And Up-Converters in 433 Mhz ISM Band

sensors

Article Indoor Positioning System Based on Global Positioning System Signals with Down- and Up-Converters in 433 MHz ISM Band

Abdulkadir Uzun 1,2 , Firas Abdul Ghani 1 , Amir Mohsen Ahmadi Najafabadi 1, Hüsnü Yenigün 3 and Ibrahim˙ Tekin 1,*

1 Electronics Engineering, Sabanci University, Istanbul 34956, Turkey; [email protected] (A.U.); ﬁ[email protected] (F.A.G.); [email protected] (A.M.A.N.) 2 Radar Electronic Warfare and Intelligence Systems Division, ASELSAN A.¸S.,Istanbul 34906, Turkey 3 Computer Science and Engineering, Sabanci University, Istanbul 34956, Turkey; [email protected] * Correspondence: [email protected]; Tel.: +90-(216)-483-9534

Abstract: In this paper, an indoor positioning system using Global Positioning System (GPS) signals in the 433 MHz Industrial Scientific Medical (ISM) band is proposed, and an experimental demonstration of how the proposed system operates under both line-of-sight and non-line-of-sight conditions on a building floor is presented. The proposed method is based on down-converting (DC) repeaters and an up-converting (UC) receiver. The down-conversion is deployed to avoid the restrictions on the use of Global Navigation Satellite Systems (GNSS) repeaters, to achieve higher output power, and to expose the GPS signals to lower path loss. The repeaters receive outdoor GPS signals at 1575.42 MHz (L1 band), down-convert them to the 433 MHz ISM band, then amplify and retransmit them to the indoor environment. The front end up-converter is combined with an off-the-shelf GPS receiver. When GPS signals at 433 MHz are received by the up-converting receiver, it then amplifies and up-converts these signals back to the L1 frequency. Subsequently, the off-the-shelf GPS receiver Citation: Uzun, A.; Ghani, F.A.; calculates the pseudo-ranges. The raw data are then sent from the receiver over a 2.4 GHz Wi-Fi link Ahmadi Najafabadi, A.M.; Yenigün, H.; Tekin, I.˙ Indoor Positioning to a remote computer for data processing and indoor position estimation. Each repeater also has an System Based on Global Positioning attenuator to adjust its amplification level so that each repeater transmits almost equal signal levels System Signals with Down- and in order to prevent jamming of the off-the-shelf GPS receiver. Experimental results demonstrate that Up-Converters in 433 MHz ISM Band. the indoor position of a receiver can be found with sub-meter accuracy under both line-of-sight and Sensors 2021, 21, 4338. https:// non-line-of-sight conditions. The estimated position was found to be 54 and 98 cm away from the doi.org/10.3390/s21134338 real position, while the 50% circular error probable (CEP) of the collected samples showed a radius of 3.3 and 4 m, respectively, for line-of-sight and non-line-of-sight cases. Academic Editor: Simon Tomažiˇc Keywords: down-conversion; GPS; indoor positioning; navigation; RF repeaters; up-conversion Received: 26 May 2021 Accepted: 18 June 2021 Published: 25 June 2021 1. Introduction Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in Global indoor positioning is an emerging market whose size is forecast to grow published maps and institutional affil- from USD 6.1 billion to USD 17.0 billion by 2025 [1]. Many researchers, in the field of iations. indoor positioning, have proposed different solutions to solve this sophisticated problem. Some of the proposed technologies for indoor positioning are based on IEEE 802.11 [2–7], Bluetooth [8–11], Zigbee [12], radio frequency identification devices (RFIDs) [13], visible light [14], acoustic [15,16], and ultrasound [17,18]. GNSS-based solutions have also proven to be a candidate for the indoor positioning problem [19–21]. Some of the GNSS-based Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. solutions are based on high sensitivity GNSS (HS-GNSS) [22], assisted GNSS (A-GNSS) [23], This article is an open access article pseudolites [24–31], GNSS repeaters [32,33], repealites [34,35], and peer-to-peer cooperative distributed under the terms and positioning [36]. Among the GNSS-based techniques, HS-GNSS and A-GNSS technologies conditions of the Creative Commons require no infrastructure within the indoor environment, while pseudolite and repeater- Attribution (CC BY) license (https:// based approaches require infrastructure. creativecommons.org/licenses/by/ Ideally, from a navigational point of view, a GPS-based solution to the indoor posi- 4.0/). tioning problem would result in integrated outdoor and indoor applications such as asset

Sensors 2021, 21, 4338. https://doi.org/10.3390/s21134338 https://www.mdpi.com/journal/sensors Sensors 2021, 21, x FOR PEER REVIEW 2 of 24

Sensors 2021, 21, 4338 tive positioning [36]. Among the GNSS-based techniques, HS-GNSS and A-GNSS tech-2 of 23 nologies require no infrastructure within the indoor environment, while pseudolite and repeater-based approaches require infrastructure. Ideally, from a navigational point of view, a GPS-based solution to the indoor posi- tioningtracking, problem vehicular would navigation, result in integrated and emergency outdoor services. and indoor However, applications GPS-based such as asset solutions tracking,currently vehicular experience navigation, a multitude and of emergency weaknesses services. due to However, (a) reflections GPS- andbased multipath solutions errors currentlyfrom the ex indoorperience environment, a multitude of (b) weaknesses attenuation due (up to (a) to reflections 30 dB [37 and]) from multipath walls errors and other fromstructural the indoor components, environment, (c) non-line-of-sight (b) attenuation (up conditions to 30 dB caused [37]) from by wallscorners and and other objects, (d)structural changes components, in the indoor (c) non environment-line-of-sight due conditions to people caused and movingby corners objects, and objects, and (e) (d) RF de- changesvices operating in the indoor in the environment same frequency due bandto people and and their moving interference objects, to and the ( indoore) RF devices positioning operatingsystem. in the same frequency band and their interference to the indoor positioning system. In addition to the listed drawbacks, GPS-based systems operating in L1, L2, and L5 bandsIn alsoaddit sufferion to the from listed the drawbacks, restrictions GPS on- thebased use systems of GNSS operating repeaters. in L1, Such L2, and restrictions L5 bandsaim to also prevent suffer repeatersfrom the restrictions from interfering on the use with of otherGNSS GNSSrepeaters. systems Such restrictions in the vicinity. aim The toElectronic prevent repeaters Communications from interfering Committee with (ECC),other GNSS European systems Telecommunication in the vicinity. The StandardsElec- trInstituteonic Communications (ETSI), and the Committee US policy (ECC), “Manual European of Regulations Telecommunication and Procedures Standards for In- Federal stituteRadio (ETSI), Frequency and the Management” US policy “Manual present of theRegulations practices and and Procedures restrictions for on Federal the use Radio of GNSS Frequency Management” present the practices and restrictions on the use of GNSS repeat- repeaters [38–41]. These standards reduce the coverage of GPS repeaters by limiting the ers [38–41]. These standards reduce the coverage of GPS repeaters by limiting the output output power and the maximum allowable amplification for repeater structures in the L1, power and the maximum allowable amplification for repeater structures in the L1, L2, and L2, and L5 frequency bands. L5 frequency bands. In [32,42], a GPS-based indoor positioning technique with three repeaters is shown; In [32,42], a GPS-based indoor positioning technique with three repeaters is shown; however, the the repeaters repeaters operate operate solely solely in inthe the L1 L1band band and and,, therefore, therefore, are restricted are restricted to the to the aforementioned restrictive restrictive policies policies in intheir their usage. usage. ThisThis paper presents a a new new GPS-based GPS-based approach approach that that does does not not contradict contradict the the restrictions re- strictionsin the aforementioned in the aforementioned standards. standards. The proposed The proposed indoor indoor positioning positioning system system in this in paper thisoperates paper in operates the 433 in MHz the ISM433 MHz band, ISM hence band, it is hence not subject it is not to subject the amplification to the amplification restrictions in restrictionsGPS frequencies. in GPS The frequencies. proposed The repeaters proposed down-convert repeaters down GPS-convert signals GPS in the signals 1575.42 in MHzthe (L1 1575.42band) to MHz the 433 (L1 MHz band) ISM to the band, 433 allowing MHz ISM signal band, coverage allowing to signal be increased, coverage with to be the in- higher creased,permitted with power the higher levels permitted in the 433 power MHZ levels ISM band.in the In433 addition MHZ ISM to band. the higher In addition power to levels thepermitted higher power in the levels 433 MHz permitted ISM band, in the the 433 free MHz space ISM pathband, loss the infree the space 433 path MHz loss frequency in theis 11.22 433 MHz dB less, frequency as presented is 11.22 indB Figure less, as1, presented while the in penetration Figure 1, while through the penetration walls is higher throughthan at 1575.42 walls is MHz higher (GPS than frequency) at 1575.42 or MHz 2.4 GHz (GPS (Wi-Fi frequency) frequency). or 2.4 Therefore, GHz (Wi-Fi in fre- terms of quency).coverage Therefore, and compatibility in terms withof coverage existing and rules compatibility on the use with of repeaters, existing therules proposed on the use system ofcould repeaters, outperform the proposed the existing system indoor could positioning outperform systems the existing that areindoor operating positioning at frequencies sys- temshigher that than are 433operating MHz. at frequencies higher than 433 MHz.

Figure 1. Free space path loss at 2.4 GHz, 1575.42 MHz, and 433 MHz. Figure 1. Free space path loss at 2.4 GHz, 1575.42 MHz, and 433 MHz.

