the physical coordinates of each position. Each measured signal-strength vector is compared against the reference map and the coordinates of the best match will be reported as the estimated position. Bahl et al. [32] improved Radar to alleviate side effects

Conclusions and future work

This paper introduced two novel localization methods based on RSSI measurements. In the first case, statistical signal-strength fingerprints are created using multivariate Gaussian distributions. The position of the device is estimated by computing the region with the training fingerprint that has the minimum KLD from the runtime fingerprint. In the second case, the localization problem was reduced in a sparse reconstruction problem in the framework of CS. The dimensionality of the original

Acknowledgments

This work was partially funded by the Marie Curie IAPP “CS-ORION” (PIAP-GA-2009-251605) grant within the 7th Framework Program of the European Community and by the Greek General Secretariat for Research and Technology (Regional of Crete) Crete-Wise Grant.

Dimitrios Milioris received the B.S. degree in Computer Science from the University of Crete (UOC), Greece, in 2009 and a double M.Sc. degree (First in Class, Honors) in Computer Science from Paris XI University and University of Crete (2011). Since 2008 he has been a research assistant at the Telecommunications and Networks Laboratory (TNL) of the Foundation for Research and Technology-Hellas (FO.R.T.H.). In 2010 he joined Hipercom Team of the Institut National de Recherche en Informatique et

References (50)

  • M. Youssef, A. Youssef, C. Rieger, U. Shankar, A. Agrawala, PinPoint: An asynchronous time-based location determination...
  • P. Bahl, V. Radmanabhan, Radar: An in-building RF-based user location and tracking system, in: IEEE InfoCom, March...
  • Y. Gwon, R. Jain, T. Kawahara, Robust indoor location estimation of stationary and mobile users, in: IEEE InfoCom, Hong...
  • Ekahau v.3.1....
  • R. Want et al.

    The active badge location system

    ACM Trans. Info. Syst.

    (1992)
  • N.B. Priyantha, A. Chakraborty, H. Balakrishnan, The cricket location-support system, in: ACM MobiCom, August...
  • N.B. Priyantha, A. Miu, H. Balakrishnan, Teller, The cricket compass for context-aware mobile applications, in: ACM...
  • U. Bandara, M. Hasegawa, M. Inoue, H. Morikawa, T. Aoyama, Design and implementation of a bluetooth signal strength...
  • S. Feldmann, K. Kyamakya, A. Zapater, Z. Lue, An indoor bluetooth-based positioning system: concept, implementation and...
  • M. Rodriguez, J.P. Pece, C.J. Escudero, In-building location using bluetooth, in: International Workshop on Wireless...
  • S. Asthana, D. Kalofonos, The problem of bluetooth pollution and accelerating connectivity in bluetooth ad-hoc...
  • A. Roy, A. Misra, S.K. Das, An information theoretic framework for optimal location tracking in multi-system 4G...
  • M. Azizyan, I. Constandache, R. Choudhury, SurroundSense: Mobile phone localization via ambience fingerprinting, in:...
  • A. Ladd, K. Bekris, A. Rudys, G. Marceau, L. Kavraki, D. Wallach, Robotics-based location sensing using wireless...
  • M. Youssef, A. Agrawala, The horus WLAN location determination system, in: ACM MobiSys, June 6–8,...
  • K. Vandikas, L. Kriara, T. Papakonstantinou, A. Katranidou, H. Baltzakis, M. Papadopouli, Empirical-based analysis of a...
  • C. Savarese, J. Rabaey, K. Langendoen, Robust positioning algorithms for distributed ad-hoc wireless sensor networks,...
  • S. Capkun, M. Hamdi, J.-P. Hubaux, GPS-free positioning in mobile ad-hoc networks, in: Proc. Hawaii International...
  • D. Niculescu, B. Nath, Ad hoc positioning system (APS), in: IEEE GlobeCom, San Antonio, TX, November...
  • C. Fretzagias, M. Papadopouli, Cooperative location sensing for wireless networks, in: IEEE PerCom, Orlando, Florida,...
  • K. Chintalapudi, R. Govindan, G. Sukhatme, A. Dhariwal, Ad-hoc localization using ranging and sectoring, in: IEEE...
  • L. Fang, W. Du, P. Ning, A beacon-less location discovery scheme for wireless sensor networks, in: IEEE InfoCom, Miami,...
  • J. Hightower, G. Borriello, A survey and taxonomy of location sensing systems for ubiquitous computing, Tech. Rep.,...
  • E. Candés et al.

    Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information

    IEEE Trans. Inform. Theory

    (2006)
  • D. Donoho

    Compressive sensing

    IEEE Trans. Inform. Theory

    (2006)
  • Cited by (82)

    • Research on Crowdsourcing network indoor localization based on Co-Forest and Bayesian Compressed Sensing

      2020, Ad Hoc Networks
      Citation Excerpt :

      With the development and optimization of mobile ad hoc network [1], indoor localization information in the fields of navigation services, mobile socialization, public safety, and smart city construction is extremely important for individuals [2], which uses radio communication network technologies and external localization methods to make the ILT gradually become a hot topic.

