A-GNSS Performance Test in Various Urban Environments by Using a Commercial Low Cost GNSS Receiver and Service

JPNT 7(4), 205-215 (2018) Journal of Positioning, https://doi.org/10.11003/JPNT.2018.7.4.205 JPNT Navigation, and Timing

Kahee Han1, Jung-Hoon Lee1, Ji-Ung Im2, Jong-Hoon Won1†

1Department of Electrical Engineering, Inha University, Incheon 22212, Korea 2Department of Future Vehicle Engineering, Inha University, Incheon 22212, Korea

ABSTRACT

The recent emergence of new Global Navigation Satellite Systems (GNSS) has resulted in a gradual improvement in the performance of positioning services. This paper verifies the degree of improvement in positioning performance of Assisted- GNSS (A-GNSS) receivers using assistance information compared to standalone-GNSS receivers that do not use assistance information in various urban environments in Korea. For this purpose, field tests are performed in various urban and indoor environments in Korea. The assistance information is provided by u-blox's AssistNow Online and low-cost commercial receivers are used for mobile station receivers. Through experiments, the Time to First Fix (TTFF), acquisition sensitivity, and position accuracy performance improvement are analyzed. The results of the experiments show that using assistance data improved the performance in all experiment locations, and, in particular, a significant performance improvement in terms of TTFF.

Keywords: navigation, positioning, A-GPS/A-GNSS, navigation performance

1. INTRODUCTION guarantee the interoperability of positioning systems that provide the location information of rescue requesters in case Recently, various Location Based Services (LBS) have of crime, fire, and disaster, as well as to improve positioning emerged due to the price drop of the Global Navigation performance (TTA 2015, 2016a,b, 2017a,b). Satellite System (GNSS) receiver chips and the widespread One of the typical emergency rescue services using of smartphones. LBS refers to all services that utilize the location information is the EU emergency call (e-call). e-Call user's location information and is available in various is an In-Vehicle System (IVS) that requests for rescue by fields such as commerce, navigation, and security. In automatically or manually sending the accident location particular, developed countries are building social safety and time information to an emergency rescue organization networks based on location information, and the Federal in the event of a serious car accident, which is required to be Communications Commission (FCC) and the European installed in all vehicles sold in Europe as of March 2018 (The Emergency Number Association of the European Union (EU) European Parliament and the Council of the European Union are establishing positioning performance requirements to 2015). Currently, the e-Call system in Europe is designed ensure the accuracy of the location information of rescue to send a Minimum Set of Data (MSD) from the e-Call IVS requesters to be provided to emergency rescue agencies to the closest Public Safety Answering Point via a cellular (FCC 2015, ECC 2016). In 2015, Korea also established network (e.g. GSM, UMTS). The MSD is standardized data standards for emergency rescue positioning systems to that includes information such as the number of passengers, the accident time, and the location of the accident vehicle. Received Sep 20, 2018 Revised Nov 06, 2018 Accepted Nov 13, 2018 The Next Generation e-Call (NG e-Call), which is recently †Corresponding Author being discussed, plans to phase out the existing GSM and E-mail: [email protected] UMTS networks and replace them with 4G LTE and 5G Tel: +82-32-860-7406 Fax: +82-32-863-5822 infrastructure. Fig. 1 shows the principle of data transfer

