ECAI 2018 - International Conference – 10th Edition

Electronics, Computers and Artificial Intelligence

28 June -30 June, 2018, Iasi, ROMÂNIA

RSSI Based Localization of Bluetooth Devices

Using Trilateration: An Improved Method for the
Visually Impaired
Muhammad Irfan Aziz Thomas Owens
College of Aeronautical Engineering Department of Electronic and Computer Engineering
National University of Sciences and Technology Brunel University
Risalpur, Pakistan West London, United Kingdom
[Link]@[Link] [Link]@[Link]

Uzair Khaleeq uz Zaman

Laboratoire de Conception Fabrication Commande
Ecole Nationale Supérieure d’Arts et Métiers
Metz, France
[Link]@[Link]

Abstract - The instantaneous and spatial localization for visually user-friendly localization and location-based services (LBS)
impaired people in dynamically changing environments with are the need of today [3,4].
unexpected hazards and obstacles, is the most demanding and The Global positioning system (GPS) has been used quite
challenging issue faced by the navigation systems today. extensively over the years for outdoor environments but it
Although time difference of arrival (TDOA), time of arrival
deteriorates considerably indoors or near high rise structures
(TOA) and received signal strength (RSS) have been used in
literature multiple times for positioning, but narrow band signals due to technical constraints such as shadowing and multi-path
such as Bluetooth cannot efficiently utilize TDOA or TOA. disturbances [5,6]. Therefore, other techniques such as time of
Received signal strength indicator (RSSI), to measure RSS, has arrival (TOA), time difference of arrival (TDOA) and received
been found to be more suitable in this case. It has been found that signal strength (RSS) have been used extensively in literature
the estimations using RSSI can be improved significantly by for localization of LBS. But since, narrow band signals such
improving the existing methodologies related to RSSI. Therefore, as Bluetooth cannot efficiently utilize TDOA or TOA,
the current paper focuses on proposing an improved method received signal strength indicator (RSSI) to measure RSS, has
using trilateration for localization of Bluetooth devices for been found to be more reliable as it caters for the
visually impaired people. To validate the method, class 2
environmental interference [7,8]. As the location of a node is
Bluetooth devices were used along with the development of a
software. Experiments were then conducted to obtain surface very important in a sensor network, triangulation, trilateration
plots that showed the signal interferences and other and multilateration have been used extensively as positioning
environmental effects. Finally, the results obtained were methods. Angles are used for positioning in triangulation
discussed. while distances are employed in trilateration. Multilateration
uses TDOA.
Keywords - Bluetooth, indoor/outdoor localization, received Bluetooth transmitters, if installed at regular intervals,
signal strength indicator, visually impaired. transmit pre-recorded voice messages and can provide
information about users’ current position and nearby points of
I. INTRODUCTION interests [9]. This kind of localization is a not a novel idea but
An estimated 253 million people in the world live with due to the limitations of the original Bluetooth specification,
vision impairment of which 36 million are blind and 217 this approach has not been widely used [10]. The Bluetooth-
million have moderate to severe vision impairment [1]. based localization Devices can also connect several other
Normally, a mixture of declarative and route knowledge is devices to Personal Area Networks (PANs) thereby reducing
used for navigation by the blind which is learnt by the positioning error [11]. In addition, the available services of
participation in orientation and mobility (O&M) sessions [2]. Bluetooth and their discovery are relatively simple which
But due to unexpected hazards and obstacles in dynamically allow smoother communications as compared to Ethernet
changing environments, time saving, adaptable, cost- and based Wi-Fi networks.