    • An accurate UWB based localization system using modified leading edge detection algorithm

      2020, Ad Hoc Networks
      Citation Excerpt :

      The most widely used ranging methods are, received signal strength indicator (RSSI), angle of arrival (AOA) and TOA [4]. RSSI based methods [5–10] estimate the distance between the base station and the mobile unit using the received signal strength. The trilateration method uses these range measurements to estimate the mobile position.

    • Wireless Localization Techniques

      2023, Wireless Networks (United Kingdom)
    • Urban Vehicle Localization in Public LoRaWan Network

      2022, IEEE Internet of Things Journal
    View all citing articles on Scopus

    Dimitrios Milioris received the B.S. degree in Computer Science from the University of Crete (UOC), Greece, in 2009 and a double M.Sc. degree (First in Class, Honors) in Computer Science from Paris XI University and University of Crete (2011). Since 2008 he has been a research assistant at the Telecommunications and Networks Laboratory (TNL) of the Foundation for Research and Technology-Hellas (FO.R.T.H.). In 2010 he joined Hipercom Team of the Institut National de Recherche en Informatique et en Automatique (I.N.R.I.A. – Rocquencourt campus) as a research assistant. His research interests lie in the field of signal processing, mobile computing, performance analysis, sensor networks, compressive sensing, indoor localization and information theory.

    George Tzagkarakis received the B.S. degree in Mathematics from the University of Crete (UOC), Greece, in 2002 (First in Class, Honors). At the same year he joined the Computer Science department (CSD) at the UOC for graduate studies with scholarships from the CSD and the Institute of Computer Science (ICS) of the Foundation for Research and Technology-Hellas (FO.R.T.H.). He received the M.Sc. (First in Class, Honors) and the Ph.D. degrees from the CSD in 2004 and 2009, respectively. Since 2000, he has been also collaborating with the Wave Propagation Group of the Institute of Applied and Computational Mathematics (FO.R.T.H.), while from 2002 he is a research assistant in the Telecommunications and Networks Lab (ICS-TNL). Currently he is a post-doctoral researcher under a Marie Curie fellowship at CEA, Saclay (France). His research interests lie in the fields of statistical signal & image processing with emphasis in non-Gaussian heavy-tailed modeling, compressive sensing with applications in video processing, distributed signal processing for sensor networks, information theory, and inverse problems in underwater acoustics.

    Artemis Papakonstantinou received a B.Sc. degree in Computer Science (2007) and a M.Sc. degree in Computer Science (2011), both from the Computer Science Department, University of Crete, Greece. Since 2008 she has been a research assistant at the Telecommunications and Networks Lab (TNL) of the Institute of Computer Science, Foundation for Research and Technology-Hellas (ICS-FORTH). Her research interests lie in the fields of wireless networks, mobile and pervasive computing with emphasis on location sensing and location based services (LBS), and smartphone applications and services.

    Maria Papadopouli (Ph.D. Columbia University, October 2002) is a tenured assistant professor in the Department of Computer Science, University of Crete, a guest professor at the EE KTH Royal Institute of Technology, in Stockholm, and a research associate in FORTH-ICS. From July 2002 until June 2006, she was a tenure-track assistant professor at UNC (on leave from July 2004 until June 2006). Her current research interests are in wireless networking, modeling and performance analysis, network measurements, cognitive radio networks, mobile peer-to-peer computing, positioning, and pervasive computing. She has co-authored a monograph on Peer-to-Peer Computing for Mobile Networks: Information Discovery and Dissemination (Springer Eds. 2009). She has been the co-chair of nine international workshops in the area of wireless networks and mobile peer-to-peer computing and has given more than 25 invited talks in research labs and universities world-wide. In 2004 and 2005, she was awarded with an IBM Faculty Award.

    Panagiotis Tsakalides received the Diploma in electrical engineering from the Aristotle University of Thessaloniki, Greece, in 1990, and the Ph.D. degree in electrical engineering from the University of Southern California (USC), Los Angeles, in 1995. He is a Professor of Computer Science at the University of Crete, and a Researcher with the Institute of Computer Science, Foundation for Research and Technology-Hellas (FORTH-ICS), Greece. From 2004 to 2006, he served as the Department Chairman. From 1999 to 2002, he was with the Department of Electrical Engineering, University of Patras, Patras, Greece. From 1996 to 1998, he was a Research Assistant Professor with the Signal and Image Processing Institute, USC, and he consulted for the US Navy and Air Force. His research interests lie in the field of statistical signal processing with emphasis in non-Gaussian estimation and detection theory, and applications in sensor networks, audio, imaging, and multimedia systems. He has coauthored over 100 technical publications in these areas, including 25 journal papers. He is the PI of the 1.3 M euros FP7 MC-IAPP ”CS-ORION” project (2010–2014) conducting research on compressed sensing for remote imaging in aerial and terrestrial surveillance. He is a member of the ERCIM Network of Innovation/Technology and Knowledge Transfer Experts (I-Board).

    View full text