Fig. 1. Legacy e-Call (left) and NG e-Call (right) data transfer principle (Rohde & Schwarz 2018a,b). in the traditional e-Call and NG e-Call systems (Rohde & papers and reports (Karunanayake et al. 2007, Singh 2007, Schwarz 2018a,b). López et al. 2010), these results are based on considering the The location information of the accident vehicle in the GNSS signal reception environments in foreign countries. MSD is key data for a quick rescue. The e-Call IVS uses Recently, Korea is also proposing policies and implementing positioning technology using satellite navigation systems standardization to develop and commercialize ICT-based such as Global Positioning System (GPS) to generate emergency rescue system technologies and aims to build a location information of rescue requesters. The location Korean e-Call system by June 2019 (Korea Transportation information obtained through satellite navigation systems Safety Authority 2016). Therefore, in order to use A-GNSS can be used globally and has the advantage of high receivers, tests on the positioning performance of A-GNSS positioning accuracy. In addition, multi-GNSS services are receivers considering the multi-GNSS satellite arrangement now available with the operation of new GNSS (or Regional status in the sky of Korea is required to verify the Navigation Satellite System, RNSS) such as EU’s Galileo, improvement of positioning performance in Korea. China’s BeiDou navigation satellite system (BDS), Japan’s This paper is structured as follows. Chapter 2 briefly Quasi-Zenith Satellite System (QZSS), and as a result, the introduces the concept and features of A-GNSS technique, positioning performance has been significantly improved and Chapter 3 analyzes the signal processing performance compared to standalone-GPS. On the other hand, GNSS- improvement according to the application of A-GNSS based positioning requires receivers to observe a certain techniques. Chapter 4 describes the field tests performed level of Line of Sight (LOS) signals. Therefore, the positioning in urban and indoor environments, and the test results are results of standalone-GNSS receivers in urban environments, analyzed from the perspective of improving the performance where the LOS is not sufficiently secured, such as an area of the Geometric Dilution of Precision (GDOP) according to concentrated with high-rise buildings, can have low accuracy the signal acquisition sensitivity and satellite arrangement as and in addition, it takes a long time to obtain positioning well as Time to First Fix (TTFF) according to the provision of results. assistance data. Lastly, Chapter 5 draws the conclusion. However, Assisted-GNSS (A-GNSS) receivers which receive assistance information from the outside via wireless networks can compensate the disadvantages of standalone-GNSS 2. ASSISTED-GNSS TECHNIQUE receivers by using the assistance information provided for positioning (van Digglen 2009). The assistance information The A-GNSS technique was developed due to the need provided externally helps GNSS receivers to acquire signals to reduce the time to obtain navigation solutions and to and derive navigation solutions, and is roughly classified increase the sensitivity of GNSS receivers. The weakness into acquisition assistance information, sensitivity assistance of standalone-GNSS positioning is that it takes a long time information, and navigation assistance information. Although to obtain satellite orbital information and satellite clock various studies on the improved positioning performance of correction information included in received signals. For A-GNSS receivers have been reported in various academic example, when using GPS L1 C/A signals having a NAV https://doi.org/10.11003/JPNT.2018.7.4.205 Kahee Han et al. A-GNSS Performance Test in Various Urban Environments 207

Fig. 2. Overview of generating and transmitting assistance data in the network.

Table 1. Classification of assistance data according to purpose (Kaplan & Hegarty 2006). Forms Purpose Assistance data Acquisition assistance to reduce the GPS receiver’s A list of visible satellites, Predicted GPS satellite Dopplers and Doppler rates, time to generate a fix-time to Azimuth and elevation angles for the visible satellites, Local oscillator offset first fix (TTFF) information, Approximate mobile location, GPS satellite ephemeris information, GPS almanac, Satellite clock correction terms, Approximate GPS time, Precise GPS time, Predicted code phases, Predicted code phase search window, Navigation data bit timing information (bit number, fractional bit), Navigation data bits Sensitivity assistance to help the GPS receiver lower Navigation data bit timing information (bit number, fractional bit), Navigation data its acquisition thresholds bits Navigation assistance to improve the accuracy or DGPS correction data, Approximate altitude of the mobile, Approximate mobile integrity of the position solution location, Real-time satellite integrity information, Fine GPS timing information, generated by the GPS receiver Satellite clock correction coefficients, GPS satellite ephemeris information

message structure, a minimum of 30 seconds is required to The assistance information can be largely classified extract, through demodulation, orbit information and clock into acquisition, sensitivity, and navigation assistance correction components required for positioning. In addition, information depending on the purpose. The acquisition if received signals are not strong enough, acquisition process assistance information is provided to reduce the TTFF at may take a long time. Meanwhile, A-GNSS receivers can the receiver and typically includes a list of visible satellites, estimate the location of the receiver in a shorter time by Doppler/code delay estimates, and ephemeris information. receiving the time information, approximate user location, The sensitivity assistance information is provided so that approximate Doppler frequency, and the ephemeris and the receiver has a low acquisition threshold, and typically almanac of visible satellite from assistant servers. The includes information about navigation data bits. The information obtained from GNSS signals, such as visible navigation assistance information is provided to improve the satellite list, ephemeris, almanac, and satellite health status accuracy and integrity of the navigation solution generated is provided to the server by permanently operating GNSS by the receiver, and includes differential GNSS (DGNSS) reference stations. In case of providing time information correction data, approximate receiver location, and real- to the assistant server from a separate source that provides time satellite integrity information. Table 1 shows a detailed accurate time, A-GNSS receivers can also receive precise time list of assistance information that corresponds to the above- assistance information and can correct the local oscillator mentioned 3 categories. of A-GNSS receivers using the reference frequency provided Meanwhile, the A-GNSS technique can be divided into from the cell tower of the mobile communication network. MS-Assisted (MSA) and MS-Based (MSB) depending on Fig. 2 roughly shows the flow of assistance information in the whether the element that finally calculates location is from A-GNSS technique. the assistance information from an external server or a

http://www.ipnt.or.kr 208 JPNT 7(4), 205-215 (2018)

Fig. 3. MS-assisted (left) and MS-based (right) positioning methods.

