improving indoor localization using bluetooth low energy beacons

Research ArticleImproving Indoor Localization Using BluetoothLow Energy Beacons

Pavel Kriz Filip Maly and Tomas Kozel

Department of Informatics and Quantitative Methods Faculty of Informatics and Management University of Hradec KraloveRokitanskeho 62 500 03 Hradec Kralove Czech Republic

Correspondence should be addressed to Pavel Kriz pavelkrizuhkcz

Received 7 January 2016 Revised 8 March 2016 Accepted 27 March 2016

Academic Editor Ioannis Papapanagiotou

Copyright copy 2016 Pavel Kriz et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

The paper describes basic principles of a radio-based indoor localization and focuses on the improvement of its results with the aidof a new Bluetooth Low Energy technology The advantage of this technology lies in its support by contemporary mobile devicesespecially by smartphones and tabletsWe have implemented a distributed system for collecting radio fingerprints bymobile deviceswith the Android operating system This system enables volunteers to create radio-maps and update them continuously NewBluetooth Low Energy transmitters (Apple uses its ldquoiBeaconrdquo brand name for these devices) have been installed on the floor ofthe building in addition to existingWiFi access pointsThe localization of stationary objects based onWiFi Bluetooth Low Energyand their combination has been evaluated using the data measured during the experiment in the building Several configurations ofthe transmittersrsquo arrangement several ways of combination of the data from both technologies and other parameters influencingthe accuracy of the stationary localization have been tested

1 Introduction

Nowadays when satellite navigation systems such as GPSGLONASS or Galileo are available for everyone it is usuallynot a problem to locate a person or a mobile device outsideA situation can get more complicated in high-density urbanareas with rare line-of-sight to the satellites of the corre-sponding system The situation is most complicated insidebuildings with no line-of-sight In such cases other solutionsare employed usually those based on radio networks (egIEEE 80211-WiFi) and fingerprints of signal strengths ofindividual WiFi devices which transmit their signals inside abuilding [1] Localization accuracy is influenced by a numberof circumstances for example by characteristics of trans-mitters and receivers and characteristics of the environmentwhich influence the radio signal propagation Another factorwhich can be adjusted quite easily is the number of radiotransmitters and their positions A typical situation is thatthere already are some WiFi access points in the buildingwhich more or less cover the building with the radio signalwhich can be used for localization To increase the accuracy

of the localization it is possible to install more transmitterswhich would enrich the individual fingerprints or cover theplaceswhich are poorly covered by the existingWiFi network

In this paper we will deal with the Bluetooth Low Energy(BLE) technology which can be a very good alternativesupplementing WiFi access points Their combination willallow more accurate localization The key advantage of BLEcomprises low energy consumption which allows the trans-mittersmdashcalled beaconsmdashto be powered continuously frombatteries from months to years This also makes it possible toplace the beacons in the spotswhereWiFi access pointswouldbe difficult to power

The rest of this paper is organized as follows We discussrelated work in Section 2 Section 3 describes the technol-ogy of BLE beacons (iBeacons) and deals with support ofBluetooth Low Energy with the Google Android platformSection 4 summarizes the use of BLE beacons in indoornavigation Section 5 describes the architecture of our indoorlocalization system based on BLE beacons and the localiza-tion method Section 6 is focused on the arrangement of BLEbeacons inside the building Section 7 describes the results

Hindawi Publishing CorporationMobile Information SystemsVolume 2016 Article ID 2083094 11 pageshttpdxdoiorg10115520162083094

2 Mobile Information Systems

of the evaluation In Section 8 we summarize discuss andinterpret the achieved results Section 9 concludes the paper

2 Related Work

Methods of indoor localization are usually based on moni-toring the radio signal strength the so-called Received SignalStrength Indicator (RSSI) The radio signals are broadcastedby transmitters (usually WiFi access points but eg Blue-tooth Low Energy beacons are also an option) covering aparticular area With the growing distance from the trans-mitter the received signal strength decreases and the traveltime from the transmitter to the receiver increases When wemeasure these values from more transmitters we are able toestimate a position of the receiver Two basic approaches arebeing usedmdashtriangulation and fingerprinting

21 Triangulation Methods based on triangulation can befurther divided into lateration and angulation [2] Thesemethods use estimation of the distance from several transmit-ters based on signal attenuation [3] time characteristics of thesignal propagation (TOA Time Of Arrival [4] TDOA TimeDifference of Arrival [5]) or the direction of the receivedsignal (AOA Angle of Arrival [6]) when using directionalantennas or antenna arrays All these methods achieve goodperformance in an open space with line-of-sight propagationbetween the transmitter and the receiver Unfortunatelythey have weak results inside buildings where the measuredvariables are highly influenced by the environmentThe radiosignal may be reflected and attenuated by several obstaclessuch aswallsmaking the estimation of distancemore difficult

22 Fingerprinting Fingerprinting is a localization methodcomprising two phases In the first phasemdashlearningmdashvectorscomposed of the RSSI values and optional extra featuresmeasured by a measuring device in the known locations arecollected [7]These reference valuesmdashthe calibrated datasetmdashare saved together with the location coordinates into thefingerprint database for the needs of localization In thesecond phasemdashlocalization itselfmdashthe device to be localizedmeasures the RSSI values and compares them with the datain the fingerprint database using a suitable method Themost widely used algorithms or methods of comparison andestimation of the position are [2]

(i) probabilistic methods(ii) 119896-Nearest Neighbors(iii) neural networks(iv) support vector machine(v) smallest M-vertex polygon

Concrete solutions based on collection of fingerprintsare described by Bahl and Padmanabhan [8] or Azizyan etal [9] who collect other features during the measurementsuch as sound intensity acceleration light intensity or colorof the light Wu et al [10] bring an interesting approachwhich assumes similarity between the so-called virtual andphysical model of the interior It automates the initial phase

of learning based on clustering of the fingerprints Then thevirtual rooms are mapped to the physical rooms

Localization accuracy can be increased if the movementof localized objects is considered Such methods utilise thehistory of previous measurements and estimate the positionbased on the known previous trajectory of the object Othersolutions use dead reckoning method based on collection ofdata from movement and orientation sensors of a mobiledevice (like accelerometer gyroscope and magnetometer)Thisway the direction ofmovement and the distance traveledcould be determined and combined with other measure-ments andor estimations Particle filters are often incorpo-rated in the process of gathering such estimations Particularexamples of these methods were published in [11 12]