TheThe system system we we propose propose in inthis this paper paper differs differs from from the theexisting existing systems systems with withits fol- its following properties. properties. The The down down-- and and up up-conversion-conversion scheme schemess have have previously previously been been pro- pro- posed in [43–45] for indoor positioning applications with GPS signals. While in [43,44], a posed in [43–45] for indoor positioning applications with GPS signals. While in [43,44], a down-converting repeater is proposed, neither indoor positioning nor GPS signal down- conversion is demonstrated. In [45], only the down-converting repeater and up-converting receiver circuits are presented, to show that GPS signal ﬁdelity is preserved. However, in [45], it is not shown how the proposed circuits perform in an indoor environment for positioning. In [46], the indoor positioning of an up-converting receiver at a point on a line between two down-converting repeaters (1-dimensional (1D) positioning) is proposed; Sensors 2021, 21, x FOR PEER REVIEW 3 of 24

down-converting repeater is proposed, neither indoor positioning nor GPS signal down- conversion is demonstrated. In [45], only the down-converting repeater and up-convert- Sensors 2021, 21, 4338 3 of 23 ing receiver circuits are presented, to show that GPS signal fidelity is preserved. However, in [45], it is not shown how the proposed circuits perform in an indoor environment for positioning. In [46], the indoor positioning of an up-converting receiver at a point on a however,line between non-line-of-sight two down-converting conditions repeaters and two-dimensional (1-dimensional (2D) (1D) positioning positioning) of the is pro- receiver areposed not; however, addressed. non In-line [47-],of we-sight refer con toditions the patent and two application-dimensional of (2D) a system positioning that can of bethe used withreceiver different are not positioning addressed. systemsIn [47], we (i.e., refer BeiDou, to the Galileo,patent application GLONASS) of bya system changing that the can regis- terbe used sets inwith the different down- andpositioning up-converters systems to(i.e. work, BeiDou, in the Galileo, 433 MHz GLONASS) frequency by changing band or even inthe other register ISM sets bands. in the In down the latter- and up case,-converters the indoor to work antennas in the should 433 MHz also frequency be changed band to the or even in other ISM bands. In the latter case, the indoor antennas should also be changed selected frequency band. to the selected frequency band. Upon review of readily available publications and to the best of the authors’ knowl- Upon review of readily available publications and to the best of the authors’ edge, this paper is the ﬁrst experimental study that demonstrates that 2D indoor positioning knowledge, this paper is the first experimental study that demonstrates that 2D indoor can be achieved by transmitting GPS signals, at 433 MHz, from three repeaters in an indoor positioning can be achieved by transmitting GPS signals, at 433 MHz, from three repeaters environment,in an indoor environment, where the position where the of the position receiver of the can receiver be estimated can beby estimated calculating by calculat- the distance betweening the distance each repeater between and each the repeater receiver, and on the the receiver, plane thaton the is formedplane that by is the formed deployed by repeaters.the deployed The indoor repeaters. positioning The indoor concept positioning that is concept proposed that in is this proposed paper is in described this paper visually is indescribed Figure2 .visually in Figure 2.

FigureFigure 2. The proposed proposed ind indooroor positioning positioning system. system. : indexi: index of the of therepeaters repeaters ( = 1, (i 2,= 3); 1, 2,: 3);indexj: index of of the satellites ( = 1, 2, …, ); : th repeater; : th satellite; : indoor distance from repeater the satellites (j = 1, 2, ... , n ); Ri: ith repeater;Sj: jth satellite; di: indoor distance from repeater Ri to to the receiver (with a clock bias ofreceiver ); Sj : distance from ith repeater (:) to the th satellite the receiver (with a clock bias of tbias ); DRi: distance from ith repeater (Ri:) to the jth satellite (Sj ); (); : propagation delay of the ith repeater;Sj : satellite clock bias of the th satellite. τi: propagation delay of the ith repeater; tbias: satellite clock bias of the jth satellite. In this particular paper, we present two experiments for 2D indoor positioning in the 433 MHzIn this ISM particular band: one paper, experiment we present is under two line experiments-of-sight conditions, for 2D indoor and another positioning is un- in the 433der MHznon-line ISM-of band:-sight oneconditions. experiment Note isthat under adding line-of-sight a 4th repeater conditions, will allow and us another to achieve is under non-line-of-sight3D indoor positioning. conditions. Note that adding a 4th repeater will allow us to achieve 3D indoorThe positioning. rest of the paper is organized as follows: Section 2 introduces the proposed in- doorThe positioning rest of the system paper and is organized describes asits follows:hardware, Section software2 introduces and algorithms. the proposed Section indoor 3 positioningdescribes the experimental system and framework describes and the its real hardware,-life environm softwareent where andwe tested algorithms. our Sectionsystem.3 Section describes 4 analyzes the experimental and discusses framework the results we and achieved the real-life when the environment proposed sys- where wetem tested and technique our system. are used Section for 42D analyzes indoor positioning. and discusses Finally, the resultsSection we5 concludes achieved this when thepaper. proposed system and technique are used for 2D indoor positioning. Finally, Section5 concludes this paper.

2. Indoor Positioning System in 433 MHz ISM Band The proposed indoor positioning system is composed of two unique subsystems: GPS down-converting repeaters and an up-converter integrated with an off-the-shelf GPS receiver. The block diagrams of the GPS down-converting repeaters and the up-converter with the off-the-shelf GPS receiver are presented in Figure3. The red dashed line represents the 433 MHz RF link. Sensors 2021, 21, x FOR PEER REVIEW 4 of 24

2. Indoor Positioning System in 433 MHz ISM Band The proposed indoor positioning system is composed of two unique subsystems: GPS down-converting repeaters and an up-converter integrated with an off-the-shelf GPS Sensors 2021, 21, 4338 receiver. The block diagrams of the GPS down-converting repeaters and the up-converter4 of 23 with the off-the-shelf GPS receiver are presented in Figure 3. The red dashed line represents the 433 MHz RF link.

FigureFigure 3. 3.RF RF blocks blocks in in down-convertingdown-converting repeater ( (leftleft) )and and up up-converting-converting receiver receiver (right (right). ).

2.1.2.1. GPS GPS Down-ConvertingDown-Converting Repeater Repeater TheThe GPSGPS down-convertingdown-converting repeater repeater subsystem subsystem comprise comprisess RF RFblocks blocks and and supporting supporting blocks,blocks, which which are are presented presented in in Table Table1. 1. The The RF RF blocks blocks refer refer to to the the components components that that together to- formgether the form RF path,the RF through path, through which which the GPS the signalGPS signal propagates, propagates, while while the supportingthe supporting blocks sustainblocks sustain the regular the regular operation operation of RF of blocks. RF blocks. The GPSThe GPS down-converting down-converting repeater repeater subsystem sub- issystem implemented is implemented using the using listed the components listed components and blocks and blocks in Table in Table1. 1.

TableTable 1. 1.Repeater Repeater subsystemsubsystem and and its its components components..

BlocksBlocks Components Components Block Block Properties Properties TangoTango 20 off 20-the off-the-shelf-shelf active active GPS antenna GPS antenna 34.45 34.45dBi gain dBi gain DirectionalDirectional GPSGPS Antenna withwith conic conic reflector reflector 60 degree60 degree 3 dB beam 3 dB beamwidth width ProvidesProvides DC to GPS DC antenna to GPS and antenna transmits and transmitsRF to down RF to DirectionalDirectional GPSGPS Antenna TB-JEBT TB-JEBT-4R2GW-4R2GW + bias + tee bias tee converterdown converter 8.7 dB8.7 block dB blockloss loss Down-ConverterDown-Converter ADRF6820ADRF6820 quadrature quadrature demodulator, demodulator, ProgrammableProgrammable over SPI over from SPI controller from controller block block Down-ConverterDown-Converter ZX10Q ZX10Q-2-5-S-2-5-S 90 degree 90 degree power power combiner combiner Combines Combines down down-converter-converter I/Q outputs I/Q outputs 22.43 22.43dB gain dB gain SignalSignal PowerPower ConditionerConditioner LHALHA-13LN-13LN + low + noise low noiseamplifier amplifier 0.9 dB0.9 noise dB noisefigure figure DATDAT-31R5A-SP-31R5A-SP + digital + digital step attenuator step attenuator 0 to 31.5 dB0 to adjustable 31.5 dB adjustable attenuation attenuation by 0.5 dB step by 0.5size dB and step SignalSignal PowerPower ConditionerConditioner (variable(variable attenuator) attenuator) cansize be and adjusted can be from adjusted controller from controller 1.7 dB insertion loss at 433 MHz, 19 MHz 3 dB SignalSignal PowerPower ConditionerConditioner DBP.433.T.A.30DBP.433.T.A.30 band band pass passfilter filterat 433 at MHz 433 MHz1.7 dB insertion loss at 433 MHz, 19 MHz 3 dB bandwidth LHA-13LN + as the second low noise 22.43 dBbandwidth gain Signal Power Conditioner LHA-13LN + as the second low noise 22.43 dB gain Signal Power Conditioner amplifier 0.9 dB noise figure amplifier 2.10.9 dBi dB gain noise figure 433 MHz Dipole 433 MHz dipole on FR4 2.1 dBi gain 433 MHz Dipole 433 MHz dipole on FR4 80 degree 3 dB beam width on both sides 80 degree 3 dB beam width on both sides Supporting Blocks Voltage Regulator Provides DC to system components and blocks Supporting Blocks Voltage Regulator Provides DC to system components and blocks Controller with Wi-Fi for programming Wi-Fi connection to remote PC, SPI connection to Supporting Blocks Controller with Wi-Fi for programming Wi-Fi connection to remote PC, SPI connection to Supporting Blocks attenuator and down-converter ADRF6820, and attenuator attenuator and down-converter ADRF6820, and attenuator The implemented GPS down-converting repeater is depicted in Figure 4. Fabricated and Themodelled implemented 1575.42 MHz GPS outdoor down-converting directional GPS repeater antenna is depicted and 433 MHz in Figure dipole4. antenna Fabricated andare modelleddemonstrated 1575.42 in Figure MHz outdoor5. The RF directional blocks of the GPS GPS antenna down and-converting 433 MHz repeater dipole sub- antenna aresystem demonstrated are as follows; in Figure a directional5. The RF outdoor blocks GPS of the antenna GPS down-converting (proposed in [32]) repeater and its bias subsys-- temtee, are a 1575.42 as follows; to 433 a directionalMHz down outdoor-converter, GPS a antennasignal power (proposed conditioner in [32 ])and and filter its bias-tee,block a 1575.42(where toan 433 LNA, MHz a variable down-converter, attenuator, a a signal 433 MHz power band conditioner pass filter, andand filtera second block LNA (where are an LNA,cascaded), a variable and an attenuator, indoor 433 a MHz 433 MHzdipole band antenna. pass The filter, supporting and a second blocks LNA are a are controller cascaded), and an indoor 433 MHz dipole antenna. The supporting blocks are a controller over Wi-Fi that interfaces between the RF blocks and the user, and a voltage regulator that provides the required DC to RF blocks. The active directional GPS antennas pick-up the GPS signals from satellites in the 3 dB beam width. The conic reflector reduces the beam width by 30 degrees. Therefore, the 3 dB beam width of the fabricated antenna is measured as 60 degrees. A bias-tee is placed before the down-converter to provide the DC voltage that is required by the active directional GPS antenna. SensorsSensors 2021 2021, 21, 21, x, xFOR FOR PEER PEER REVIEW REVIEW 5 5of of 24 24

overover Wi Wi-Fi-Fi that that interfaces interfaces between between the the RF RF blocks blocks and and the the user, user, and and a avoltage voltage reg regulatorulator that that providesprovides the the required required DC DC to to RF RF blocks. blocks. TheThe active active directional directional GPS GPS antennas antennas pick pick-up-up the the GPS GPS signals signals from from satellites satellites in in the the 3 3 dBdB beam beam width width. The. The conic conic reflector reflector reduces reduces the the beam beam width width by by 30 30 d egrees.degrees. Therefore, Therefore, the the Sensors 2021, 21, 4338 3 3dB dB beam beam width width of of the the fabricated fabricated antenna antenna is is measured measured as as 60 60 degrees. degrees. A A bias bias-tee-tee is is placed placed5 of 23 beforebefore the the down down-converter-converter to to provide provide the the DC DC voltage voltage that that is is required required by by the the active active direc- direc- tionaltional GPS GPS antenna. antenna.