Mobile Station (MS). Fig. 3 shows the flow of assistance data according to MSA or MSB method. First, in the case of MSA, which calculates the user’s location at the server, there is an advantage of sending only a small amount of assistance data from the server to the MS. At this time, the assistance data provided by the server is the list of visible satellites, Doppler, and code phase information corresponding to the acquisition assistance information. These assistance data allows the receiver to rapidly acquire signals by reducing the number of search satellites and the search area. The Doppler, code phase, and C/N0 information obtained are transmitted back to the server, and the server calculates the location of the receiver based on such information. On the other hand, Fig. 4. TTFF in A-GNSS receiver. the MSB method calculates the location at the MS based on the assistance information provided by the server. The assistance information provided to the MS from the server by the movement of the user. Therefore, it is possible to to use the MSB technique includes ephemeris, approximate operate as a normal A-GNSS receiver without additional location and time, and almanac. The A-GNSS receiver can communication with the server for a certain period of time use the assistance information to calculate its location before in which the assistance information provided is valid (about completing tasks such as acquisition and synchronization. 2 hours in case of ephemeris), making the MSB method When using the MSA method, some of the functions of suitable for cases that demand rapid location estimation such conventional GNSS receivers are performed at the assistant as navigation in vehicles. However, the disadvantage is that a server, thereby reducing the processing power consumption relatively large amount of data needs to be provided from the required for signal processing at the MS. In addition, server at the initial stage of operation. It also requires higher when multiple rescue agencies want to acquire location hardware specifications and processing power as the MS information from the MS for emergency rescue activities, calculates the location. it is more efficient to provide the MS location information to each rescue agency from the assistant server than each agency requesting the information. Meanwhile, when the 3. EFFECTS OF ASSISTED DATA ON MS is moving, ongoing communication between the MS and GNSS PERFORMANCE the server is necessary for accurate positioning. Therefore, if a sufficient data rate is not secured, it is not suitable for 3.1 Time to First Fix continually estimating the location of the moving MS. On the other hand, when using the MSB method, the assistance TTFF is the time required for GNSS receivers to acquire information provided by the server is not greatly affected satellite signals and navigation data, and then to calculate https://doi.org/10.11003/JPNT.2018.7.4.205 Kahee Han et al. A-GNSS Performance Test in Various Urban Environments 209

Table 2. Jon's interpretation of dilution of precision values (Person 2008). GDOP value Rating Description 1 Ideal This is the highest possible confidence level to be used for applications demanding the highest possible precision at all times. 2-3 Excellent At this confidence level, positional measurements are considered accurate enough to meet all but the most sensitive applications. 4-6 Good Represents a level that marks the minimum appropriate for making business decisions. Positional measurements could be used to make reliable in-route navigation suggestions to the user. 7-8 Moderate Positional measurements could be used for calculations, but the fix quality could still be improved. A more open view of the sky is recommended. 9-20 Fair Represents a low confidence level. Positional measurements should be discarded or used only to indicate a very rough estimate of the current location. 21-50 Poor At this level, measurements are inaccurate by as much as half a football field and should be discarded.