Another approach is presented by the Ubicarse project[13] where the emulation of a large antenna array is usedfor localization purposes on a tablet device with two MIMOantennas Note that there is no public API that could readRSSI from multiple MIMO chains in high speed at theAndroid platform

23 Bluetooth-Based Localization Bluetooth-based indoorlocalization is not a novel idea [14 15] Due to the limitationof the original Bluetooth specification (now called BluetoothClassic) this approach has not been widely used The timerequired for obtaining a sufficient number of nearby Blue-tooth devices was not satisfactory due to the lengthy processof discovery Likewise energy and economic demands ofBluetooth infrastructure were high compared to WiFi-basedinfrastructure which also served other purposes

The situation changed with the advent of Bluetooth40 (including BLEBluetooth Smart) in 2010 Due to lowenergy consumption and configuration options (regardingthe advertising interval and the transmitter output power)the utilisation of this technology is much more promisingnot only in comparison with previous versions of Blue-tooth but also in comparison with todayrsquos widespread WiFi-positioning In [16] the authors focus on proximity esti-mation based on signal strength Furthermore [17] directlycompares the BLE-based localization to theWiFi-localizationby deploying BLE beacons at the same spots where WiFiaccess points were originally placed The results show thatBLE is more accurate at identical places than WiFi

In this paper we focus on an appropriate combinationof both technologies rather than on their direct comparisonWe deploy additional BLE beacons in order to improve thelocalization accuracy while utilising both technologies at thesame time

3 iBeacon Technology

iBeacon is Applersquos brand name of the technology based onthe microlocalization and the interaction of a mobile devicein the physical world This technology can be considered tobe the next development stage of the QR code technology oralternatively theNFC technology iBeacon uses the BluetoothLow Energy standard which is a part of a new versionof Bluetooth 40 Sometimes the terms Bluetooth SmartBluetooth LE BTLE and just BLE are used It is a technology

Mobile Information Systems 3

developed by Nokia (originally the technology was namedWibree in 2010 BLE was standardized) and in contrast to theprevious versions of Bluetooth dramatically lower consump-tion is typical for BLE [18 19] Also the way how the(peripheral) device announces its existence to the otherdevices is the opposite fromhow it is in the original BluetoothClassic BLE enables a peripheral device to transmit anadvertisement packet without being paged by the master(central) device Thanks to this communication model itis possible to construct energy-efficient transmittersmdashBLEbeacons or iBeacons according to Apple

iBeacon is a small device which transmits particularinformation in a defined radius and in regular intervals Assoon as a mobile device (a smartphone) gets within thisradius it can receive such information and based on this itcan perform an action Considering low consumption of BLEsuch a device can be powered by a coin battery for up to twoyears Of course the battery life depends on the transmitteroutput power (TX power) and advertising interval settings

The iBeacon technology is going to be adopted by shopmarketers A visitor with a BLE-enabled smartphone may benotified of special offers discounts information and so forthbased on hisher position or proximity to a beacon It findssimilar use in museums and exhibition halls

31 Hardware Solution BLE beacons are devices made byEstimote Kontakt Gimbal and other manufacturers [20]A beacon consists of a Bluetooth chipset (including itsfirmware) a battery providing power supply and an antennaTexas Instruments Nordic Semiconductor Bluegiga andQualcomm are the main current producers of the BLE chips

We have used beacons made by Estimote in our projectEstimote beacons can be attached to any location or objectThey broadcast BLE radio signals which can be received andinterpreted by a smartphone unlocking microlocation andcontextual awareness To be able to listen to these beaconsit is necessary to have a device that supports Bluetooth 40or higher The Estimote beacon contains an nRF51822 chipa powerful highly flexible multiprotocol System-on-a-Chip(SoC)The nRF51822 is built around a 32-bit ARM CortexM0 CPU with 256 kB128 kB flash + 32 kB16 kB RAM [21]The whole SoC is highly optimized to be energy-efficientThus the stable TX power of the beacon is ensured whilethe battery voltage may drop When the voltage finally dropsfrom 3V to 17 V a brown-out reset is generated and thedevice stops broadcasting [21] In its basic mode a beaconsimply transmits Bluetooth packets with identification datamdashso-called advertisementsmdashin regular intervals It does notcommunicate with the surrounding devices by any othersophisticatedwayAdvertisements contain the following data

(i) MAC address(ii) Universally unique identifier (UUID)mdashcommon for

a single deployment at a venue(iii) Major numbermdashdesignated for dividing the beacon

sets into smaller segments(iv) Minor numbermdashdesignated for dividing the seg-

ments into smaller subsegments

In the configuration mode beaconrsquos broadcasting param-eters (which include the above stated data transmitted in apacket and other parameters such as the TX power or theadvertising interval) can be configured In the configurationmode beacons use advanced bidirectional communicationwith a master device (eg a smartphone) with the aid ofwhich they are configured

At a physical layer BLE transmits in the 24GHz indus-trial scientific and medical (ISM) band with 40 channelseach 20MHz wide 37 channels are used to exchange thedata among paired devices and 3 channels are designatedfor broadcasting advertisements These 3 channels are thusprimarily used by beacons and are chosen deliberately so thatthey collide with the WiFi channels as little as possible Thebeacon broadcasts its advertisement packet repetitively basedon the selected advertising interval while hopping over the 3designated channels [18]

32 Android Support for BLE Android platform was chosenfor testing the whole solution because it is the most widelyused operating system for smartphones

Android offers BLE support from version 43 (APIlevel 18) From version 50 (API level 21) the BLE-related API had been revised and extracted to a separateandroidbluetoothle package The applications have tobe granted both BLUETOOTH and BLUETOOTH ADMIN sys-tem permissions to use BLE API API level 18 supportscommunication with BLE peripheral devices onlymdashthat isscanning devices enumerating devicersquos services and sendingor receiving the data to or from such devices API level 21further opens the possibility for a smartphone or a tablet(depending on hardware support) to act as a Bluetooth LowEnergy peripheral device that is to advertise itself as a BLEdevice and to offer services to other devices