978-1-5386-4901-5/18/$31.00 ©2018 IEEE

 The work in this paper, therefore, proposes a RSSI based were calculated. Frequency Modulation (FM) and Wi-Fi
improved methodology using trilateration for localization of vectors for RSSI were also combined to check the effect on
Bluetooth devices for the visually impaired people. The main localization accuracy by Chen et al. [19].
contribution to the field is in terms of making a system which
can easily be incorporated with a smart phone of the affected C. Bluetooth, RSSI and LBS
person. The novelty lies in the utilization of more than one
techniques to improve the overall results significantly, after For different indoor environments, Bluepass was
their fusion. The methodology can be adopted by shopping developed by Diaz et al. [20]. Different characteristics of
malls, etc. to help these visually impaired people. To validate Bluetooth such as RSSI, link quality (LQ), transmit power
the method, class 2 Bluetooth devices (Blue Giga WT-12) were level, etc., were evaluated by Hossain and Soh [21] who
used. National Instruments LabView was used to develop the further concluded that a combination of LQ and RSSI is a
required software. Experiments were then conducted to viable option for localization purposes. Moreover, Feldmann
generate surface plots of Bluetooth modules that can show the et al. [22] utilized the golden receiver power rank (GRPR) to
signal interferences and other environmental effects. ascertain that to obtain functional approximation of the
This paper is organized as follows: section II presents the distance between a receiver and a transmitter, only positive
theoretical background for this work; section III presents the values of RSSI should be used. Altini et al. [23] also described
proposed improved methodology; section IV describes the a region-based localization method to secure accurate location
implementation and conduct of practical experiments, and services without triangulation to handle the situations in which
finally, section V discusses the results and the conclusions more than three Bluetooth base stations are involved.
drawn. Furthermore, Li and Wang [24] found the required location by
identifying the point of intersection of all the three circles
II. BACKGROUND corresponding to three Bluetooth transmitters. In reference to
the visually impaired, Bohonos et al. [25] used an
A. Techniques to find Direction of Arrival (DOA) implementation on a mock setup to demonstrate how
Bluetooth can be used as an aid for the blind when crossing a
In multiple signal sources, DOA is very important. Various road intersection. Last, Liu et al. [26] improved the overall
methods are discussed in literature to find it. To minimize the reliability and accuracy of the system by coupling fluorescent
effects of ground reflection, multiple signal classification light communication (FLC) with Bluetooth and radio
(MUSIC) algorithm has been used extensively by calibrating frequency identification (RFID) tags installed at different
commercial electromagnetic compatibility (EMC) antennas locations.
[12] and understanding the physical meaning of the presence
of noise and other adverse effects [13]. Beamspace et al. [14] III. PROPOSED METHODOLOGY
used Estimation method of Signal Parameters via Rotational In this paper, a different approach was adopted from that
Invariance Technique (ESPRIT) to utilize a self-initiating available in literature for trilateration. Conventionally,
MUSIC based direction finder (DF) in an Acoustic Particle trilateration measures the distances between two points and
Velocity-Field. DOA can also be found using an Adcock DF then solves for the desired results using geometry of circles,
(ADF) [15]. spheres and/or triangles. In the methodology proposed, instead
of finding the circles of power, the normal to line connecting
B. Received Signal Strength Indicator (RSSI) the circles was determined. If we assume that the object of
interest is transmitting an RF signal which can be heard by at
RSSI is a ‘distance-dependent’ complex function that is least three different receivers, then depending upon the power
easily interfered with by noisy wireless channels. A Radio received by each of them, trilateration can be done, as shown in
Frequency (RF) fingerprint of the region is constructed by Fig. 1.
systems that utilize RSSI as demonstrated by Ibrahim and
Youssef [16] who developed ‘Cellsense’ using a combination
of deterministic and probabilistic techniques. It also implies
that trilateration and fingerprinting are two basic techniques P2
used in many indoor positioning systems (IPSs). A model
21