positioning results. GNSS receivers perform several signal that is, the navigation data bit and timing information, processing procedures until obtaining the first positioning enables an increase in the Predetection Integration Time TTFF is the time required for GNSS receivers to acquire satellite signals and navigation data, and thenresult to calculate after power positioning is supplied. results. GNSS This receiveris showns perform in Fig. several 4, and signal the processing(PIT) procedures in a weak until signal environment such as indoor location. TTFF is the time required for GNSS receivers to acquire satellite signals and navigation data, and obtainingdetailed the formula first positioning is as follows result after(Won power et al. is 2008). supplied. This is shown in Fig.A 4,priori and the knowledge detailed of full navigation data message then then to calculate positioning results.formula GNSS isreceiver as followss perform (Won severalet al. 2008). signal processing procedures until obtaining the first positioning result after power is supplied. This is shown in Fig. 4, and the detailed enables data wipe-off if necessary for an additional signal formula is as follows (Won et al. 2008). (1) processing gain. Therefore, if the navigation data bits 4 RX setting 𝑖𝑖=1 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑏𝑏 𝑏𝑏 𝑏𝑏 𝑠𝑠𝑠𝑠𝑠𝑠 𝑠𝑠 𝑓𝑓𝑓𝑓 (1)𝑓𝑓𝑓𝑓𝑓𝑓 𝑠𝑠𝑠𝑠𝑠𝑠(1)𝑠𝑠 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑provided 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 as assistance information can be synchronized to where isTTFF the receiver= T warm+-up∑ time𝑇𝑇 for +proper𝑇𝑇 operation+ 𝑇𝑇 , is+ the𝑇𝑇 acquisition time for the 4 th TTFF = TRX settingi +satellite∑𝑖𝑖=1 𝑇𝑇 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎signal+, 𝑇𝑇𝑏𝑏𝑏𝑏𝑏𝑏 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 +is 𝑇𝑇the𝑓𝑓𝑓𝑓𝑓𝑓 𝑓𝑓𝑓𝑓bit 𝑠𝑠𝑠𝑠𝑠𝑠synchronization𝑠𝑠 + 𝑇𝑇𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 time for navigation data demodulationthe data, bit edges of the satellite signals for signals intended where is the receiver warm-up 𝑇𝑇time𝑅𝑅𝑅𝑅 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 for𝑠𝑠𝑠𝑠 proper operation, is the acquisition time for the𝑇𝑇 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 th is thewhere starting point search is the time receiver of the navigation warm-up message time frame for (or properpage) for valid datato be bit acquired,demodulation, then the PIT can be extended beyond a single i satellite signal, is the bit synchronizationTRX setting time𝑏𝑏𝑏𝑏𝑏𝑏 𝑠𝑠𝑠𝑠𝑠𝑠 for𝑠𝑠 navigation data demodulation, 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑅𝑅𝑅𝑅 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 and is the𝑇𝑇 navigation message𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 demodulation time. th 𝑇𝑇 is the starting𝑇𝑇 point search time of the navigationoperation, message T frameis the (or acquisition page)𝑇𝑇 for valid time data bit for demodulation, the i satellite navigation data bit duration (20 ms for GPS L1 CA), which 𝑏𝑏𝑏𝑏𝑏𝑏 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 TTFF can be acqicategorized as cold start, warm start, and𝑓𝑓𝑓𝑓𝑓𝑓 𝑓𝑓𝑓𝑓hot 𝑠𝑠𝑠𝑠𝑠𝑠 start𝑠𝑠 depending on the level of initial and is the𝑇𝑇 navigation message 𝑑𝑑demodulation𝑑𝑑𝑑𝑑𝑑𝑑 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 time. 𝑇𝑇 informationsignal,𝑇𝑇 Tofbit a sync GNSS is the receiver bit synchronization (US Coast Guard 1996). time In for particular navigation for cold startresults corresponding in improvements to the in signal sensitivity. TTFF can be categorized as coldfactory start,-default warm state, start, the and acquisition hot start dependingtime can be on reduced the level by of using initial the list of visible satellites, Doppler, 𝑇𝑇𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 data demodulation, is the starting point search time information of a GNSS receiver (USand Coast code Guard phase 1996). provided In particularbyT theframe server. sync for cold In addition, start corresponding the TTFF can to thebe further significantly reduced, since factory-default state, the acquisitionsynchronization timeof the can navigationbe reduced and demodulation by messageusing the list areframe ofnot visible required(or satellites,page) to extractfor Doppler valid information, data such3.3 as Positioningephemeris from Accuracy and code phase provided by the server. In addition, the TTFF can be further significantly reduced, since navigationbit demodulation, data. and T is the navigation message synchronization and demodulation are not required to extract informationdata read such as ephemeris from navigation data. 3.2demodulation Acquisition Sensitivity time. The GNSS positioning accuracy is determined by the