Themost important function for BLE indoor localizationis scanning of the available BLE devices in the neighborhoodFor this purpose API level 18 offers startLeScan() andstopLeScan() methods of the BluetoothManager classThe scanning process is asynchronous and every devicefound is reported to an instance of the LeScanCallbackcallback class The scanned device is represented by theBluetoothDevice class which includes its MAC addressbyte-array scan record (containing UUID etc) and RSSIAPI level 21 moves the process of low energy scanninginto the separate BluetoothLeScanner class Its instanceis obtained by calling the getBluetoothLeScanner()method of the BluetoothAdapter class In contrast to APIlevel 18 it is possible to specify even more detailed para-meters of scanning Unfortunately implementation of theabove mentioned classes and underlying system libraries canvary across different vendorsThemost common issue is thatBLE devices are not reported repeatedly during the scanningprocess which is a condition necessary for localizationFor this reason it is necessary to implement a mechanismwhich starts and stops scanning repeatedly in a giventime interval It is also possible to use available librariesfor example Android Beacon Library (httpsgithubcomAltBeaconandroid-beacon-library) which providesCycledLeScanner class that encapsulates this mechanism


4 Utilising BLE in Indoor Localization

WiFi networks are commonly being used for localizationinside buildings A building is usually a complicated systemregarding WiFi signal propagation due to the materials usedThat is why areas with no WiFi signal may appear in thebuildings despite high concentration of efficient WiFi accesspoints In such areas it is not possible to collect fingerprintsbecause they would contain no signals measured from thesurrounding WiFi networks These areas can additionally becovered by other transmitters For this purpose BLE beaconscan be used They transmit a Bluetooth signal instead of aWiFi signal While powered by batteries they can be placedin less accessible places where there are no power socketsor other forms of supply such as ethernet cables allowingto use power-over-ethernet When placing the beacons itis necessary to care about the radiation pattern of a givendevice and also about possible attenuation elements in theenvironment Reference [22] deals with this topic in detail

Due to the low price of beacons their small size andindependence of an external power supply (no additionalcables required) they seem to be suitable supplements toan existing WiFi network in a building Areas covered withweakWiFi signal or with a small number ofWiFi transmitterscontained in one fingerprint can thus be enriched by newBLE transmitters Then the fingerprint can also contain themeasured RSSIs of these BLE devices in addition to the RSSIsof WiFi signals

Beacons have another advantage thanks to support bymobile operating systems they can be used for energy-efficient geofencing enabling a mobile application to beactivated based on approaching an iBeacon by a smartphoneThe whole process at the iOS platform does not require theapplication to be active and thanks to this it is possible tooptimize it so the energy consumption of the mobile deviceis minimized and the endurance of the battery is maximizedEstimote has also established a new term in this fieldmdashnearablesmdashfor their BLE beacons equipped with additionalsensors

5 Methods and Architecture

Our goal is to evaluate an improvement in the localizationusing BLE beacons In this paper we are going to compareWiFi-based stationary localization with a stationary local-ization using combination of BLE and WiFi We suggestedand performed an experiment where the originalWiFi accesspoints and additionally deployed BLE beacons were used forlocalization of a stationary device As a suitable localizationmethod we used a method based on collecting fingerprintscomposed of measured signals ofWiFi access points and BLEtransmitters

51 Learning Data Acquisition Smartphones were used toacquire the learning data and the state of their sensors(accelerometer compass and gyroscope) was also recordedfor future processing Volunteers used their smartphoneswith a digitized map of a building to acquire the learningdataset A smartphone scans signals of all available networks

and beacons around and the user creates the fingerprint of thegiven place with the aid of our application The applicationrecords strengths of individual signals in a given place for 10seconds (which should be a sufficient time [23]) A fingerprintcreated in this way is recorded into the fingerprint databaseThe rest of the system is described in Section 53

52 Positioning Method The localization inside a building isdone using collection of more fingerprints but these are notnecessarily recorded in a database The user who wishes tobe localized measures a fingerprint of a place where heshe isusing an application in hisher smartphone This fingerprintis then compared with all fingerprints in the database andone or more fingerprints with the highest similarity aresearched The fingerprints in the database are tagged withcorresponding positions inside the building The accuracy ofthe localization depends on factors such as the quality of fin-gerprints saved in the database (especially radio interferenceand the accuracy of the determination of the place wherethe tagged fingerprint was acquired) and the algorithms usedfor calculation of a similarity of the tagged fingerprint in thedatabase with the measured untagged fingerprint

To compare the measured fingerprint with the databasethe 119896-Nearest Neighbors (119896-NN) in Signal Space method wasused This method tries to find 119896 of the nearest fingerprintsfrom the database by means of for example Euclidean dis-tance In this way we get 119896 locations and by their combina-tions we estimate the position of the device to be localizedThe Euclidean distance of the measured vector of the finger-print 119898 = (119898

1 1198982 119898

119899) from the 119894th fingerprint 119878

119894=

(1199041198941 1199041198942 119904

119894119899) in the database can be expressed by the fol-

lowing formula

119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

(1)

where119873 is a number of unique transmitters in the measure-ment

After sorting the tagged fingerprints according to thedistances119863

119894from the measured fingerprint the first 119896 finger-

prints are chosen From their known positions 119875119894[119909119894 119910119894] the

weighted estimate of a position 119875 of the measured fingerprintis calculated according to the following formula

119875 =sum119896

119894=1119875119894119876119894

sum119896

119894=1119876119894

where 119876119894=1

119863119894

(2)

The Weighted 119896-Nearest Neighbors (W119896NN) in SignalSpace method was chosen especially because of its easyimplementation and the fact that its results are not difficult tointerpret If unexpected results occur they are easy to analyze

Themeasurement itself takes several seconds During themeasurement the measuring device can receive the signalof the same WiFi access points or BLE beacon several timeswith different signal strength This set of signals from onetransmitter (identified by an ID Tx eg a unique MACaddress) within onemeasurementwill bemarked119883Tx see thefollowing definition

119883Tx = 1199091Tx 1199092Tx 119909119872Tx (3)


For further processing only one value is chosen from thisset of different valuesmdashthemedian value Tx which is furthermarked as 119904 or119898 values according to its meaning in formula(1)