FigureFigureFigure 4. 4. ImplementedImplemented Implemented GPS GPSGPS down down-convertingdown-converting-converting repeater repeater repeater subsystem subsystem subsystem excluding excluding excluding the the ante the antennas antennas.nnas. .

FigureFigureFigure 5. 5.Modelled 5. Modelled Modelled directional directional directional GPS GPS GPS antenna antenna antenna ( a(,a b(,ab),,),b fabricated ),fabricated fabricated directional directional directional GPSGPS GPS antennaantenna antenna ((cc ),(),c ), modelledmodelled modelled 433 433 MHzMHz MHz dipole dipole an- antenna an- (d),tennatenna fabricated (d (),d ),fabricated fabricated 433 MHz 433 433 dipole MHz MHz dipole antenna dipole antenna antenna (e). (e ().e ).

WhenWhen the the the GPS GPS GPS signal signalsignal at at 1575.42 1575.42 MHz MHz MHz is is received is received received by by bythe the theantenna, antenna, antenna, the the signal the signal signal passes passes passes to to to thethethe down-converter.down down-converter.-converter. The The down down down-converter-converter-converter converts converts converts the the thesignal signal signal from from from 1575.42 1575.42 1575.42 to to 433 433 to MHz. 433MHz. MHz. TheTheThe amplifiersamplifiers amplifiers inin in thethe the signalsignal signal powerpower power conditionerconditioner conditioner and and filterfilter filter block block then then amplify amplify the the signal. signal. The amplificationTheThe amplification amplification level level canlevel becan can adjusted be be adjusted adjusted by changing by by changing changing the attenuation the the attenuation attenuation of the of digitalof the the digital stepdigital attenuator, step step whichattenuator,attenuator, is able which which to attenuate is is able able to theto attenuate attenuate signal fromthe the signal 0signal to 31.5 from from dB. 0 0to By to 31.5 adjusting31.5 dB. dB. By By theadjusting adjusting attenuation the the atten- atten- value of eachuationuation attenuator, value value of of each each the attenuator, gain attenuator, of the the overallthe gain gain of down-converter of the the overall overall down down subsystem-converter-converter maysubs subsystem beystem set may to may different be be set to different values. The bandpass filter at 433 MHz is also deployed to eliminate signals values.set to different The bandpass values. filter The bandpass at 433 MHz filter is alsoat 433 deployed MHz is also to eliminate deployed signalsto eliminate and harmonicssignals andand harmonics harmonics out out of of the the band. band. Furthermore, Furthermore, the the signal signal power power conditioner conditioner block block has has an- an- out of the band. Furthermore, the signal power conditioner block has another amplifier otherother amplifier amplifier following following the the band band pass pass filter. filter. After After the the second second amplifier, amplifier, the the down down-con--con- following the band pass filter. After the second amplifier, the down-converted GPS signals vertedverted GPS GPS signals signals are are retransmitted retransmitted to to indoors indoors via via a a433 433 MHz MHz dipole dipole antenna antenna at at each each are retransmitted to indoors via a 433 MHz dipole antenna at each repeater. repeater.repeater. In terms of adjustments, the repeater board attenuation levels can be set to different InIn terms terms of of adjustments, adjustments, the the repeater repeater board board attenuation attenuation levels levels can can be be set set to to different different valuesvaluesvalues fromfrom from aa remotearemote remote computercomputer computer usingusing using aa a Wi-FiWi Wi-Fi-Fi connection.connec connection.tion. In In this this way, way, thethe the gaingain gain ofof of aa repeaterare- re- canpeaterpeater be adjustedcan can be be adjusted adjusted to prevent to to prevent prevent near-far near near effects-far-far effects andeffects also and andkeep also also keep the keep signal the the signal signal level level fromlevel from eachfrom each repeatereach atrepeaterrepeater a similar at at a level. asimilar similar With level. level. this With With structure, this this structure, structure, the down-conversion the the down down-conversion-conversion of GPS of of GPS signals GPS signals signals enables enables enables users to deployusersusers to to higher deploy deploy gain higher higher GPS gain gain receivers GPS GPS rece rece thaniversivers those than than that those those are that limited that are are limited by limited international by by international international standards. Additionally, operating in 433 MHz allows us to further increase the gain of the proposed repeaters.

2.2. Up-Converting Receiver The up-converting receiver subsystem is comprised of the RF blocks and supporting blocks, which are presented in Table2. Sensors 2021, 21, 4338 6 of 23

Table 2. Receiver subsystem and its components.

Blocks Components Block Properties 2.1 dBi gain, 433 MHz Dipole 433 MHz dipole on FR4 80 degree 3 dB beam width on both sides 1.7 dB insertion loss at 433 MHz, 19 MHz 3 dB Signal Power Conditioner DBP.433.T.A.30 band pass filter at 433 MHz bandwidth 22.43 dB gain Signal Power Conditioner LHA-13LN + low noise amplifier 0.9 dB noise figure DAT-31R5A-SP + digital step attenuator 0 to 31.5 dB adjustable attenuation by 0.5 dB step Signal Power Conditioner (variable attenuator) size and can be adjusted from controller LHA-13LN + as the second low noise 22.43 dB gain Signal Power Conditioner amplifier 0.9 dB noise figure Provides I/Q inputs and required DC levels to Up-Converter I/Q power divider, bias tees ADRF6720-27 1.2 dB block loss Up-Converter ADRF6720-27 quadrature modulator Programmable over SPI from controller block Off-the-shelf GPS Receiver LEA-6T chipset Supporting Blocks Voltage Regulator Provides DC to system components and blocks Controller with Wi-Fi for programming Wi-Fi connection to remote PC, SPI connection to Supporting Blocks attenuator and down-converter ADRF6820, and attenuator

The up-converting receiver is implemented using the listed components and blocks in Table2 and is presented in Figure6. The aforementioned 433 MHz dipole antenna in Figure5e is also used as the receiver indoor antenna. The RF blocks of the up-converting receiver are (a) a 433 MHz indoor dipole antenna, (b) a signal power conditioner block where a 433 MHz band pass ﬁlter, (c) an LNA, (d) a variable attenuator, (e) a second, cascaded LNA, (f) an I/Q power divider, (g) bias tees that provide the required DC to I/Q inputs of the up-converter, (h) a 433 to 1575.42 MHz up-converter, and (i) an off-the-shelf GPS receiver (u-Blox LEA-6T®). The controller over Wi-Fi provides interfaces between RF blocks and the user. Moreover, a voltage regulator is designed to provide the required DC voltage to the RF blocks. In this subsystem, the controller is also connected to the custom Sensors 2021, 21, x FOR PEER REVIEWGPS receiver, and the raw data from the GPS receiver can be sent over Wi-Fi to7 aof remote24

computer.

FigureFigure 6. ImplementedImplemented up up-converting-converting receiver receiver excluding excluding the theantenna antenna..

The retransmitted 433 433 MHz MHz positioning positioning signals signals are are picked picked up by up the by 433 the MHz 433 MHz indoor indoor dipoledipole antennas. In In the the receiver receiver subsystem, subsystem, the the received received signals signals are first are firstfiltered filtered and then and then amplified.amplified. Sim Similarilar to to the the repeater repeater subsystem, subsystem, thethe amplification amplification levels levels are adjustedare adjusted by by changingchanging the attenuation attenuation of of the the digital digital step step attenuator, attenuator, which which is able is able to attenuate to attenuate the signal the signal from 0 to 31.5 dB. By adjusting the attenuation value of the attenuator, the gain of the overall up-converter subsystem may be set to different values. The subsystem also utilizes a second amplifier to further amplify the signal. These up-converted GPS signals then propagate to the off-the-shelf GPS receiver, which is integrated within the up-converting receiver subsystem. As mentioned previously, for further flexibility, the up-converting repeater attenuation levels can be set to different values from a remote computer using a Wi-Fi connection. Using a remote computer with an internet connection, where the ephemeris data of GPS satellites are available, the estimation of the indoor position starts when the raw data are sent from the receiver. Therefore, one can conclude that the established Wi-Fi link between the proposed up-converting receiver, and the computer where calculations are done will provide a hot start to the system.

2.3. Algorithm for Indoor Position Estimation In an indoor environment, although the signal loss and GPS coverage problems can be overcome with the proposed repeaters, this solution requires additional algorithms that takes the non-line-of-sight propagation and repeater delay into account. The proposed technique introduces a new path in that the distance between satellite and the receiver becomes different from that of the normal operation of an off-the-shelf receiver during which there is a line-of-sight distance between the satellite and the receiver. The GPS signals in the proposed scheme come to the repeater first and, then, reach to the receiver as seen in Figure 2. As part of the indoor position estimation, an algorithm (run on MATLAB®) is designed and run on a remote computer that is connected to the system via Wi-Fi. The Wi- Fi link is established with a Wi-Fi module on the controller block of the receiver board. The raw data (such as pseudo-range, carrier-to-noise ratio, satellite azimuth, and elevation angles, etc.) obtained from the off-the-shelf receiver are transmitted to the remote computer through the Wi-Fi connection. The routine summarized in Figure 7 has the following four steps: raw data reception, satellite selection and satellite-to-repeater distance calculations, cleaning pseudo-range from satellite-to-repeater distances and satellite biases, and

Sensors 2021, 21, 4338 7 of 23

from 0 to 31.5 dB. By adjusting the attenuation value of the attenuator, the gain of the overall up-converter subsystem may be set to different values. The subsystem also utilizes a second ampliﬁer to further amplify the signal. These up-converted GPS signals then propagate to the off-the-shelf GPS receiver, which is integrated within the up-converting receiver subsystem. As mentioned previously, for further ﬂexibility, the up-converting repeater attenuation levels can be set to different values from a remote computer using a Wi-Fi connection. Using a remote computer with an internet connection, where the ephemeris data of GPS satellites are available, the estimation of the indoor position starts when the raw data are sent from the receiver. Therefore, one can conclude that the established Wi-Fi link between the proposed up-converting receiver, and the computer where calculations are done will provide a hot start to the system.