TTFF can be categorized as cold start, warm start, and User Equivalent Ranging Error (UERE) and GDOP, and is 3.2 Acquisition Sensitivity The signal acquisition sensitivity is defined as the minimum signal power required to acquire a certainhot reliabilitystart depending (i.e., detection on theprobability level ofand initial false detection information probability) of (Weilexpressed 2011). When through using statistical distribution. The UERE is an The signal acquisition sensitivity is defined as the minimum signal power required to acquire a acquisitiona GNSS assistance receiver information, (US Coast it is Guard possible 1996). to reduce In the particular Doppler and for code delayindex uncertainty, that showsthereby the distance measurement error for each certain reliability (i.e., detection probabilityreducing theand search false areadetection in the probability) acquisition stage.(Weil Therefore,2011). When signal using detection sensitivity can be increased acquisition assistance information, itwithin iscold possible a startgiven to reducetimecorresponding in the signal Doppler processing, and to codethe by delayfactory-default allocating uncertainty, a long thereby state,dwell time the around satellitesearch cells from that arethe perspective of the receiver. The UERE reducing the search area in the acquisitionexpectedacquisition stage. to have Therefore, timea signal can signal or beby detectionreducingreduced sensitivitythe by size using of can the the be search increasedlist ofcell visible by further subdividingvaries according the reduced to the satellite signal, signal propagation within a given time in signal processing,search area.by allocating a long dwell time around search cells that are expected to have a signal or by reducingsatellites,Meanwhile, the size Doppler,of the the use search of and sensitivity cell code by further assistancephase subdividing provided information, the by reduc thatthe edis,server. the navigation characteristics, data bit and timing and the random changes in the user search area. information,In addition, enables the an TTFF increase can in thebe Predetectionfurther significantly Integration Time reduced, (PIT) in a weakmeasurement signal environment process, and also varies over time for each Meanwhile, the use of sensitivitysuch assistance as indoor information,location. A priori that is, knowledge the navigation of full data navigation bit and datatiming message then enables data wipe-off if information, enables an increase in necessarythesince Predetection synchronizationfor an Integration additional Timesignal and (PIT) demodulationprocessing in a weak gain. signal areTherefore, environment not required if the na tovigation datasatellite. bits provided The GDOP as is a non-dimensional index representing such as indoor location. A priori knowledgeassistanceextract of information informationfull navigation can suchdata be messagesynchronized as ephemeris then enablesto fromthe datadata navigation wipebit edges-off if data.of the satellitethe signals error for caused signals by the satellite’s geometric relationship necessary for an additional signal processing gain. Therefore, if the navigation data bits provided as intended to be acquired, then the PIT can be extended beyond a single navigation datafrom bit duration the perspective (20 ms of the receiver. For a given value, a assistance information can be synchronizedfor GPS L1 to CA), the whichdata bitresults edges in improveof the satellitements in signalssignal sensitivity.for signals intended to be acquired, then the PIT can3.2 be Acquisition extended beyond Sensitivity a single navigation data bit duration (20 ms small GDOP value means more accurate location and time. for GPS L1 CA), which results in improvements in signal sensitivity. 3.3 Positioning Accuracy Since the relative geometrical relationship between satellites

3.3 Positioning Accuracy TheThe GNSS signal positioning acquisition accuracy sensitivityis determined byis thedefined User Equivalent as the Rangingchanges Error (UERE) over andtime, the GDOP also changes over time. As

GDOP,minimum and is expressed signal throughpower statistical required distribution. to acquire The UERE a certain is an index thata result,shows the the distance GDOP can vary according to time and user The GNSS positioning accuracymeasurement is determined error by for the each User satellite Equivalent from Rangingthe perspective Error (UERE) of the receiver. and The UERE varies according to GDOP, and is expressed through statisticalthereliability satellite distribution. signal, (i.e., signal Thedetection propagationUERE is probabilityan characteristics,index that showsand and falsethe the distance random detection changes in thelocation, user measurement and is used to estimate real-time accuracy since measurement error for each satelliteprocess, fromprobability) the and perspective also varies(Weil of over 2011).the receiver.time When for eachThe using UEREsatellite. acquisition varies The accordingGDOP assistance is ato non -dimensional it canindex be representing easily measured by the receiver (U.S. Coast Guard the satellite signal, signal propagationthe errorcharacteristics, caused by andthe satellite’sthe random geometric changes relationshipin the user measurementfrom the perspective of the receiver. For a given process, and also varies over time forvalue, information,each a satellite.small GDOP The it isGDOP valuepossible ismeans a non to -more dimensionalreduce accurate the index Dopplerloca representingtion and and time. code Since the 1996).relative Accordinggeometrical to Person (2008), GDOP values can be the error caused by the satellite’s geometricrelationshipdelay relationship uncertainty, between fromsatellites therebythe perspective changes reducing over of thetime, thereceiver. the search GDOP For a area alsogiven changesin the over time.interpreted As a result, as shown the in Table 2. value, a small GDOP value meansGDOP more can accurate vary according location andto time time. and S inceuser location,the relative and geometricalis used to estimate real-time accuracy since it relationship between satellites changescanacquisition beover easi time,ly measured the stage. GDOP by Therefore, thealso receiver changes ( signalU.S.over Coasttime. detection AsGuard a result, 1996). sensitivity the According to PersonEq. (2008), (2) shows GDOP that the GPS positioning accuracy is due to GDOP can vary according to time valuesandcan user can be location, be increased interpreted and is withinasused shown to estimatea in given Table real 2.time -time in accuracy signal processing,since it the relative geometrical arrangement of satellites used for can be easily measured by the receiver Eq.(U.S. (2) Coast shows Guard that the 1996). GPS Accordingpositioning to accuracy Person (2008),is due toGDOP the relative geometrical arrangement of values can be interpreted as shown insatellites Tableby allocating 2. used for posit a longioning dwell and the time pseudo around-range measurementsearch cells error. that are positioning and the pseudo-range measurement error. Eq. (2) shows that the GPS positioningexpected accuracy to have is due a signal to the orrelative by reducing geometrical the arrangement size of the of search satellites used for positioning and the pseudo-range measurement error. cell by further subdividing the reduced search area. (2) (2) Meanwhile, the use of sensitivity assistance information, For GPS, the systemerror requirement in Positioning accuracy Solution has been= GDOPdefined∗ byUERE the Department of Defense and the North Atlantic Treaty Organization (US Coast Guard 1996). When using navigation assistance information, a lower UERE value for the corresponding satellite can be obtained by lowering the level of satellite ephemeris errors and satellite clock errors. In addition,http://www.ipnt.or.kr more satellites can be obtained as a result of good acquisition sensitivity, and a lower GDOP value can be obtained by using precise time assistance data.