53 System Architecture The system for acquisition of thedata obtained during the measurements on mobile devicesis based on the Couchbase database It is a NoSQL databasewhich has no fixed schema and enables saving of recordsconstituted by objects in a JSON format under unique keysThen it supports searching primarily according to the keysand secondarily with the aid of the so-called views (theycorrespond to indexes in relational databases) which can bebased on any data from the records

One of the advantages of schemaless databases is a highflexibility when addition of more attributes in newly acquiredrecords does not require any modifications of the schemaor a major restructuring of the database This flexibility isespecially useful in research projects when a detailed analysisof all requirements at the very beginning of the project cannotbe expected

Support of replication across several servers is anotheradvantage of Couchbase Besides the server environment thisdatabase can be operated onmobile devices using the Couch-base Lite edition Couchbase Sync Gateway allows the data-base to be replicated among the server and mobile devicesincluding selective replication of selected records only Thisfeature was used for replication of the measured data frommobile devices to the server where they were further pro-cessed and the evaluation described below was performed

Direct access to Couchbase or Couchbase Sync Gatewayfrom the Internet is not recommendeddue to security reasons[24] That is why the Apache reverse proxy is put in frontof the Sync Gateway The Apache reverse proxy mediatescommunication among clients (mobile devices) and servercomponents The whole system at the server is describedin Figure 1 The small JavaScript application deployed to theNodeJS server provides external authentication of users forCouchbase SyncGateway usingGoogle accounts It facilitatesthe authentication based on userrsquos Google account whenusing hisher smartphone or tablet with the Android operat-ing system A session token is the result of the authenticationprocess

6 Test Site The Campus Building

As a test site one floor of the building of the Faculty ofInformatics and Management University of Hradec Kralove(FIM UHK) was chosen The main walk-through corridorsare in a rectangular arrangement Classrooms and offices aresituated inwards and outwards in relation to the corridorsThere is a roofed atrium in the center of the buildingExperiments have been conducted in a 52m times 43m area

Several WiFi transmitters of the eduroam network madeby Cisco are permanently deployed on every floor Theirlocations are marked with letters in Figure 2

In every place marked there are more radio units typi-cally at least two of themmdashone in a 24GHz and the other onein a 5GHz bandTheir TX power is automatically adjusted by

Apache reverse proxy

Mobile application

Google

Couchbase

NodeJS

Couchbase Sync Gateway

8091

3000auth

4985

4984beacongw

80auth80beacongw

WiFiaccess-points

BLE beacons

Figure 1 System architecture (based on [25])

the central radio resource management unit to help mitigatecochannel interference and signal coverage problems

Then 17 new Bluetooth Low Energy beacons made byEstimote were evenly placed in the corridors and classroomson the floor Beacons were originally put behind the droppedceiling (see Figure 3) in a similar way as WiFi access pointsbut the performance was not sufficient Later on we movedthem from behind the ceiling and attached them to thebottom side of the mineral fiber ceiling tile It improvedthe performance and enabled the line-of-sight propagationBeacon broadcasting parameters were set to the advertisinginterval of 100ms and the TX power of 0 dBm Locationsof beacons are marked with numbers A to 8 in Figure 2Individual beacons in the corridor are about 10m apart

7 Evaluation

The dataset of calibrated points was acquired by volunteersduring several weeks In total 680 measurements wereperformed consisting of 115511 individual RSSI samples(signal strength + transmitter-id pairs) The exact positionon the floor was known for every measurement Two deviceswere used for measurement Sony Xperia Z3 Compact andMotorola Nexus 6 Each measurement took 10 seconds

A chart in Figure 4 shows numbers of different (unique)transmitters received within one fingerprint for bothtechnologiesmdashWiFi and BLE To be complete we also showthe sum of both technologies because in the followingevaluation we will consider combination of signals from both


109

W 8 7

17

11

13

W

12

1 2 W 34

5

W

616

15

14

10m

Figure 2 Floor plan with WiFi access points and BLE beacons

types of transmitters The median of the number of uniquetransmitters is 5 for WiFi and 4 for BLE

To evaluate the localization using WiFi BLE and acombination of both technologies the leave-one-out cross-validation technique was applied From the set of 680 cali-brated points one of themwas chosen in each iteration and itsposition was estimated based on the other calibrated pointsThis procedure was then repeated for all pointsThe accuracy(estimated position compared to the real position) was thencalculated for every estimation of the position

A Weighted 119896-Nearest Neighbors in Signal Space algo-rithm was used for estimation of the position We tested 119896 isin1 2 5 Several authors recommend 119896 to be chosen as3 or 4 [23] Our results of 119896 isin 2 3 4 were similar but weachieved the highest accuracy using 119896 = 2 This value is usedin our experiments

Figure 5 shows the results of the cross-validationmdashon the119910 axis there is an accuracy of the estimation of the position(an error in meters) when using WiFi networks only BLEtransmitters only and finally both technologies combinedtogether The median accuracy improved from 1m when

using WiFi to 077m when combining both technologiesHowever the elimination of the accuracy variance and thereduction of outliers is more interestingThemaximum errorof the localization (excluding outliers) in a given sample waslowered from 427m when using WiFi to 282m in a com-bined method

We analyzed estimations with the highest errors in detailto be able to discuss the possible causes Most of the incor-rectly localized points were situated at the dead ends of thecorridors where there was no beacon at the very end ofthe corridor Localization algorithm obviously estimates theposition better when it can approximate the position betweentwo beacons Longer corridors also allow goodpropagation ofthe signal causing less significant differences among signalsespecially at the dead ends

71 Weight of BLE Signals versus WiFi Signals Several testsverifying a suitable way to combine signals of WiFi and BLEtransmitters in Signal Space have been performed BothWiFiand BLE signals were put into common Signal Space Thestrengths of BLE signals in individual tests were multiplied


Figure 3 Physical deployment of the BLE beacon

0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi

Figure 4 Number of unique transmitters in a single measurement

by different coefficients 119888 isin [02 18] The total distance inSignal Space containing both WiFi and BLE signals was thencalculated according to the formula

119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898

119873) is the measured vector of the

fingerprint of WiFi signals and 119878119894= (1199041198941 1199041198942 119904

119894119873) is the

calibrated vector of the fingerprint of WiFi signals from thedatabase Analogously 1198981015840 = (1198981015840