made specifically for RSSI fingerprints was proposed by

Rx

D2
Rx

23
D1

Danis and Cemgil. To give location aware services, RSSI

L1

Rx
12

32
L2

schemes are being used widely almost everywhere in WLAN

Rx

(X,Y) L3 P3
Rx31

[17]. Moreover, finding the maximum value for RSSI is P1 Rx13

imperative in localization and was found by Sahu et al. [18] D3

who suggested that as the terrain changed from plain to

uneven, the fluctuations in RSSI values increased, and to
determine the sensor position, the point of intersection of the
perpendiculars that pass through the maximum RSSI point, Fig. 1 Overlapping areas of three receivers depending upon theoretical
RSSI
 Fig. 1 shows three unique receivers: P1, P2 and P3, The solution of the above lines (their intersection with each
marked by their area of listening as yellow, green and blue, other) will result in the error triangle. The area inside the
respectively. The object of interest (A) for which trilateration triangle is the probable location of the transmitter.
is intended, is at point (X, Y). Since the coordinates of P1, P2
and P3 are fixed and known, the distances between them (i.e., IV. IMPLEMENTATION AND EXPERIMENTS
D1, D2 and D3) are also known. P1, P2 and P3 make a
triangle and three perpendiculars are drawn from each side of
A. Implementation
the triangle to the location of ‘A’. These perpendiculars are
L1, L2 and L3, respectively, and each divide lines D1, D2 and
To implement the improved method of trilateration
D3 in such a way that D1 = Rx12 + Rx21, D2 = Rx23 + Rx32; and
explained in section III, a Bluetooth module (Blue Giga WT-
D3 = Rx13 + Rx31. If equations of lines L1, L2 and L3 are
12) was used which was a class 2 Bluetooth device with a
found, the error-region or probable location of ‘A’ can be
range of 10 meters and a built-in chip antenna. The module
known, as shown in Fig. 2.
was coupled with an evaluation board which was used as an
inquiry generator for all the Bluetooth receivers. A graphical
user interface (GUI) was then developed using National
Instruments LabView for calculating RSSI. The GUI was
designed in such a way that it could record all the values from
the register of the Bluetooth Module after the Inquiry cycle. In
addition to this, MAC addresses of all the Bluetooth modules
along with their location on the 15 x 15 grid could also be
entered before the start of the tests. The inquiry cycle was run
25 times at each spot, whose coordinates were manually
changed after each run.

Fig. 2 Error region depicting position of actual target B. Experiments

Since RSSI is a direct gauge of the distance, i.e., P , A grid of 15 x 15 feet was used on which five Bluetooth
modules were placed, shown in Fig. 3.
the relationship between receivers P1 and P2 can be depicted
by Eq. 1.

(1)

where PRSSI – 1 and PRSSI – 2 are the received powers at P1 and

P2, respectively. Eq. 1 can be re-written as shown in Eq. 2.

Rx12 = (2)

Similarly, Rx23, Rx31, and remaining distances can be found.

Considering the crossing points of L1(xD1,yD1) on D1,
L2(xD2,yD2) on D2 and L3(xD3,yD3) on D3, the final equations
of L1, L2 and L3 can be known as shown in Eqs. 3, 4 and 5,
respectively.
Fig. 3 A 15x15 grid for RSSI experiment

(3) These modules acted as receivers whose calibration was

already carried out. The inquiry module (WT-12 Evaluation
Board connected to the Laptop) was placed at each grid point
starting from (2,0), (4,0) and so on till (14,0) and then it was
(4) moved to (0,2), (2,2) and so on up to (15,2) and this process
continued until all the 81-grid points were covered except
(0,0), (15,0) and (0,15) and (15,15) since they were occupied
(5)
by the four receiver modules with the fifth one at (7.5,7.5). At
each point 25 inquiry recordings were made for each receiver.