4. FIELD TEST

4.1 Test Environment

The field test was performed using EVK-M8T receivers from u-blox as shown in Fig. 5a and the data from the free AssistNow A-GNSS service by u-blox was used as assistance information. AssistNow data is collected by u-blox's global satellite receivers and is maintained in real-time on u-blox AssistNow servers accessible via the Internet. Fig. 6 shows the GUI used to request assistance information from the AssistNow server. The information that the user can randomly enter into this GUI includes reference location, systems that will receive the assistance information (e.g. GPS, BeiDou, etc.), the type of assistance information to be provided, and the type of time assistance information. A detailed description of each input value is also provided by the manufacturer (Ublox 2015). Meanwhile, the EVK-M8T receiver cannot use GLONASS and BeiDou simultaneously due to its characteristics (Ublox 2015). Therefore, GPS, Galileo, BeiDou and QZSS are used in the experiment in this study, and hereinafter GNSS refers to these systems. The experimental sites were selected as shown in Fig. 5 and Table 3, in order to compare the performance of standalone-GNSS receiver and A-GNSS receiver in various reception environments in urban areas. As shown in Fig. 5, the building lobby and underpass are harsh environments to obtain the LOS between reception antenna and visible satellites. The LOS between satellites can be secured through one side of the glass in the lobby. But in the underpass, it is difficult to secure the LOS between antenna and visible satellites. In apartment complexes or residential areas, the LOS for high elevation satellites can be obtained, but it is difficult obtain the LOS for low elevation satellites. It is possible to obtain more LOS for low elevation satellites in apartment complexes compared to residential areas due to the wide space between buildings. The Rooftop was selected as the highest place around, where it is possible to secure the LOS with all visible satellites at all times. In order to compare performance improvements over standalone-GNSS receivers when using only using assistance data on GPS (i.e. A-GPS) and using assistance data on all GNSS (i.e. A-GNSS), this study performed additional tests using A-GPS receivers in the lobby and open sky (rooftop) environment.

4.2 Test Results

The experiments were performed 10 times for each scenario and the TTFF was analyzed by averaging the TTFF values for each scenario. The signal acquisition sensitivity was analyzed using the average values for each scenario, after selecting the minimum among the satellite signals used at each TTFF point. The positioning accuracy, as shown in Eq. (2), is closely related to the GDOP. Therefore, this paper replaces positioning accuracy with the 𝐶𝐶 /GD𝑁𝑁0OP. The GDOP was analyzed by averaging the values from the TTFF point to the time outputting positioning results 30 times. The reason why the analysis of location accuracy was omitted in this paper is that it was difficult to accurately analyze location errors because of the absence of the precise reference position for indoor positioning. Therefore, the analysis was replaced by determining whether it is possible to estimate the approximate 210 JPNT 7(4), 205-215 (2018)

Fig. 5. Test scenarios and equipment: (a) test equipment, (b) lobby, (c) underpass, (d) apartment complex, (e) residential area, (f) rooftop (open-sky).