1 1198981015840

2 119898

1015840

1198731015840) is the mea-

sured vector of the fingerprint of BLE signals and 1198781015840119894= (1199041015840

1198941 1199041015840

1198942

1199041015840

1198941198731015840) is the calibrated vector of the fingerprint of BLE

signals from the database 119873 is the number of unique WiFi

WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)

Figure 5 Comparison of localization accuracy

transmitters in a measurement and 1198731015840 is the number ofunique BLE beacons in a measurement

These tests revealed that the best results in the given setof measurements were achieved by a ratio of 1 1 (thus acoefficient 119888 = 1) There is no reason to give more weight toone technology or the other

72 Scanning Duration Signal scanning (measuring) dura-tion in a given place is another important parameter Duringour experiment every measurement always took 10 secondsOur goal was to find out howmany seconds themeasurementshould last to provide good results

Because all the data acquired about individual measuredsignals were time-stamped even shorter scanning durationcan be considered For example the results of measurementtaking 2 seconds can be achieved by ignoring signals mea-sured after a lapse of 2 seconds (thus ignoring the remaining8 seconds from the total scanning interval) Due to the factthat we calculate the median value (119898 or 119904 values informula (4)) from several signal strengths acquired from thesame transmitter within one measurement shortening of theconsidered scanning duration does not have to necessarilymean reduction of the number of 119873 or 1198731015840 due to completeldquolossrdquo of signals of some transmitters After longer time ofmeasurement the 119883 sets from which the median values Txare calculated for each transmitter are rather smaller becauseof a reduction of the duration

We evaluated the localization again with different scan-ning duration considered from 1 second to 10 secondsFigure 6 shows the influence of the scanning duration on theaccuracy of the estimation of the position Only median andmaximum (excluding outliers) values are displayed for clarityTwo factors probably affect the accuracy first the speed ofdelivery of the measured signal strengths of transmitters tothe Android application and second the significance of theparticular technology in the localization process We noticed


Median WiFiMedian BLEMedian combined

Max WiFiMax BLEMax combined

0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110

Scanning duration (s)

Figure 6 Localization accuracy depending on scanning duration(scanning started at time 0)

faster delivery of signal strengths of the BLE transmitters incontrast to WiFi access points Thus BLE promises fasterinitial localization than WiFi does This effect becomes evenstronger in combination with WiFi For example in the 2ndsecond of the scanning we were unable to localize themobiledevice in 168 positions usingWiFi in 36 positions using BLEand in 20 positions using combination of BLE and WiFi Itis because the Android operating system may delay deliveryof the WiFi scanning results until the first scan cycle isdoneOur stationary localizationmay significantly help in theinitial phase of other methods based on distance estimationand pedestrian dead reckoning [12]

We have also investigated a situation when we estimatethe location using particular scanning duration while thescanning was started 4 seconds in advance We have chosen4 seconds because we observed a ldquowarm-uprdquo period ofapproximately 4 seconds before the scanning results werecontinuously delivered to the Android application after thescanning had been initially startedThe results have improvedenabling the localization algorithm to be applicable to mov-ing object localization while doing continuous scanning thatwas started at least 4 seconds in advance see Figure 7

73 BLE Beacons Density Density of BLE beacons will alsoinfluence the quality of the localization Due to the fact thatbeacons were firmly attached the experiment verifying theimpact of beaconsrsquo density was performed using existingdata Some beacons were not considered when processed bythe localization algorithm (ie they were ignored) in orderto simulate lower density This experiment was performedtwice For the first time only 6 beacons in the corridors

0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 70Scanning duration (s)



Figure 7 Localization accuracy depending on scanning duration(scanning started 4 s before time 0)

marked by numbersACEG0 and2 in Figure 2 wereconsideredmdashthis experiment was marked as an A optionFor the second time only 12 beacons marked A to 3 wereconsideredmdashthis experiment was marked as a B option

A boxplot in Figure 8 shows the final accuracy of theestimated position in individual cases A and B The resultsof configuration A reveal a substantial deterioration of theaccuracy of the localization in the test set when using BLEbeacons only This is understandable due to a reduction inthe number of the unique BLE beacons scanned within onemeasurement to less than 50 (in average from 44 to 19)The number of nonlocalized points thus increases

8 Discussion

Figure 8 also shows the results of the combined local-ization using WiFi and BLE beacons In this method inconfigurations A and B the median accuracy worsened from077m to 099m (A configuration) and to 087m (B con-figuration) respectively Let us remember that the medianaccuracy of the WiFi-based localization is 1m in our experi-mentThese results are also summarized in Table 1 Improve-ment in accuracy is relative to the accuracy of theWiFi-basedlocalization in the table

Thanks to addition of 17 BLE beacons the accuracy ofthe localization in a given dataset improved by 23 Besidesthe median value of the localization error the maximumerror (outliers not considered) also improved from 427m to282mThe average error improved from 181m to 108mWeassume that the number of BLE beacons scanned in onemea-surement (which was 44 on average) has the main impact


Table 1 Localization results summary

Median accuracy Improvement Improvement pctWiFi 100m NA NACombined WiFi + 6 BLE beacons conf A 099m 001m 1Combined WiFi + 12 BLE beacons conf B 087m 013m 13Combined WiFi + 17 BLE beacons (all) 077m 023m 23

A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)

Figure 8 Comparison of localization accuracy among different BLEdeployment configurations

on the resulting improvement in the localization accuracyWe also assume that by increasing the number of beaconsit will be possible to achieve more substantial improvementconsidering the observed impact of further reduction ofBLE beacons Increase of the TX power of the BLE beaconsshould also increase the number of beacons detected duringmeasurement thanks to the extended coverage and overlapof areas covered by individual beacons But this option is lessefficient for two reasons First the higher TXpowerwill resultin substantially faster discharge of batteries in the beaconsSecond a denser network of less performing beacons willincrease significant differences among individual places com-pared to a sparser network of more performing beacons