2.3. Algorithm for Indoor Position Estimation In an indoor environment, although the signal loss and GPS coverage problems can be overcome with the proposed repeaters, this solution requires additional algorithms that takes the non-line-of-sight propagation and repeater delay into account. The proposed technique introduces a new path in that the distance between satellite and the receiver becomes different from that of the normal operation of an off-the-shelf receiver during which there is a line-of-sight distance between the satellite and the receiver. The GPS signals in the proposed scheme come to the repeater ﬁrst and, then, reach to the receiver as seen in Figure2. As part of the indoor position estimation, an algorithm (run on MATLAB®) is designed and run on a remote computer that is connected to the system via Wi-Fi. The Wi-Fi link is established with a Wi-Fi module on the controller block of the receiver board. The raw data (such as pseudo-range, carrier-to-noise ratio, satellite azimuth, and elevation angles, etc.) obtained from the off-the-shelf receiver are transmitted to the remote computer through the

Sensors 2021, 21, x FOR PEER REVIEWWi-Fi connection. The routine summarized in Figure7 has the following four steps:8 of 24 raw data reception, satellite selection and satellite-to-repeater distance calculations, cleaning pseudo-range from satellite-to-repeater distances and satellite biases, and ﬁnally, running the least squares navigation (LSNAV) algorithm, which is a least squares solution that finally, running the least squares navigation (LSNAV) algorithm, which is a least squares minimizes the sum of the square of the residual errors [48]. solution that minimizes the sum of the square of the residual errors [48].

FigureFigure 7.7. Flow of the the a algorithm.lgorithm.

In the first part, we remove the biases from the GPS pseudo-range measurement using the models of the troposphere, ionosphere, GPS satellite clocks, GPS satellite movement during signal propagation, and Earth rotation. Navigation data are also collected as part of raw data from the off-the-shelf receiver, which contains GPS time of the week (ITOW) of the navigation epoch. ITOW field indicates the GPS time at which the navigation epoch occurred. Each navigation solution is triggered by the tick of the 1 kHz clock nearest to the desired navigation solution time. This tick is referred to as a navigation epoch. If the navigation solution attempt is successful, one of the results is an accurate measurement of time in the time base of the chosen GNSS system, called GNSS system time. The difference between the calculated GNSS system time and receiver local time is called the clock bias (and the clock drift is the rate at which this bias is changing). Navigation data also contain an Earth-centered Earth-fixed (ECEF) coordinates solution of the off-the-shelf receiver; however, this solution would not be correct in our system, since the satellite signals come from three different repeaters. This result in erroneous the off-the-shelf receiver position. The data of each satellite, received by the off-the-shelf receiver on the up-converting receiver, are recorded. The recorded data include GPS satellite ID numbers, satellite azimuth and elevation angles, carrier-to-noise ratios (CNR) related to each satellite signal, and the GPS time of the week. GPS satellites transmit ephemeris through which the receiver can estimate the position of the satellites in the Earth-centered Earth-fixed (ECEF) coordinates system. In addition to a GPS satellite’s location (current and predicted), ephemeris includes the orbital parameters, clock bias, date, timing, health, and an almanac (a reduced subset of ephemeris of all satellites) exist. The ephemeris data can be collected using an online server to enable hot start in the proposed system. Moreover, the ephemeris data are used to calculate the satellite’s clock bias. Each of the collected pseudo-range is a sum of the following distance terms: satellite- to-repeater distance, the receiver clock bias, satellite clock bias, repeater delay, the indoor distance from the corresponding repeater to the indoor receiver position. The second step of the MATLAB® algorithm is designed for choosing the GPS satellites, which are seen from the repeaters. For each repeater, specific satellites are chosen

Sensors 2021, 21, 4338 8 of 23

In the first part, we remove the biases from the GPS pseudo-range measurement using the models of the troposphere, ionosphere, GPS satellite clocks, GPS satellite movement during signal propagation, and Earth rotation. Navigation data are also collected as part of raw data from the off-the-shelf receiver, which contains GPS time of the week (ITOW) of the navigation epoch. ITOW field indicates the GPS time at which the navigation epoch occurred. Each navigation solution is triggered by the tick of the 1 kHz clock nearest to the desired navigation solution time. This tick is referred to as a navigation epoch. If the navigation solution attempt is successful, one of the results is an accurate measurement of time in the time base of the chosen GNSS system, called GNSS system time. The difference between the calculated GNSS system time and receiver local time is called the clock bias (and the clock drift is the rate at which this bias is changing). Navigation data also contain an Earth-centered Earth-fixed (ECEF) coordinates solution of the off-the-shelf receiver; however, this solution would not be correct in our system, since the satellite signals come from three different repeaters. This result in erroneous the off-the-shelf receiver position. The data of each satellite, received by the off-the-shelf receiver on the up-converting receiver, are recorded. The recorded data include GPS satellite ID numbers, satellite azimuth and elevation angles, carrier-to-noise ratios (CNR) related to each satellite signal, and the GPS time of the week. GPS satellites transmit ephemeris through which the receiver can estimate the position of the satellites in the Earth-centered Earth-fixed (ECEF) coordinates system. In addition to a GPS satellite’s location (current and predicted), ephemeris includes the orbital parameters, clock bias, date, timing, health, and an almanac (a reduced subset of ephemeris of all satellites) exist. The ephemeris data can be collected using an online server to enable hot start in the proposed system. Moreover, the ephemeris data are used to calculate the satellite’s clock bias. Each of the collected pseudo-range is a sum of the following distance terms: satellite- to-repeater distance, the receiver clock bias, satellite clock bias, repeater delay, the indoor distance from the corresponding repeater to the indoor receiver position. The second step of the MATLAB® algorithm is designed for choosing the GPS satellites, which are seen from the repeaters. For each repeater, specific satellites are chosen based on their CNR levels and locations. Each satellite’s position, clock bias, and distance from the corresponding repeater are calculated using the satellite’s ephemeris data. In the third step of the algorithm, the pseudo-range (PR) from each repeater to the selected satellite for the corresponding repeater is cleaned by subtraction operation on the right-hand side of Equation (1). Thus, the indoor distance from the receiver to each repeater is calculated. The calculated indoor distances also include the receiver clock bias. The indoor distances between each repeater and the indoor receiver is calculated using Equation (1).

receiver Sj Sj Sj di + tbias = PRRi − DRi + τi × c + tbias × c (1)

In Equation (1), i is the index of the repeaters (i = 1, 2, 3), j is the index of the satellites th th (j = 1, 2, ... , n), Ri represents the i repeater, Sj represents the j satellite, di represents receiver Sj the indoor distance from repeater Ri to the receiver (with a clock bias of tbias ), DRi stands th for the distance from ith repeater (Ri) to the j satellite (Sj), τi is the propagation delay of Sj the ith repeater, tbias is the satellite clock bias of the jth satellite, and c represents the speed receiver of the light. When more than one satellite is chosen for a repeater, the di + tbias value is calculated by averaging the calculations from each satellite selected for that repeater. Thus, all the distances in Figure2 can be solved with the proposed system. The ﬁnal step of the proposed algorithm is using the LSNAV algorithm to calculate the 2D indoor position for each sample collected. The indoor distances between the receiver and each repeater with receiver clock bias are calculated using Equation (1). Then, these calculated values and the repeater positions are given as the inputs to the LSNAV algorithm. It is important to note that the calculated indoor distances still have a receiver clock bias. receiver Since the receiver clock bias (tbias ) is the same in all indoor distances, it is removed by Sensors 2021, 21, x FOR PEER REVIEW 9 of 24

based on their CNR levels and locations. Each satellite’s position, clock bias, and distance from the corresponding repeater are calculated using the satellite’s ephemeris data. In the third step of the algorithm, the pseudo-range (PR) from each repeater to the selected satellite for the corresponding repeater is cleaned by subtraction operation on the right-hand side of Equation (1). Thus, the indoor distance from the receiver to each repeater is calculated. The calculated indoor distances also include the receiver clock bias. The indoor distances between each repeater and the indoor receiver is calculated using Equation (1).

+ = − ( + × + × ) (1) In Equation (1), is the index of the repeaters ( = 1, 2, 3), is the index of the satel- th th lites ( = 1, 2,…, ), represents the repeater, represents the satellite, represents the indoor distance from repeater to the receiver (with a clock bias of ), th stands for the distance from ith repeater () to the satellite (), is the propagation delay of the th repeater, is the satellite clock bias of the th satellite, and represents the speed of the light. When more than one satellite is chosen for a repeater, the + value is calculated by averaging the calculations from each satellite selected for that repeater. Thus, all the distances in Figure 2 can be solved with the proposed system. The final step of the proposed algorithm is using the LSNAV algorithm to calculate the 2D indoor position for each sample collected. The indoor distances between the re- Sensors 2021, 21, 4338 9 of 23 ceiver and each repeater with receiver clock bias are calculated using Equation (1). Then, these calculated values and the repeater positions are given as the inputs to the LSNAV algorithm. It is important to note that the calculated indoor distances still have a receiver clockthe LSNAV bias. Since algorithm the receiver by subtracting clock bias ( them from) is the each same other, in all and indoor the distances, indoor position it is is removed by the LSNAV algorithm by subtracting them from each other, and the indoor calculated. This algorithm is run for each sample. The obtained results are filtered with a position is calculated. This algorithm is run for each sample. The obtained results are fil- moving average filter, and the indoor position is estimated as an average of these filtered tered with a moving average filter, and the indoor position is estimated as an average of points. More details regarding the estimation are provided under Section4. these filtered points. More details regarding the estimation are provided under Section 4. 3. Experimental Setups for Indoor Positioning 3. Experimental Setups for Indoor Positioning The indoor positioning experiments were performed in the Sabancı University Faculty The indoor positioning experiments were performed in the Sabancı University Fac- of Engineering and Natural Sciences. The floorplan is depicted in Figure8. ulty of Engineering and Natural Sciences. The floorplan is depicted in Figure 8.

Figure 8. The location of repeaters on the floor plan of Sabancı University Faculty of Engineering Sensors 2021, 21, x FOR PEER REVIEWFigure 8. The location of repeaters on the ﬂoor plan of Sabancı University Faculty of10 Engineering of 24 and Natural Natural Sciences Sciences (FENS) (FENS) Building Building 2nd 2nd Floor Floor..