V. RESULTS AND CONCLUSIONS

A. Results
(e)
Surface plots of all the five Bluetooth modules are shown
in Fig. 4. In each surface plot, strong and weak points of the Fig. 4 Surface plot of Bluetooth modules as seen from TOP (a) BT-1
Bluetooth signal strength are shown. The red area depicts the (b) BT-2 (c) BT-3 (d) BT-4 (e) BT-5
strongest signal strength whereas blue area shows the weakest
signal strength. The yellow area shows between strongest and B. Conclusions
weakest signal strength. In ideal cases, the transition from red
to yellow to blue should be smooth, with the transmitter closer Localization for visually impaired people both in the
to the receiver there should be a lot of red and once the indoor and outdoor environments is the need of today. Various
transmitter is moved further away, the surface plot should show methods of localization were discussed in this paper and it was
more blue. But in practice it is not the case due to many found that the RSSI of Bluetooth systems can further be
reasons like atmospheric conditions, environmental noise, improved by improving the existing methodologies. This
multipath fading, reflections, etc. These factors affect the paper, similarly, proposed an improved method of measuring
actual signal being received at the receiver end, which could RSSI using trilateration for localization of Bluetooth devices
make calculations much more complicated. for visually impaired people. To validate the method, class 2
Bluetooth devices were used along with an evaluation board.
A software was also developed in National Instruments
LabView. For the experimental part, five Bluetooth modules
were placed on a grid of 15 x 15 feet which acted as receivers
whose calibration was already carried out. Surface plots of all
the five Bluetooth modules were then generated. In each
surface plot, strong and weak points of the Bluetooth were
shown. It was also highlighted how many environmental
(a) conditions like, atmospheric conditions, environmental noise,
multipath fading, reflections, etc., can change the required
outputs by affecting the actual signal being received at the
receiver end.
ACKNOWLEDGMENT REFERENCES
[1] World Health Organization. (2018) Visual impairment and
blindness. [Online]. Available:
(b) [Link]
[2] R. G. Golledge, J. M. Loomis, R. L. Klatzky A. Flury, and X. L.
Yang, “Designing a personal guidance system to aid navigation
without sight: progress on the GIS component,” International
Journal of Geographical Information Systems, vol. 5(4), pp.
373-395, 1991.
[3] S. Halde and A. Ghosal, “Mobility-assisted localization
techniques in wireless sensor networks: issues, challenges and
approaches,” Cooperative Robots and Sensor Networks, pp. 43–
(c) 64, 2014.
[4] F. Pflaum, S. Erhardt, R. Weigel, and A. Koelpin, “RSSI-based
localization with minimal infrastructure using multivariate
statistic techniques,” in Proc. of IEEE Topical Conference on
Wireless Sensors and Sensor Networks, 2017, p. 69-72.
[5] R. Kowalik and S. Kwasniewski, “Navigator: A Talking GPS
Receiver for the Blind,” Lecture Notes in Computer Science,
Computers Helping People with Special Needs, vol. 3118, pp.
446-449, 2004.
(d)
[6] J. Loomis, R. Golledge, and R. Klatzky, “GPS-based navigation
systems for the visually impaired,” in Fundamentals of
Wearable Computers and Augmented Reality, W. Barfield and
T. Caudell Eds, pp. 429-446, 2001.
[7] Q. Wang, I. Balasingham, M. Zhang, and X. Huang, “Improving [23] M. Altini, D. Brunelli, E. Farella, and L. Benini, “Bluetooth
RSS-Based Ranging in LOS-NLOS Scenario Using GMMs,” indoor localization with multiple neural networks,” 5th IEEE
IEEE Communications Letters, vol. 15(10), pp. 1065–1067, International Symposium on Wireless Pervasive Computing, doi:
2011. 10.1109/ISWPC.2010.5483748, 2010.