Several experiments were also performed with differentways of beaconsrsquo placement Despite the fact that placementof beacons behind a dropped ceiling is the technically easiestsolution its disadvantage is the fact that the beacons arecompletely covered by the ceiling tilesWe have also put someBLE beacons inside teachersrsquo tables in computer laboratories(marked by4 to8 in Figure 2) These tables are situated inthe front part of the laboratory and they are wooden withmetal sides Compared to the beacons with the same settingsin the dropped ceilings these beacons in the tables covered a

wider area while also completely hidden inside the table Butour original expectationwas the opposite because the ceilingswere composed of mineral fiber tiles which promised lowerattenuation thanmetal sides of the tables It was expected thatthese metal components of the tables would be a substantialobstacle for BLE signal propagation Based on this observa-tion wemoved beacons from behind the ceiling and attachedthem to the bottom side of themineral fiber ceiling tile whichimproved their performance In the future we plan to conductmore experiments with placement of individual beacons andto verify the results of the subsequent localization

Note that besides the improvement in the accuracy BLEbeacons bring another advantagemdashenergy-efficient geofenc-ingThanks to BLE it is possible to develop applicationswhichreact to approach of a mobile phone towards a beacon andwhich can bring new userrsquos experience In contrast to theAndroid operating system the iOS operating systembyAppleis more advanced in this field it has direct support of detec-tion of beacon regions while the device is in a standby mode

9 Conclusion

In this paper we have introduced a way to improve theaccuracy of the radio-based indoor stationary localizationoriginally based on WiFi signals We have designed andimplemented a distributed system for acquisition of radiofingerprints The system consists of server(s) and mobiledevices with the Android operating system which supportBluetooth Low Energy The system is designed to enablevolunteers to create a radio-map and update it continuouslyEvaluation of the solution was based on the Weighted 119896-Nearest Neighbors in Signal Space algorithm New BluetoothLow Energy transmitters by Estimote were installed on thefloor of the building where WiFi access points used bythe eduroam network had been installed before Based onthe data acquired in this real world scenario the resultsof the localization using WiFi Bluetooth Low Energy andtheir combination were evaluated We have tested severalconfigurations of positions of transmitters or their densityWe have also made experiments with how to combine signalsfrom both technologies within one Signal Space Further wehave tested the influence of the scanning duration on theaccuracy of the localization The resulting data have shownthat it is possible to improve themedian accuracy by 23 andto reduce the variance

In the future we will deal with testing the influence ofbroadcasting parameters of beacons such as the advertisinginterval and the TX power It will also be suitable to testeven higher density of beacons We will also focus on testingthe influence of other features associated with particular


measurements such as the orientation of the mobile deviceand the difference of their impact in both technologiesFurther attention will also be paid to the incorporation ofdevice movement aspects (using particle filters accordingto [11 12]) and to their potential use in the fingerprintingapproach

Competing Interests

The authors declare that they have no competing interests

Acknowledgments

The authors of this paper would like to thank DominikMatoulek and Matej Danicek students of Applied Informat-ics at the University of Hradec Kralove for implementationof the mobile applicationrsquos prototype They would also liketo thank Tereza Krizova and James Buchanan White forproofreading This work was supported by the SPEV projectfinanced from the Faculty of Informatics and ManagementUniversity of Hradec Kralove

References

[1] P Brida F Gaborik J Duha and J Machaj ldquoIndoor positioningsystem designed for user adaptive systemsrdquo in New Challengesfor Intelligent Information and Database Systems N T NguyenB Trawinski and J J Jung Eds vol 351 of Studies in Computa-tional Intelligence pp 237ndash245 Springer Berlin Germany 2011

[2] H Liu H Darabi P Banerjee and J Liu ldquoSurvey of wirelessindoor positioning techniques and systemsrdquo IEEE Transactionson Systems Man and Cybernetics Part C vol 37 no 6 pp 1067ndash1080 2007

[3] Q Dong and W Dargie ldquoEvaluation of the reliability of RSSIfor indoor localizationrdquo in Proceedings of the International Con-ference on Wireless Communications in Unusual and ConfinedAreas (ICWCUCA rsquo12 ) pp 1ndash6 Clermont-Ferrand FranceAugust 2012

[4] N Patwari J NAsh S Kyperountas AOHero III R LMosesand N S Correal ldquoLocating the nodes cooperative localizationin wireless sensor networksrdquo IEEE Signal Processing Magazinevol 22 no 4 pp 54ndash69 2005

[5] X Li K Pahlavan M Latva-aho and M Ylianttila ldquoCompari-son of indoor geolocationmethods inDSSS andOFDMwirelessLAN systemsrdquo in Proceedings of the 52nd Vehicular TechnologyConference (IEEE-VTS Fall VTC rsquo00) vol 6 pp 3015ndash3020Boston Mass USA September 2000

[6] R Peng and M L Sichitiu ldquoAngle of arrival localization forwireless sensor networksrdquo in Proceedings of the 3rd Annual IEEECommunications Society on Sensor andAdHocCommunicationsand Networks (Secon rsquo06) vol 1 pp 374ndash382 Reston Va USASeptember 2006

[7] A Zhang Y Yuan Q Wu S Zhu and J Deng ldquoWireless local-ization based on RSSI fingerprint feature vectorrdquo InternationalJournal of Distributed Sensor Networks vol 2015 Article ID528747 7 pages 2015

[8] P Bahl and V N Padmanabhan ldquoRADAR an in-building rf-based user location and tracking systemrdquo in Proceedings of theINFOCOM pp 775ndash784 2000

[9] MAzizyan I Constandache andR R Choudhury ldquoSurround-sense mobile phone localization via ambience fingerprintingrdquo

in Proceedings of the 15th Annual International Conference onMobile Computing and Networking (MobiCom rsquo09) pp 261ndash272ACM New York NY USA 2009

[10] C Wu Z Yang Y Liu andW Xi ldquoWILL wireless indoor loca-lization without site surveyrdquo IEEE Transactions on Parallel andDistributed Systems vol 24 no 4 pp 839ndash848 2013

[11] F Zafari and I Papapanagiotou ldquoEnhancing ibeacon basedmicro-locationwith particle filteringrdquo inProceedings of the IEEEGlobal Communications Conference (GLOBECOM rsquo15) pp 1ndash7IEEE San Diego Calif USA December 2015

[12] Z Chen Q Zhu H Jiang and Y C Soh ldquoIndoor localizationusing smartphone sensors and ibeaconsrdquo in Proceedings of theIEEE 10th Conference on Industrial Electronics and Applications(ICIEA rsquo15) pp 1723ndash1728 IEEE Auckland New Zealand June2015