The satellite positions in the sky have been observed as depicted in Figure9, and the directionalThe satellite GPS p antennaositions orientationsin the sky have are been set accordingly.observed as depicted in Figure 9, and the directional GPS antenna orientations are set accordingly.

Figure 9. The satellite locations in the sky during the experiments. Figure 9. The satellite locations in the sky during the experiments. The hardware delays from GPS antennas, ampliﬁers, attenuators, ﬁlters, converters, The hardware delays from GPS antennas, amplifiers, attenuators, filters, converters, bias-tees, and cables are measured and presented as group delays (τi) in Table3. The bias-tees, and cables are measured and presented as group delays () in Table 3. The group delay of each repeater is measured with a 20 GHz oscilloscope. As a result, for each repeater, GPS antenna and cabling have different measured values in the order of nanoseconds. These values are used in Equation (1) to calibrate the system. Moreover, each repeater position is also provided in Table 3. As mentioned earlier, two experiments are presented in this paper. While the repeater positions are fixed in the same location in both experiments, as presented in Table 3, the receiver position is different for each experiment. In the first experiment, the receiver is located at the intersection of 2 corridors, in that it is on the line-of-sight of all 3 repeaters. However, in the second experiment, the receiver is located closer to Repeater 1 (R1), in that it is not on the line-of-sight of Repeater 2 (R2).

Table 3. Repeater configurations.

Position Repeater Number (Ri) Measured Group Delay ( ) (Latitude, Longitude) R1 48.5 ns 40.89072, 29.37955 R2 103 ns * 40.89048, 29.37941 R3 86 ns 40.89044, 29.37966 The latitudes and longitudes are given in degrees. * Repeater 2 has a third LHA-13LN+ amplifier.

In all experiments, there has been no switching, and all three repeaters have been transmitted simultaneously.

3.1. Setup for Experiment 1 Figure 10 shows the first experimental setup and indoor distances between each receiver and repeater. The repeater configurations are kept as provided in Table 3. Although Repeater 2 has a third amplifier, its attenuator is set to 15 dB. Therefore, its gain is 7 dB higher from Repeaters 2 and 3. The outdoor GPS antenna directions are set such that the beam of an antenna does not overlap in azimuth with other GPS antennas in the setup.

Sensors 2021, 21, 4338 10 of 23

group delay of each repeater is measured with a 20 GHz oscilloscope. As a result, for each repeater, GPS antenna and cabling have different measured values in the order of nanoseconds. These values are used in Equation (1) to calibrate the system. Moreover, each repeater position is also provided in Table3. As mentioned earlier, two experiments are presented in this paper. While the repeater positions are ﬁxed in the same location in both experiments, as presented in Table3, the receiver position is different for each experiment. In the ﬁrst experiment, the receiver is located at the intersection of 2 corridors, in that it is on the line-of-sight of all 3 repeaters. However, in the second experiment, the receiver is located closer to Repeater 1 (R1), in that it is not on the line-of-sight of Repeater 2 (R2).

Table 3. Repeater conﬁgurations.

Position Repeater Number (Ri) Measured Group Delay (τ ) i (Latitude, Longitude) R1 48.5 ns 40.89072, 29.37955 R2 103 ns * 40.89048, 29.37941 R3 86 ns 40.89044, 29.37966 The latitudes and longitudes are given in degrees. * Repeater 2 has a third LHA-13LN+ ampliﬁer.

In all experiments, there has been no switching, and all three repeaters have been transmitted simultaneously.

3.1. Setup for Experiment 1 Figure 10 shows the first experimental setup and indoor distances between each receiver and repeater. The repeater configurations are kept as provided in Table3. Although Repeater 2 has a third amplifier, its attenuator is set to 15 dB. Therefore, its gain is 7 dB Sensors 2021, 21, x FOR PEER REVIEW 11 of 24 higher from Repeaters 2 and 3. The outdoor GPS antenna directions are set such that the beam of an antenna does not overlap in azimuth with other GPS antennas in the setup. The directional GPS antennas in Repeaters 1, 2, and 3 are, respectively, towards the geographic Theeast, directional south, and GPS west. antennas In this in experiment, Repeaters 1,the 2, and receiver 3 are, isrespectively, located at thetowards intersection the geo- of two graphicperpendicular east, south, corridors and west. to provideIn this experiment, a line-of-sight the receiver condition. is located at the intersection of two perpendicular corridors to provide a line-of-sight condition.

Figure 10. Distance between the receiver and each repeater, directional GPS antenna azimuth (Az) Figure 10. Distance between the receiver and each repeater, directional GPS antenna azimuth (Az) and elevation (El), and the attenuator value in each repeater in Experiment 1. and elevation (El), and the attenuator value in each repeater in Experiment 1.

The real values of the indoor distances , , and are determined with a physical measurement using a laser pointer for a later comparison with the estimated values. The distance measurements are performed with respect to the known coordinates of the building.

3.2. Setup for Experiment 2 Figure 11 shows the second experimental setup and indoor distances between each receiver and repeater. In this experiment, a non-line-of-sight condition is formed by locating the receiver closer to Repeater 1. Repeater 2 does not directly see the receiver due to the corner at the intersection of two perpendicular corridors. The repeater configurations are kept as provided in Table 3. Repeater 2, which has an additional amplifier, is used with 5 dB attenuation to compensate for the non-line-of-sight condition and the scattering from the corner. Therefore, its gain is 17 dB higher than Repeaters 2 and 3. The outdoor GPS antenna directions are the same as Experiment 1.

Sensors 2021, 21, 4338 11 of 23

The real values of the indoor distances d1, d2, and d3 are determined with a physical measurement using a laser pointer for a later comparison with the estimated values. The distance measurements are performed with respect to the known coordinates of the building.

3.2. Setup for Experiment 2 Figure 11 shows the second experimental setup and indoor distances between each receiver and repeater. In this experiment, a non-line-of-sight condition is formed by locating the receiver closer to Repeater 1. Repeater 2 does not directly see the receiver due to the corner at the intersection of two perpendicular corridors. The repeater conﬁgurations are kept as provided in Table3. Repeater 2, which has an additional ampliﬁer, is used with 5 dB attenuation to compensate for the non-line-of-sight condition and the scattering from Sensors 2021, 21, x FOR PEER REVIEWthe corner. Therefore, its gain is 17 dB higher than Repeaters 2 and 3. 12 of 24

The outdoor GPS antenna directions are the same as Experiment 1.

FigureFigure 11. 11. DistanceDistance between between the the receiver receiver and and each each repeater, repeater, directional directional GPS antenna GPS antenna azimuth azimuth (Az) (Az) andand elevation elevation (El), (El), and and the the attenuator attenuator value value in each in each repeater repeater in Experiment in Experiment 2. 2.

TheThe real real values values of ofindoor indoor dist distancesances , d1,, andd2, and ared3 aredetermined determined with witha physical a physical measurementmeasurement using using a laser a laser pointer pointer for fora later a later comparison comparison with with the estimated the estimated values. values. The The distancedistance measurements measurements are are performed performed with with respect respect to the to well the- well-knownknown coordinates coordinates of the of the building.building.

4.4. Results Results and and Di Discussionscussion InIn this this section, section, the the techniques techniques performed performed for the for indoor the indoor position position estimation estimation with the with the proposedproposed hardware, hardware, algorithm algorithm,, and and methods methods are arepresented presented along along with withthe experimental the experimental datadata gathered gathered in in the the experiments experiments in in a a real indoor environment. Among Among these, these, t thehe esti- estimated matedindoor indoor position, position, the satellitesthe satellites seen seen by by each each directional directional outdooroutdoor GPS GPS antenna antenna during during the theexperiments, experiments, estimated estimated distances distances from from each each repeater repeater toto the receiver with with receiver receiver clock clock bias, bias,the CNRthe CNR of each of each satellite satellite signal signal within within the anglethe angle of view of view of the of correspondingthe corresponding repeater, re- and peater,the 50% and CEP the from50% CEP the estimatedfrom the estimated position areposition graphically are graphically visualized visualized and presented and pre- in this ® sentedsection. in this The section. MATLAB The® MATLABroutine with routine LSNAV with algorithm,LSNAV algorithm, presented presented in Section in Section2, has been 2, has been utilized for the position estimation of each sample collected in the experiments. utilized for the position estimation of each sample collected in the experiments. The satellites, in the angle of view of each repeater in Experiments 1 and 2, are pre- The satellites, in the angle of view of each repeater in Experiments 1 and 2, are sented in Figure 12. It should be noted that for this approach to work properly, it is im- presented in Figure 12. It should be noted that for this approach to work properly, it is portant to know which repeater transfers the signal, coming from a particular satellite, into the building. This is in fact one of the reasons for the decision to utilize directional GPS antennas. The association of the repeater, used for the signal of a particular GPS satellite, is accomplished by using the position, the azimuth, and the elevation of the directional GPS antenna and considering the GPS satellites that are in the targeted region of that particular antenna. If a satellite falls into the targeted region of multiple GPS antennas, we do not use any measurement based on that satellite, since we cannot be sure which repeater transferred the signal of that satellite into the building. In these experiments, therefore, the satellite labeled as G19 in Figure 12 (red point) has been ignored, while the satellites represented with green points have been utilized in position estimation for both experiments.

Sensors 2021, 21, 4338 12 of 23 Sensors 2021, 21, x FOR PEER REVIEW 13 of 24

important to know which repeater transfers the signal, coming from a particular satellite, Additionally, although the directional antenna beam widths are 60 degrees, the experi- into the building. This is in fact one of the reasons for the decision to utilize directional ment results have shown that the GPS antennas can receive signals from a 90-degree angle GPS antennas. in azimuth (Figure 12).

Figure 12. Satellites in each repeater’s angle of view in Experiments (a) 1 and (b) 2. Figure 12. Satellites in each repeater’s angle of view in Experiments (a) 1 and (b) 2. The association of the repeater, used for the signal of a particular GPS satellite, is 4.1. Results of Experiment 1 accomplished by using the position, the azimuth, and the elevation of the directional GPSThe antenna first experiment and considering is performed the GPS in the satellites real indoor that environment, are in the targetedwhich is presented region of that inparticular Figure 10, antenna. to demonstrate If a satellite the performance falls intothe of the targeted proposed region indoor of multiplepositioning GPS system antennas, underwe do line not-of use-sight any conditions. measurement The position based onfor thateach satellite,sample is sinceestimated we cannot with the be LSNAV sure which algorithm using the GPS satellite’s signal, which is within the angle of view of each re- repeater transferred the signal of that satellite into the building. In these experiments, peater. As mentioned previously, G19 is ignored because it is seen by more than one re- therefore, the satellite labeled as G19 in Figure 12 (red point) has been ignored, while the peater. Under these conditions, the CNR of the received GPS signals, from the selected satellites represented with green points have been utilized in position estimation for both satellites for the indoor position estimation, is provided in Figure 13. experiments. Additionally, although the directional antenna beam widths are 60 degrees, the experiment results have shown that the GPS antennas can receive signals from a 90-degree angle in azimuth (Figure 12).