[8] Z. Farid, R. Nordin, and M. Ismail, “Recent advances in [24] D. Li and J. Wang, “Research of indoor local positioning based
wireless indoor localization techniques and system,” Journal of on Bluetooth technology,” in 5th International Conference on
Computer Networks and Communications, doi: Wireless Communications, Networking and Mobile Computing,
10.1155/2013/185138, 2013. doi: 10.1109/WICOM.2009.5302300, 2009.
[9] Dabrowski A, Kardys P, Marciniak T (2005) Bluetooth [25] S. Bohonos, A. Lee, A. Malik, C. Thai, and R. Manduchi,
technology applications dedicated to supporting blind and “Universal real-time navigational assistance (URNA): an urban
hearing as well as speech handicapped people. In ELMAR 47th bluetooth beacon for the blind,” in Proc. of the 1st ACM
International Symposium: 295–298. SIGMOBILE international workshop on Systems and
[10] Kriz P, Maly F, Kozel T (2016) Improving indoor localization networking support for healthcare and assisted living
using Bluetooth low energy beacons. Mobile Information environments, pp. 83–88, 2007.
Systems, doi: 10.1155/2016/2083094. [26] X. Liu, H. Makino, S. Kobayashi, and Y. Maeda, “Design of an
[11] Altini M, Brunelli D, Farella E, Benini L (2010) Bluetooth Indoor Self-Positioning System for the Visually Impaired-
indoor localization with multiple neural networks. 5th IEEE Simulation with RFID and Bluetooth in a Visible Light
International Symposium on Wireless Pervasive Computing, Communication System,” in Engineering in Medicine and
doi: 10.1109/ISWPC.2010.5483748. Biology Society, EMBS, 29th Annual International Conference
of the IEEE, pp. 1655–1658, 2007.
[12] H. Chen and K. Lin, “An improved method for free-space
antenna-factor measurement by using the MUSIC algorithm,”
IEEE Transactions on Electromagnetic Compatibility, vol.
53(2), pp. 274–282, 2011.
[13] K. Agarwal and X. Chen, “Applicability of MUSIC-type
imaging in two-dimensional electromagnetic inverse problems,”
IEEE Transactions on Antennas Propagation, vol. 56(10), pp.
3217–3223, 2008.
[14] V. Beamspace, K. T. Wong, and M. D. Zoltowski, “Self-
initiating MUSIC-based direction finding in underwater acoustic
particle,” IEEE Journal of Oceanic Engineering, vol. 25(2), pp.
262–273, 2000.
[15] S. Henault, Y. M. M. Antar, S. Rajan, R. Inkol, and S. Wang,
“Impact of mutual coupling on wideband Adcock direction
finders,” Canadian conference on Electrical and Computer
Engineering, doi: 10.1109/CCECE.2008.4564755, 2008.
[16] M. Ibrahim and M. Youssef, “CellSense: An accurate energy-
efficient GSM positioning system,” IEEE Transactions on
Vehicular Technology, vol. 61(1), pp. 286–296, 2012.
[17] K. Wu, J. Xiao, S. Member, Y. Yi, and D. Chen, “CSI-based
indoor localization,” IEEE Transactions on Parallel and
Distributed Systems, vol. 24(7), pp. 1300–1309, 2013.
[18] P. K. Sahu, E. H. K. Wu, and J. Sahoo, “DuRT: Dual RSSI
trend-based localization for wireless sensor networks,” IEEE
Sensors Journal, vol. 13(8), pp. 3115–3123, 2013.
[19] Y. Chen, D. Lymberopoulos, J. Liu, B. Priyantha, “Indoor
localization using FM signals,” IEEE Transactions on Mobile
Computing, vol. 12(8), pp. 1502–1517, 2013.
[20] J. J. M. Diaz, R. D. A. Mau, R. B. Soares, E. F. Nakamura, and
C. M. S. Figueiredo, “Bluepass: an indoor Bluetooth-based
localization system for mobile applications,” IEEE Symposium
on Computers and Communications, doi:
10.1109/ISCC.2010.5546506, 2010.
[21] A. K. M. M. Hossain and W. S. Soh, “A comprehensive study of
bluetooth signal parameters for localization,” IEEE 18th
International Symposium on Personal, Indoor and Mobile Radio
Communications, doi: 10.1109/PIMRC.2007.4394215, 2007.
[22] S. Feldmann, K. Kyamakya, A. Zapater, and Z. Lue, “An indoor
Bluetooth-based positioning system: concept, implementation
and experimental evaluation,” in International Conference on
Wireless Networks, pp. 109–113, 2003.