[13] S Kumar S Gil D Katabi and D Rus ldquoAccurate indoor loca-lization with zero start-up costrdquo in Proceedings of the 20thACM Annual International Conference on Mobile Computingand Networking (MobiCom rsquo2014) pp 483ndash494 ACM MauiHawaii USA September 2014

[14] R Bruno and F Delmastro ldquoDesign and analysis of a bluetooth-based indoor localization systemrdquo in PersonalWireless Commu-nications vol 2775 of Lecture Notes in Computer Science pp711ndash725 Springer Berlin Germany 2003

[15] M Munoz-Organero P J Munoz-Merino and C DelgadoKloos ldquoUsing bluetooth to implement a pervasive indoor posi-tioning system with minimal requirements at the applicationlevelrdquoMobile Information Systems vol 8 no 1 pp 73ndash82 2012

[16] A KwiecienMMackowskiM Kojder andMManczyk ldquoReli-ability of bluetooth smart technology for indoor localizationsystemrdquo inComputer Networks P Gaj A Kwiecien and P SteraEds vol 522 of Communications in Computer and InformationScience pp 444ndash454 Springer Berlin Germany 2015

[17] X Zhao Z Xiao A Markham N Trigoni and Y Ren ldquoDoesBTLE measure up against WiFi A comparison of indoorlocation performancerdquo in Proceedings of 20th EuropeanWirelessConference pp 1ndash6 IEEE Barcelona Spain May 2014

[18] E Vlugt Bluetooth Low Energy Beacons and Retail 2013 httpglobalverifonecommedia3603729bluetooth-low-energy-bea-cons-retail-wppdf

[19] Bluetooth SIG Bluetooth Technology Basics 2015 httpswwwbluetoothcomwhat-is-bluetooth-technologybluetooth-tech-nology-basicslow-energy

[20] Aislelabs ldquoThe Hitchhikers Guide to iBeacon HardwareA Comprehensive Reportrdquo 2015 httpwwwaislelabscomreportsbeacon-guide

[21] Nordic Semiconductor nRF51822 Product Specification v312014 httpswwwnordicsemicomengProductsBluetooth-Smart-Bluetooth-low-energynRF51822

[22] J Budina O Klapka T Kozel and M Zmitko ldquoMethodof iBeacon optimal distribution for indoor localizationrdquo inModeling and Using Context H Christiansen I Stojanovicand G A Papadopoulos Eds vol 9405 of Lecture Notes inComputer Science pp 105ndash117 Springer Berlin Germany 2015

[23] V Honkavirta T Perala S Ali-Loytty and R Piche ldquoA com-parative survey of WLAN location fingerprinting methodsrdquoin Proceedings of the 6th Workshop on Positioning NavigationandCommunication (WPNC rsquo09) pp 243ndash251 IEEEHannoverGermany March 2009


[24] Couchbase ldquoDeploying Sync Gateway behind a reverse proxyrdquo2015 httpdevelopercouchbasecomdocumentationmobile110developguidessync-gatewaynginxindexhtml

[25] DMatoulek Indoor localization (Supervisor Pavel Kriz) [Bach-elor thesis] University of Hradec Kralove 2015

Submit your manuscripts athttpwwwhindawicom

Computer Games Technology

International Journal of

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Distributed Sensor Networks


Advances in

FuzzySystems

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014


ReconfigurableComputing

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014


Applied Computational Intelligence and Soft Computing

thinspAdvancesthinspinthinsp

Artificial Intelligence

HindawithinspPublishingthinspCorporationhttpwwwhindawicom Volumethinsp2014

Advances inSoftware EngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Electrical and Computer Engineering

Journal of

Journal of

Computer Networks and Communications


Hindawi Publishing Corporation

httpwwwhindawicom Volume 2014

Advances in

Multimedia


Biomedical Imaging


ArtificialNeural Systems

Advances in


RoboticsJournal of



Computational Intelligence and Neuroscience

Industrial EngineeringJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Human-ComputerInteraction

Advances in

Computer EngineeringAdvances in



of the evaluation In Section 8 we summarize discuss andinterpret the achieved results Section 9 concludes the paper

2 Related Work

Methods of indoor localization are usually based on moni-toring the radio signal strength the so-called Received SignalStrength Indicator (RSSI) The radio signals are broadcastedby transmitters (usually WiFi access points but eg Blue-tooth Low Energy beacons are also an option) covering aparticular area With the growing distance from the trans-mitter the received signal strength decreases and the traveltime from the transmitter to the receiver increases When wemeasure these values from more transmitters we are able toestimate a position of the receiver Two basic approaches arebeing usedmdashtriangulation and fingerprinting

21 Triangulation Methods based on triangulation can befurther divided into lateration and angulation [2] Thesemethods use estimation of the distance from several transmit-ters based on signal attenuation [3] time characteristics of thesignal propagation (TOA Time Of Arrival [4] TDOA TimeDifference of Arrival [5]) or the direction of the receivedsignal (AOA Angle of Arrival [6]) when using directionalantennas or antenna arrays All these methods achieve goodperformance in an open space with line-of-sight propagationbetween the transmitter and the receiver Unfortunatelythey have weak results inside buildings where the measuredvariables are highly influenced by the environmentThe radiosignal may be reflected and attenuated by several obstaclessuch aswallsmaking the estimation of distancemore difficult

22 Fingerprinting Fingerprinting is a localization methodcomprising two phases In the first phasemdashlearningmdashvectorscomposed of the RSSI values and optional extra featuresmeasured by a measuring device in the known locations arecollected [7]These reference valuesmdashthe calibrated datasetmdashare saved together with the location coordinates into thefingerprint database for the needs of localization In thesecond phasemdashlocalization itselfmdashthe device to be localizedmeasures the RSSI values and compares them with the datain the fingerprint database using a suitable method Themost widely used algorithms or methods of comparison andestimation of the position are [2]

(i) probabilistic methods(ii) 119896-Nearest Neighbors(iii) neural networks(iv) support vector machine(v) smallest M-vertex polygon