Sensors 2021, 21, 4338 13 of 23

4.1. Results of Experiment 1 The ﬁrst experiment is performed in the real indoor environment, which is presented in Figure 10, to demonstrate the performance of the proposed indoor positioning system under line-of-sight conditions. The position for each sample is estimated with the LSNAV algorithm using the GPS satellite’s signal, which is within the angle of view of each repeater. As mentioned previously, G19 is ignored because it is seen by more than one repeater. Sensors 2021, 21, x FOR PEER REVIEWUnder these conditions, the CNR of the received GPS signals, from the selected14 satellites of 24

for the indoor position estimation, is provided in Figure 13.

FigureFigure 13. 13.CNR CNR of of each each GPS GPS signal signal received received byby thethe repeaters.repeaters.

The sum of the indoor distances ( , , and ), including receiver clock receiver bias The sum of the indoor distances (d , d2, and d3), including receiver clock bias (t ), 1 bias are( calculated ), are in calculated Equation in(1) Equation and subtracted (1) and subtracted from each from to each obtain to obtain the difference the difference terms termsreceiver ( + receiver)−( + ) andreceiver ( + receiver)−( + ) for each sample. (d2 + tbias )−(d1 + tbias ) and (d3 + tbias )−(d1 + tbias ) for each sample. Figure 14a,b showsFigure the 14a indoor,b shows distance the indoor difference distance terms difference in which terms the in receiver which the clock receiver biases clock are removed biases are removed due to the subtraction, while in Figure 14c, it demonstrates the individual due to the subtraction, while in Figure 14c, it demonstrates the individual distances with distances with the receiver clockreceiver bias ( receiver+ , receiver + , + ). the receiver clock bias (d1 + tbias , d2 + tbias , d3 + tbias ). As seen in Figure 14c, the indoor distances with receiver clock bias ( + , + , + ) can go up to thousands of meters. When the receiver clock bias is removed, with subtraction, the indoor differences can be found in the range of tens of meters. The calculated indoor distance difference terms (blue curve in Figure 14a,b) are averaged with a moving average filter of a five-sample window size. The resulting terms (red curve in Figure 14a,b) show a closer result to the measured indoor distance differences (black dashed line in Figure 14a,b), which can be calculated using the measured indoor distances , , and in Figure 10.

Sensors 2021, 21, 4338 14 of 23 Sensors 2021, 21, x FOR PEER REVIEW 15 of 24

(a)

(b)

(c)

FigureFigure 14. 14.The The indoor indoor distance distance differences differences ( a(,ab,b)) and and indoor indoor distances distances with with the the receiver receiver clock clock bias bias term term (c ().c).

Figure 15a depicts the GPS repeater and outdoor GPS antenna positions accordingreceiver to As seen in Figure 14c, the indoor distances with receiver clock bias (d1 + tbias , d2 theirreceiver latitude andreceiver longitude in Experiment 1. These results show that the LSNAV algo- + tbias , d3 + tbias ) can go up to thousands of meters. When the receiver clock bias is removed,rithm has with resulted subtraction, in a solution the indoor for 74 differencesout of 100 collected can be found samples. in the The range reason of why tens ofthe meters.LSNAV The does calculated not result indoor in a solution distance for differenceevery sample terms collected (blue curve is addressed in Figure subsequently 14a,b) are averagedin the CNR with provided a moving in average Figure 13. ﬁlter For of some a ﬁve-sample samples, window due to environmental size. The resulting changes terms or (redmultipath, curve in an Figure abrupt 14a,b) change show in a CNR closer occurs. result to For the these measured samples, indoor that distance do not differences have good (blackCNR from dashed all linethree in repeaters, Figure 14 LSNAa,b), whichV does can not be result calculated in a solution. using the measured indoor distancesMoreover,d1, d2, anda movingd3 in Figureaverage 10 filter. with a five-sample window size has been applied overFigure the 74 15 resultsa depicts from the LSNAV, GPS repeater which and then outdoor give 69 GPS location antenna estimations. positions accordingThe resulting to their69 locations latitude andare depicted longitude in in Figure Experiment 15a,b with 1. These red resultsdots. Figure show that15a presents the LSNAV the algorithm results on hasa latitude–longitude resulted in a solution graph, for 74whereas out of 100Figure collected 15b presents samples. the The results reason on why a local the reference LSNAV doesframe not where result the in acenter solution is the for real every position sample of collected the receiver, is addressed which was subsequently previously in deter- the CNRmined provided by physical in Figure measurements 13. For some in samples, the expe duerimental to environmental setup with respect changes to or the multipath, building whose coordinates and plan are well known. We calculated the radius of 50% CEP as 3.3

Sensors 2021, 21, x FOR PEER REVIEW 16 of 24

Sensors 2021, 21, 4338 m and plotted it in Figure 15a. Moreover, the indoor position is estimated as the mean15 of of 23 these 69 points obtained with moving average filtering. The real position of the receiver is represented with a black square, while the estimated position is presented with a blue square in Figure 15a,b. The estimated location is obtained by averaging 69 points, which an abruptare the change output in of CNR the moving occurs. average For these filter, samples, and is that54 cm do away not havefrom goodthe real CNR position, from allas threecan repeaters, be seen LSNAVin Figure does 15b. not result in a solution.

Figure 15. The results of the Experiment 1: (a) repeater positions, outdoor GPS antenna positions, collected samples (red Figure 15. The results of the Experiment 1: (a) repeater positions, outdoor GPS antenna positions, collected samples dots), real position of the receiver (black square), estimated position (blue square), and CEP (blue circle) on latitude– (red dots),longitude real position graph; of(b) the collected receiver samples (black (red square), dots), estimated real position position of the (blue receiver square), (black and square), CEP (blueestimated circle) position on latitude– (blue longitudesquare) graph; on local (b) collected reference samplesframe whose (red center dots), is real the positionreal position of the of the receiver receiver. (black square), estimated position (blue square) on local reference frame whose center is the real position of the receiver.

Moreover, a moving average ﬁlter with a ﬁve-sample window size has been applied

over the 74 results from LSNAV, which then give 69 location estimations. The resulting 69 locations are depicted in Figure 15a,b with red dots. Figure 15a presents the results on a latitude–longitude graph, whereas Figure 15b presents the results on a local reference frame where the center is the real position of the receiver, which was previously determined Sensors 2021, 21, 4338 16 of 23

by physical measurements in the experimental setup with respect to the building whose coordinates and plan are well known. We calculated the radius of 50% CEP as 3.3 m and plotted it in Figure 15a. Moreover, the indoor position is estimated as the mean of these 69 points obtained with moving average ﬁltering. The real position of the receiver is represented with a black square, while the estimated position is presented with a blue Sensors 2021, 21, x FOR PEER REVIEW 17 of 24 square in Figure 15a,b. The estimated location is obtained by averaging 69 points, which are the output of the moving average ﬁlter, and is 54 cm away from the real position, as can be seen in Figure 15b. BecauseBecause the estimated the estimated position position from 69 from points 69 pointsis 54 cm is away 54 cm from away the from real theposition real position (Figure(Figure 15), 15 a ),sub a sub-meter-meter accuracy accuracy has hasbeen been achieved achieved for line for- line-of-sightof-sight conditions. conditions. In addi- In addition tionto to the the 50% 50% CEPCEP withwith a radius radius of of 3.3 3.3 m m from from the the estimated estimated position, position, we we also also calculated calculated each eachof of the the 69 69 points points distances distances from from the the real real position position andand presentedpresented itit inin FigureFigure 1616.. TheThe average averageerror error of the of individualthe individual samples, samples, from from the the real real position, position, is is 3.37 3.37 m. m.

FigureFigure 16. Distance 16. Distance from fromthe real the position real position for each for of each 69 points. of 69 points.

4.2.4.2. Results Results of Experiment of Experiment 2 2 ExperimentExperiment 2 setup 2 setup in Figure in Figure 11 is11 set is setto demonstrate to demonstrate the theperformance performance of the of thepro- proposed posedindoor indoor positioning positioning system system under under non-line-of-sight non-line-of-sight conditions conditions inin aa real indoor envi- environment. ronment.The CNR The ofCNR GPS of signals GPS signals received received from from the selected the selected GPS satellitesGPS satellites are providedare provided in Figure 17. in FigureSimilar 17. to Experiment 1, the average position is estimated with the LSNAV algorithm. InSimilar this experiment to Experiment also, 1, G19 the is average ignored, position as it is is within estimated the angle with ofthe view LSNAV of more algo- than one rithm.repeater. In this Additionally,experiment also, although G19 is ignored G15 is seen, as it by is Repeaterwithin the 2, angle its CNR of view is much of more lower than thanG24; onetherefore, repeater. itAdditionally, is also ignored although in the G15 calculations. is seen by Repeater 2, its CNR is much lower than G24; therefore, it is also ignored in the calculations.

Sensors 2021, 21, 4338 17 of 23 Sensors 2021, 21, x FOR PEER REVIEW 18 of 24

FigureFigure 17. 17.CNR CNR of of each each GPS GPS signal signal received received byby thethe repeaters.repeaters.

TheThe indoor indoor distances distances (d(,, d,, andand d),), withwith receiver receiver clock clock bias bias ( (treceiver) )and and their their 1 2 3 bias difference at each sample (receiver + )−(receiver + ) andreceiver ( + receiver)−( + difference at each sample (d2 + t )−(d1 + t ) and (d3 + t )−(d1 + t ) are ) are plotted and demonstratedbias in Figurebias 18. The resultingbias terms (red biascurve in plotted and demonstrated in Figure 18. The resulting terms (red curve in Figure 18a,b) Figure 18a,b) show a closer result to the measured indoor distance differences (black show a closer result to the measured indoor distance differences (black dashed line in dashed line in Figure 18a,b), which can be calculated using the measured indoor distances Figure 18a,b), which can be calculated using the measured indoor distances d , d , and d , , and in Figure 11. 1 2 3 in Figure 11.

Sensors 2021, 21, 4338 18 of 23 Sensors 2021, 21, x FOR PEER REVIEW 19 of 24

(a)

(b)

(c)

FigureFigure 18. 18.The The indoorindoor distancedistance differences differences ( a(a,b,b)) and and indoor indoor distances distances with with the the receiver receiver clock clock bias bias term term ( c().c).