For GPS, the system requirement accuracy has been Table 3. Test locations. defined by the Department of Defense and the North Atlantic Test locations Describes Treaty Organization (US Coast Guard 1996). When using Lobby (Fig. 5b) On the 1st floor of the 16-story building One side is made of glass and the other three sides navigation assistance information, a lower UERE value for are clogged with a wall the corresponding satellite can be obtained by lowering the Underpass (Fig. 5c) A narrow passage under the overpass level of satellite ephemeris errors and satellite clock errors. The three sides are clogged with thick concrete walls Apartment complex High-rise apartment complex In addition, more satellites can be obtained as a result of (Fig. 5d) Surrounded by apartments on all sides, and there good acquisition sensitivity, and a lower GDOP value can be are many trees around Distance between buildings is wide obtained by using precise time assistance data. Residential area Low-rise residential area where flats are (Fig. 5e) concentrated Distance between flats is short Rooftop (Fig. 5f) Rooftop of 15-story building 4. FIELD TEST There is no high-rise building in the surroundings and the view is open 4.1 Test Environment

The field test was performed using EVK-M8T receivers provided, and the type of time assistance information. A from u-blox as shown in Fig. 5a and the data from the free detailed description of each input value is also provided by AssistNow A-GNSS service by u-blox was used as assistance the manufacturer (Ublox 2015). Meanwhile, the EVK-M8T information. AssistNow data is collected by u-blox's global receiver cannot use GLONASS and BeiDou simultaneously satellite receivers and is maintained in real-time on u-blox due to its characteristics (Ublox 2015). Therefore, GPS, AssistNow servers accessible via the Internet. Fig. 6 shows Galileo, BeiDou and QZSS are used in the experiment in this the GUI used to request assistance information from study, and hereinafter GNSS refers to these systems. the AssistNow server. The information that the user can The experimental sites were selected as shown in Fig. 5 and randomly enter into this GUI includes reference location, Table 3, in order to compare the performance of standalone- systems that will receive the assistance information (e.g. GNSS receiver and A-GNSS receiver in various reception GPS, BeiDou, etc.), the type of assistance information to be environments in urban areas. As shown in Fig. 5, the building https://doi.org/10.11003/JPNT.2018.7.4.205 Kahee Han et al. A-GNSS Performance Test in Various Urban Environments 211

Fig. 7. Comparison of TTFF [sec] for each scenario; Standalone-GNSS vs. A-GPS vs. A-GNSS.

Table 4. TTFF [sec] for each scenario. Receiver TTFF [sec] Difference of TTFF Standalone- between cold start A-GPS A-GNSS Location GNSS and A-GNSS [sec] Lobby 90.209 9.807 9.217 -61.305 Underpass 329.716 - 16.192 -313.524 Apartment complex 28.646 - 2.432 -26.214 Residential area 30.377 - 3.065 -27.312 Rooftop 27.664 2.524 2.214 -25.45

Fig. 6. The GUI of AssistNow online. Eq. (2), is closely related to the GDOP. Therefore, this paper lobby and underpass are harsh environments to obtain the replaces positioning accuracy with the GDOP. The GDOP LOS between reception antenna and visible satellites. The was analyzed by averaging the values from the TTFF point to LOS between satellites can be secured through one side of the the time outputting positioning results 30 times. The reason glass in the lobby. But in the underpass, it is difficult to secure why the analysis of location accuracy was omitted in this the LOS between antenna and visible satellites. In apartment paper is that it was difficult to accurately analyze location complexes or residential areas, the LOS for high elevation errors because of the absence of the precise reference satellites can be obtained, but it is difficult obtain the LOS for position for indoor positioning. Therefore, the analysis was low elevation satellites. It is possible to obtain more LOS for replaced by determining whether it is possible to estimate the low elevation satellites in apartment complexes compared approximate location on the map. to residential areas due to the wide space between buildings. The Rooftop was selected as the highest place around, where 4.2.1 Time to first fix it is possible to secure the LOS with all visible satellites at all times. In order to compare performance improvements over The comparison of TTFF according to the use of a standalone-GNSS receivers when using only using assistance standalone-GNSS receiver, A-GPS receiver or A-GNSS data on GPS (i.e. A-GPS) and using assistance data on all receiver in each experimental site is shown in Fig. 7. GNSS (i.e. A-GNSS), this study performed additional tests Considering that the low-cost commercial GNSS chipset using A-GPS receivers in the lobby and open sky (rooftop) is the multiple GNSS, this study only tested the Assisted- environment. GNSS in underpass, apartment, and residential area. In the case of poor reception environments such as lobby 4.2 Test Results and underpass, the difference was more than 1 minute depending on the use of assistance information. On the The experiments were performed 10 times for each other hand, in the apartment complex, residential area, and scenario and the TTFF was analyzed by averaging the TTFF rooftop where the reception environment is relatively good, values for each scenario. The signal acquisition sensitivity the difference in TTFF was observed to be the amount of was analyzed using the average values for each scenario, after time required to extract the ephemeris. Table 4 summarizes selecting the minimum C/N0 among the satellite signals used the TTFF obtained for each experimental site and for each at each TTFF point. The positioning accuracy, as shown in receiver used. This study confirmed that the performance

http://www.ipnt.or.kr 212 JPNT 7(4), 205-215 (2018)