Concrete solutions based on collection of fingerprintsare described by Bahl and Padmanabhan [8] or Azizyan etal [9] who collect other features during the measurementsuch as sound intensity acceleration light intensity or colorof the light Wu et al [10] bring an interesting approachwhich assumes similarity between the so-called virtual andphysical model of the interior It automates the initial phase

of learning based on clustering of the fingerprints Then thevirtual rooms are mapped to the physical rooms

Localization accuracy can be increased if the movementof localized objects is considered Such methods utilise thehistory of previous measurements and estimate the positionbased on the known previous trajectory of the object Othersolutions use dead reckoning method based on collection ofdata from movement and orientation sensors of a mobiledevice (like accelerometer gyroscope and magnetometer)Thisway the direction ofmovement and the distance traveledcould be determined and combined with other measure-ments andor estimations Particle filters are often incorpo-rated in the process of gathering such estimations Particularexamples of these methods were published in [11 12]

Another approach is presented by the Ubicarse project[13] where the emulation of a large antenna array is usedfor localization purposes on a tablet device with two MIMOantennas Note that there is no public API that could readRSSI from multiple MIMO chains in high speed at theAndroid platform

23 Bluetooth-Based Localization Bluetooth-based indoorlocalization is not a novel idea [14 15] Due to the limitationof the original Bluetooth specification (now called BluetoothClassic) this approach has not been widely used The timerequired for obtaining a sufficient number of nearby Blue-tooth devices was not satisfactory due to the lengthy processof discovery Likewise energy and economic demands ofBluetooth infrastructure were high compared to WiFi-basedinfrastructure which also served other purposes

The situation changed with the advent of Bluetooth40 (including BLEBluetooth Smart) in 2010 Due to lowenergy consumption and configuration options (regardingthe advertising interval and the transmitter output power)the utilisation of this technology is much more promisingnot only in comparison with previous versions of Blue-tooth but also in comparison with todayrsquos widespread WiFi-positioning In [16] the authors focus on proximity esti-mation based on signal strength Furthermore [17] directlycompares the BLE-based localization to theWiFi-localizationby deploying BLE beacons at the same spots where WiFiaccess points were originally placed The results show thatBLE is more accurate at identical places than WiFi

In this paper we focus on an appropriate combinationof both technologies rather than on their direct comparisonWe deploy additional BLE beacons in order to improve thelocalization accuracy while utilising both technologies at thesame time

3 iBeacon Technology

iBeacon is Applersquos brand name of the technology based onthe microlocalization and the interaction of a mobile devicein the physical world This technology can be considered tobe the next development stage of the QR code technology oralternatively theNFC technology iBeacon uses the BluetoothLow Energy standard which is a part of a new versionof Bluetooth 40 Sometimes the terms Bluetooth SmartBluetooth LE BTLE and just BLE are used It is a technology



























1 1198982 119898


119894=

(1199041198941 1199041198942 119904


lowing formula

119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

(1)






119875 =sum119896

119894=1119875119894119876119894

sum119896

119894=1119876119894

where 119876119894=1

119863119894

(2)



119883Tx = 1199091Tx 1199092Tx 119909119872Tx (3)












Mobile application

Google

Couchbase

NodeJS


8091

3000auth

4985

4984beacongw

80auth80beacongw

WiFiaccess-points

BLE beacons




7 Evaluation




109

W 8 7

17

11

13

W

12

1 2 W 34

5

W

616

15

14

10m











0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi



119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898



119894119873) is the


1 1198981015840

2 119898

1015840

1198731015840) is the mea-


1198941 1199041015840

1198942

1199041015840



WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)










0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in





























1 1198982 119898


119894=

(1199041198941 1199041198942 119904


lowing formula

119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

(1)






119875 =sum119896

119894=1119875119894119876119894

sum119896

119894=1119876119894

where 119876119894=1

119863119894

(2)



119883Tx = 1199091Tx 1199092Tx 119909119872Tx (3)












Mobile application

Google

Couchbase

NodeJS


8091

3000auth

4985

4984beacongw

80auth80beacongw

WiFiaccess-points

BLE beacons




7 Evaluation




109

W 8 7

17

11

13

W

12

1 2 W 34

5

W

616

15

14

10m











0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi



119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898



119894119873) is the


1 1198981015840

2 119898

1015840

1198731015840) is the mea-


1198941 1199041015840

1198942

1199041015840



WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)










0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in














Mobile application

Google

Couchbase

NodeJS


8091

3000auth

4985

4984beacongw

80auth80beacongw

WiFiaccess-points

BLE beacons




7 Evaluation




109

W 8 7

17

11

13

W

12

1 2 W 34

5

W

616

15

14

10m











0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi



119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898



119894119873) is the


1 1198981015840

2 119898

1015840

1198731015840) is the mea-


1198941 1199041015840

1198942

1199041015840



WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)










0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in




109

W 8 7

17

11

13

W

12

1 2 W 34

5

W

616

15

14

10m











0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi



119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898



119894119873) is the


1 1198981015840

2 119898

1015840

1198731015840) is the mea-


1198941 1199041015840

1198942

1199041015840



WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)










0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in





0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

Num

ber o

f uni

que t

rans

mitt

ers

CombinedBLEWiFi



119863119894= radic

119873

sum

119895=1

(119898119895minus 119904119894119895)2

+

1198731015840

sum

1198951015840=1

119888 sdot (11989810158401198951015840minus 11990410158401198941198951015840)2

(4)

where 119898 = (1198981 1198982 119898



119894119873) is the


1 1198981015840

2 119898

1015840

1198731015840) is the mea-


1198941 1199041015840

1198942

1199041015840



WiFi BLE Combined0

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Loca

lizat

ion

erro

r (m

)










0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)

1 2 3 4 5 6 7 8 9 10 110






0

1

2

3

4

5

6

7

8

9

10

11

Loca

lizat

ion

erro

r (m

)







8 Discussion






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in






A-BL

E

A-co

mbi

ned

B-BL

E

B-co

mbi

ned

All-

BLE

All-

com

bine

d

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

34

36

Loca

lizat

ion

erro

r (m

)






9 Conclusion





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in





Competing Interests


Acknowledgments


References



































Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in













Advances in

FuzzySystems


Volume 2014












Journal of

Journal of





Advances in

Multimedia


Biomedical Imaging



Advances in


RoboticsJournal of










Advances in



improving indoor localization using bluetooth low energy beacons

Documents