FigureFigure 19 19aa depicts depicts the the GPS GPS repeater repeater and and outdoor outdoor GPS GPS antenna antenna positions, positions, according according to theirto their latitude latitude and and longitude, longitude, in in Experiment Experiment 2. 2. In In this this experiment, experiment, the the LSNAV LSNAV algorithm algorithm hashas resultedresulted in solutions for for 35 35 out out of of 100 100 collected collected samples. samples. Then, Then, a moving a moving average average filter ﬁlterwith witha five-sample a ﬁve-sample window window size sizehas been has been applied applied over over these these 35 converged 35 converged solutions, solutions, and and30 location 30 location estimations estimations have have been been obtained obtained (Figure (Figure 19a,b 19a,b with with red dots).dots). FigureFigure 1919aa demonstratesdemonstrates thethe resultsresults onon aa latitude–longitude latitude–longitude graph,graph, whereaswhereas FigureFigure 19 19bb presents presents the the resultsresults onon aa locallocal referencereference frameframe inin whichwhich thethe centercenter isis thethe realreal positionposition ofof thethe receiver,receiver, whichwhich waswas previouslypreviously determineddetermined byby distancedistance measurementmeasurement inin thethe experimental experimental setupsetup withwith respectrespect toto thethe known known coordinates coordinates of of the the building building structure. structure. TheThe indoorindoor positionposition isis estimatedestimated asas thethe meanmean ofof thesethese 3030 points.points. TheThe realreal positionposition ofof thethe receiverreceiver (black(black square)square) andand estimatedestimated positionposition (blue(blue square)square) have have been been plotted plotted in in Figure Figure 19 19a,b.a,b. ItIt is is seen seen that that the the estimatedestimated positionposition is is 98 98 cm cm away away from from the the real real position position in in Figure Figure 19 19b.b. The 50% CEP is calculated to have a radius of 4 m and drawn also in Figure 19a.

Sensors 2021, 21, 4338 19 of 23 Sensors 2021, 21, x FOR PEER REVIEW 20 of 24

FigureFigure 19. The 19. resultsThe results of the of the Experiment Experiment 2: 2: (a ()a repeater) repeater positions,positions, o outdoorutdoor GPS GPS antenna antenna positions, positions, collected collected samples samples (red dots), real position of the receiver (black square), estimated position (blue square), and CEP (blue circle) on latitude– (red dots), real position of the receiver (black square), estimated position (blue square), and CEP (blue circle) on latitude– longitude graph; (b) collected samples (red dots), real position of the receiver (black square), estimated position (blue longitudesquare) graph; on local (b) collectedreference samplesframe whose (red center dots), is real the real position position of theof the receiver receiver. (black square), estimated position (blue square) on local reference frame whose center is the real position of the receiver. Moreover, the distance error from the real receiver position for each of the 30 points hasThe been 50% depicted CEP is calculated in Figure 20. to The have average a radius error of 4of m the and individual drawn also samples in Figure from 19thea. real positionMoreover, is 4.17 the m distance. error from the real receiver position for each of the 30 points has been depicted in Figure 20. The average error of the individual samples from the real position is 4.17 m.

SensorsSensors 20212021, 21, ,21 x ,FOR 4338 PEER REVIEW 21 of 24 20 of 23

FigureFigure 20. Deviation 20. Deviation from fromthe real the position real position for each for of each 30 points of 30. points.

UnderUnder non- non-line-of-sightline-of-sight conditions conditions in the inexperiments, the experiments, while estimated while estimated position positionerror error increases,increases, it is itstill is possible still possible to achieve to achieve sub-meter sub-meter accuracy. accuracy. To quantify To quantify the increased the increased error error amount,amount, one onecan compare can compare the radius the radius of the of50% the CEP 50% circles CEP circlesin Figure insFigures 15a and 15 19a.a and The 19 a. The radiusradius of the of 50% the CEP 50% is CEPhigher is in higher Experiment in Experiment 2. Additionally 2. Additionally,, when CNR values when in CNR Figure values in 17 areFigure compared 17 are with compared those in withFigure those 13, we in can Figure conclude 13, wethat can the concludenon-line-of that-sight the condi- non-line-of- tionsight reduces condition the CNR. reduces In the proposed the CNR. system, In the proposedvariable attenuators system, variable are used attenuators to compensate are used to for the non-line-of-sight conditions. To adjust, we can easily set the attenuation level, the compensate for the non-line-of-sight conditions. To adjust, we can easily set the attenuation retransmitted GPS power levels, and adjust the CNR levels so that the receivers can receive level, the retransmitted GPS power levels, and adjust the CNR levels so that the receivers almost equal signal levels from each direction and each satellite. That is the reason why the can receive almost equal signal levels from each direction and each satellite. That is the attenuation level of Repeater 2 is set to a lower level in Experiment 2. reason why the attenuation level of Repeater 2 is set to a lower level in Experiment 2. 5. Conclusions 5. Conclusions The implemented 433 MHz down-converting repeaters with variable attenuators and The implemented 433 MHz down-converting repeaters with variable attenuators and 433 MHz up-converting receiver with variable attenuator have been used in 2D position- 433 MHz up-converting receiver with variable attenuator have been used in 2D positioning ing successfully for the first time. It is also shown that, with the help of cascaded LNA and variablesuccessfully attenuator, for the the proposed ﬁrst time. system It is also for 2D shown positioning that, with has thegained help immunity of cascaded to the LNA and nonvariable-line-of-sight attenuator, conditions. the proposedThe power systemlevel of forthe 2Dsignals positioning that are transmitted has gained to immunity the re- to the ceivernon-line-of-sight can be adjusted conditions.by varying the The attenuation power level so that of thea receiver signals can that pick are up transmitted different to the satellitereceiver signals can comingbe adjusted from by different varying indoor the attenuation paths almost so that equally. a receiver The canGPS pick signals, up different whichsatellite are down signals-converted coming and from retransmitted different indoor by repeaters, paths almost are sent equally. to the indoor The GPS environ- signals, which mentare and down-converted picked-up by an and up retransmitted-converting receiver. by repeaters, After up are-conversion, sent to the indoorthe raw environment data are and collectedpicked-up by an by off an-the up-converting-shelf receiver receiver.and then Aftertransmitted up-conversion, over Wi-Fi the to a raw remote data com- are collected puterby for an processing. off-the-shelf The receiver estimated and positions then transmitted are found overto be Wi-Fionly 54 to and a remote 98 cm away computer for fromprocessing. the actual The receiver estimated position, positions for line- areof-sight found and to non be only-line-of 54-sight and 98 cases, cm respec-away from the tively.actual Therefore, receiver expe position,riment for results line-of-sight show that and sub non-line-of-sight-meter accuracy cases,can be respectively.achieved with Therefore, the experimenttransmission results of GPS show signals that in sub-meterthe 433 MHz accuracy ISM band can in be an achieved indoor environment. with the transmission of GPS signals in the 433 MHz ISM band in an indoor environment. Author Contributions: Conceptualization, İ.T. and H.Y.; funding acquisition, İ.T. and H.Y.; software,Author F.A.G., Contributions: A.U., and A.M.A.N.;Conceptualization, investigation, I.T.A.U.˙ and and H.Y.; F.A.G.; funding testing, acquisition, A.U., F.A.GI.T.˙ ., A.M.A and H.Y.;.N., software, H.YF.A.G.,., and İ.T.; A.U., writing and A.M.A.N.;—original draft investigation, preparation, A.U. A.U. and and F.A.G.; F.A.G.; testing, writing A.U.,—review F.A.G., and A.M.A.N., editing, H.Y., and A.U., F.A.G., H.Y., and İ.T.; project administration, İ.T. and H.Y. All authors have read and agreed I.T.;˙ writing—original draft preparation, A.U. and F.A.G.; writing—review and editing, A.U., F.A.G., to the published version of the manuscript. H.Y., and I.T.;˙ project administration, I.T.˙ and H.Y. All authors have read and agreed to the published Funding:version This of research the manuscript. has been supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK), Grant No: 116E752. Funding: This research has been supported by the Scientiﬁc and Technological Research Council of Turkey (TÜBITAK),˙ Grant No: 116E752.

Institutional Review Board Statement: Not Applicable. Sensors 2021, 21, 4338 21 of 23

Informed Consent Statement: Not Applicable. Data Availability Statement: Not Applicable. Conﬂicts of Interest: The authors declare no conﬂict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References 1. MarketsandMarkets. Indoor Location Market by Component (Hardware, Solutions, and Services), Deployment Mode, Organiza- tion Size, Technology, Application, Vertical (Retail, Transportation and Logistics, Entertainment), and Region—Global Forecast to 2025 (Report Code: TC 2878). May 2020. Available online: https://www.marketsandmarkets.com/Market-Reports/indoor- location-market-989.html (accessed on 20 May 2021). 2. Vasisht, D.; Kumar, S.; Katabi, D. Decimeter-Level Localization with A Single Wifi Access Point. In Proceedings of the 13th USENIX Symposium on Networked Systems Design and Implementation (NSDI’16), Santa Clara, CA, USA, 16–18 March 2016; pp. 165–178. 3. Cypriani, M.; Lassabe, F.; Canalda, P.; Spies, F. Open Wireless Positioning System: A Wi-Fi-Based Indoor Positioning System. In Proceedings of the 2009 IEEE 70th Vehicular Technology Conference Fall (VTC 2009-Fall), Anchorage, AK, USA, 20–23 September 2009; pp. 1–5. 4. Zou, H.; Jiang, H.; Lu, X.; Xie, L. An Online Sequential Extreme Learning Machine Approach to Wifi Based Indoor Positioning. In Proceedings of the 2014 IEEE World Forum on Internet of Things (WF-IoT), Seoul, Korea, 6 March 2014; pp. 111–116. 5. Ciurana, M.; Cugno, S. WLAN Indoor Positioning Based on TOA with Two Reference Points. In Proceedings of the 4th Workshop on Positioning, Navigation and Communication 2007 (WPNC’07), Hannover, Germany, 22–22 March 2007; pp. 23–28. 6. Hoang, M.K.; Haeb-Umbach, R. Parameter Estimation and Classification of Censored Gaussian Data with Application to Wifi Indoor Positioning. In Proceedings of the 38th International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2013), Vancouver, BC, Canada, 26–31 May 2013; pp. 3721–3725. 7. Faheem, F. Ibeacon Based Proximity and Indoor Localization System. Master’s Thesis, Purdue University, West Lafayette, IN, USA, 2016. 8. Faragher, R.; Harle, R. Location Fingerprinting with Bluetooth Low Energy Beacons. IEEE J. Sel. Areas Commun. 2015, 33, 2418–2428. [CrossRef] 9. Madhavapeddy, A.; Tse, A. A Study of Bluetooth Propagation Using Accurate Indoor Location Mapping. In Proceedings of the UbiComp 2005: Ubiquitous Computing, Tokyo, Japan, 11–14 September 2005; pp. 105–122. 10. Zafari, F.; Papapanagiotou, I.; Christidis, K. Microlocation for Internet of Things Equipped Smart Buildings. IEEE Internet Things J. 2016, 3, 96–112. [CrossRef] 11. Castillo-Cara, M.; Lovón-Melgarejo, J.; Bravo-Rocca, G.; Orozco-Barbosa, L.; García-Varea, I. An Analysis of Multiple Criteria and Setups for Bluetooth Smartphone-Based Indoor Localization Mechanism. J. Sens. 2017, 2017.[CrossRef] 12. Aykac, M.; Ergun, E.; Aldin, N.B. Zigbee-Based Indoor Localization System with The Personal Dynamic Positioning Method and Modified Particle Filter Estimation. Analog. Integr. Circuits Signal Process. 2017, 92, 263–279. [CrossRef] 13. Ni, L.M.; Liu, Y.; Lau, Y.C.; Patil, A.P. LANDMARC: Indoor Location Sensing Using Active RFID. Wirel. Netw. 2004, 10, 701–710. [CrossRef] 14. Kuo, Y.-S.; Pannuto, P.; Hsiao, K.-J.; Dutta, P. Luxapose: Indoor Positioning with Mobile Phones and Visible Light. In Proceedings of the 20th Annual International Conference on Mobile Computing and Networking, Maui, HI, USA, 7–11 September 2014; pp. 447–458. 15. Liu, K.; Liu, X.; Li, X. Guoguo: Enabling Fine-Grained Smartphone Localization Via Acoustic Anchors. IEEE Trans. Mob. Comput. 2016, 15, 1144–1156. [CrossRef] 16. Huang, W.; Xiong, Y.; Li, X.Y.; Lin, H.; Mao, X.; Yang, P.; Liu, Y.; Wang, X. Swadloon: Direction Finding and Indoor Localization Using Acoustic Signal by Shaking Smartphones. IEEE Trans. Mob. Comput. 2015, 14, 2145–2157. [CrossRef] 17. Priyantha, N.B. The Cricket Indoor Location System; Massachusetts Institute of Technology: Cambridge, MA, USA, 2005. 18. Hazas, M.; Hopper, A. Broadband Ultrasonic Location Systems for Improved Indoor Positioning. IEEE Trans. Mob. Comput. 2006, 5, 536–547. [CrossRef] 19. Peterson, B.; Bruckner, D.; Heye, S. Measuring GPS Signals Indoors. In Proceedings of the 10th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1997), Kansas City, MO, USA, 16–19 September 1997; pp. 615–624. 20. Moeglein, M.; Krasner, N. An Introduction to SnapTrack Server-Aided GPS Technology. In Proceedings of the 11th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1998), Nashville, TN, USA, 15–18 September 1998; pp. 333–342. 21. Garin, L.J.; Chansarkar, M.; Miocinovic, S.; Norman, C.; Hilgenberg, D. Wireless Assisted GPS-SiRF Architecture and Field Test Results. In Proceedings of the 12th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS 1999), Nashville, TN, USA, 14–17 September 1999; pp. 489–498. 22. Andrianarison, M. New Methods and Architectures for High Sensitivity Hybrid GNSS Receivers in Challenging Environments. Ph.D. Thesis, University of Toulouse, Toulouse, France, 2018. Sensors 2021, 21, 4338 22 of 23

23. Monnerat, M.; Couty, R.; Vincent, N.; Huez, O.; Chatre, E. The Assisted GNSS, Technology and Applications. In Proceedings of the 17th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2004), Long Beach, CA, USA, 21–24 September 2004. 24. Klein, D.; Parkinson, B.W. The Use of Pseudo-Satellites for Improving GPS Performance. J. Institude Navig. 1984, 31, 303–315. [CrossRef] 25. Rapinski, J.; Cellmer, S.; Rzepecka, Z. Modiﬁed GPS/Pseudolite Navigation Message. J. Navig. 2012, 65, 711–716. [CrossRef] 26. Rizos, C.; Roberts, G.; Barnes, J.; Gambale, N. Experimental Results of Locata: A High Accuracy Indoor Positioning System. In Proceedings of the 2010 International Conference on Indoor Positioning and Indoor Navigation (IPIN), Zurich, Switzerland, 15–17 September 2010. [CrossRef] 27. Kee, C.; Jun, H.; Yun, D. Indoor Navigation System Using Asynchronous Pseudolites. J. Navig. 2003, 56, 443–455. [CrossRef] 28. Xingli, G.; Baoguo, Y.; Xianpeng, W.; Yongqin, Y.; Ruicai, J.; Heng, Z.; Chuanzhen, S.; Lu, H.; Boyuan, W. A New Array Pseudolites Technology for High Precision Indoor Positioning. IEEE Access 2019, 7, 153269–153277. [CrossRef] 29. Petrovski, I.; Okano, K.; Kawaguchi, S.; Torimoto, H.; Suzuki, K.; Toda, M.; Akita, J. Indoor Code and Carrier Phase Positioning with Pseudolites and Multiple GPS Repeaters. In Proceedings of the 16th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GPS/GNSS 2003), Portland, OR, USA, 9–12 September 2003; pp. 1135–1143. 30. Xu, R.; Chen, W.; Xu, Y.; Ji, S. A New Indoor Positioning System Architecture Using Gps Signals. Sensors 2015, 15, 10074–10087. [CrossRef] 31. Ma, C.; Yang, J.; Chen, J.; Tang, Y. Indoor and Outdoor Positioning System Based on Navigation Signal Simulator and Pseudolites. Adv. Space Res. 2018, 62, 2509–2517. [CrossRef] 32. Ozsoy, K.; Bozkurt, A.; Tekin, I. Indoor Positioning Based on Global Positioning System Signals. Microw. Opt. Technol. Lett. 2013, 55, 1091–1097. [CrossRef] 33. Fluerasu, A.; Jardak, N.; Vervisch-picois, A.; Samama, N. GNSS Repeater Based Approach for Indoor Positioning: Current Status. In Proceedings of the European Navigational Conference ENC-GNSS 2009, Naples, Italy, 3–6 May 2009. 34. Vervisch-Picois, A.; Samama, N. First Experimental Performances of The Repealite Based Indoor Positioning System. In Proceedings of the 2012 International Symposium on Wireless Communication Systems, Paris, France, 28–31 August 2012; pp. 636–640. 35. Selmi, I.; Samama, N.; Vervisch-Picois, A. A New Approach for Decimeter Accurate GNSS Indoor Positioning Using Carrier Phase Measurements. In Proceedings of the 2013 International Conference on Indoor Positioning and Indoor Navigation, Montbeliard, France, 28–31 October 2013. 36. Morosi, S.; Martinelli, A.; Del´zRe,E. Peer-to-peer cooperation for GPS positioning. Int. J. Satell. Commun. Netw. 2017, 35, 323–341. [CrossRef] 37. Mautz, R. Overview of Current Indoor Positioning Systems. Geod. Cartogr. 2009, 35, 18–22. [CrossRef] 38. Electronic Communications Committee (ECC) within the European Conference of Postal and Telecommunications Administration (CEPT). Technical and Operational Provisions Required for the Use of GNSS Repeaters; ECC REPORT 129; ECC: Dublin, Ireland, 2009. Available online: https://docdb.cept.org/download/f9323e3e-0577/ECCREP129.PDF (accessed on 20 May 2021). 39. Electronic Communications Committee (ECC) within the European Conference of Postal and Telecommunications Administration (CEPT). Regulatory Framework for Global Navigation Satellite System (GNSS) Repeaters. ECC REPORT 145. Available online: https://docdb.cept.org/download/582 (accessed on 20 May 2021). 40. European Telecommunications Standard Institute (ETSI). Electromagnetic Compatibility and Radio Spectrum Matters (ERM); Short Range Devices; Global Navigation Satellite Systems (GNSS) Repeaters; Harmonized EN Covering the Essential Require- ments of Article 3.2 of the R&TTE Directive (ETSI EN 302 645). Available online: https://www.etsi.org/deliver/etsi_en/302600 _302699/302645/01.01.01_60/en_302645v010101p.pdf (accessed on 20 May 2021). 41. U.S. Department of Commerce National Telecommunications and Information Administration (NTIA). Manual of Regulations and Procedures for Federal Radio Frequency Management; Redbook. Available online: https://www.ntia.doc.gov/page/2011/manual- regulations-and-procedures-federal-radio-frequency-management-redbook (accessed on 20 May 2021). 42. Tekin, I.; Bozkurt, A.; Ozsoy, K. Indoor Positioning System Based on GPS Signals and Pseudolites with Outdoor Directional Antennas. U.S. Patent US20120286992A1, 15 November 2012. 43. Orhan, O.; Tekin, I.; Bozkurt, A. Down-converter for GPS applications. In Proceedings of the 12th Mediterranean Microwave Symposium, Istanbul, Turkey, 3–4 September 2012. 44. Orhan, O. An Indoor Positioning System Based on Global Positioning System. Master’s Thesis, Sabanci University, Istanbul, Turkey, 2013. 45. Uzun, A.; Ghani, F.A.; Yenigun, H.; Tekin, I. A Novel GNSS Repeater Architecture for Indoor Positioning Systems in ISM Band. In Proceedings of the 2020 IEEE International Symposium on Antennas and Propagation and North American Radio Science Meeting, Montreal, QC, Canada, 5–10 July 2020; pp. 1631–1632. 46. Uzun, A. A Novel Indoor Positioning System Based on GPS Repeaters in 433 MHZ ISM Band. Master’s Thesis, Sabanci University, Istanbul, Turkey, 2020. Sensors 2021, 21, 4338 23 of 23

47. Tekin, I.; Yenigun, H.; Uzun, A.; Ghani, F.A. A GNSS Repeater Architecture and Location Finding Method for Indoor Positioning Systems Using Lower Frequencies than GNSS Signals. No. PCT/TR2020/050202, 13 March 2020. 48. Constell, Inc. Constellation Toolbox User’s Guide, Version 7.00. Available online: https://docplayer.net/64176453-Constellation- toolbox.html (accessed on 20 May 2021).