Table 5. Sensitivity in different receiver mode. System Sensitivity [dBm] Scenario GPS & BDS GPS BDS GAL Cold start -148 -148 -143 -138 Aided acquisition -157 -157 -146 -142

Table 6. Minimum C/N0 in different receiver mode.

System Minimum C/N0 [dB-Hz] Scenario GPS & BDS GPS BDS GAL Cold start 20 ~ 23 20 ~ 23 25 ~ 28 30 ~ 33 Aided acquisition 11 ~ 14 11 ~ 14 22 ~ 25 26 ~ 29

location on the map. Fig. 9. Comparison of GDOP for each scenario; Standalone-GNSS vs. A-GPS vs. A-GNSS. location on the map. 4.2.1 Time to first fix

4.2.1 Time to first fix The comparisonTable 7. GDOPof TTFF for eachaccording scenario. to the use of a standalone-GNSS receiver, A-GPS receiver or A- GNSS receiver in each experimental site is shown in Fig. 7. Considering that the low-cost commercial The comparison of TTFF according to the use ofGNSS a standalone chipset- GNSSis the multiplereceiver,Receiver GNSS, A-GPS this receiver studyGDOP oronly A- tested the DifferenceAssisted- ofGNSS GDOP in underpass, apartment, GNSS receiver in each experimental site is shown in Fig. 7. Considering that the Standalone-low-cost commercial between standalone- and residential area. In the case of poor receptionA-GPS A-GNSSenvironments such as lobby and underpass, the GNSS chipset is the multiple GNSS, this study onlydifference tested th waseLocation Assisted more than-GNSS 1 minute in underpass,GNSS depending apartment, on the use of assistanceGNSS and information. A-GNSS On the other hand, in and residential area. In the case of poor receptionthe environmentsapartmentLobby complex, such asresidential lobby and 11.24area, underpass, and rooftop2.174 the where 2.937 the reception-8.303 environment is relatively good, difference was more than 1 minute depending on thethe use difference of assistanceUnderpass in TTFF information. was observed On3.958 the to otherbe the hand, amount- in 2.316of time required-1.642 to extract the ephemeris. Table 4 the apartment complex, residential area, and rooftop where the reception environment is relatively good, summarizesApartment the TTFF complex obtained for5.43 each experimental- 1.873 site and for-3.557 each receiver used. This study the difference in TTFF was observed to be the amountconfirmed of time that required the performan to extractce the was ephemeris. improved Tablein terms 4 of TTFF when using assistance information for all summarizes the TTFF obtained for each experimental site Residentialand for each area receiver4.381 used. This -study 1.636 -2.745 experimentalRooftop sites. In particular, in1.477 the underpasses1.436 and 1.653lobby where the0.176 reception is very poor, the TTFF confirmed that the performance was improved in termsgains of wereTTFF 313 when seconds using andassistance 61 seconds, information respectively, for all when using assistance information, which meet the experimental sites. In particular, in the underpasses 30and-second lobby whererequirement the reception by the FCCis very (2015). poor, theMeanwhile, TTFF the reason why the TTFF gain varies depending gains were 313 seconds and 61 seconds, respectively, when using assistance information, which meet the on experimental site is that signals with low C/N0 take a relatively longer amount of time to acquire 30-second requirement by the FCC (2015). Meanwhile, the reason why the TTFF gain varies depending ephemeris and signals compared to signals with high C/N0. In conclusion, as shown in Fig. 7 and Table 4, Fig. 8. Comparisonon of experimental minimum value site of is used that signals signals C/N0 with [dB-Hz]; low C/N although0 take a relativelythe provision longer of assistanceamount of informationtime to acquire leads to a large TTFF performance difference, whether Standalone-GNSSephemeris vs. A-GNSS. and signals compared to signals with high C/N0. In conclusion, as shown in Fig. 7 and Table 4, the assistancerelatively information low reception provided was C/N from among a large the number 5 experimental of GNSS (i.e.,sites. whether it is an A-GPS although the provision of assistance information leadsreceiver to a largeor an ATTFF-GNSS performance receiver) did difference, not have0 whethera significant impact. the assistance information provided was from a large number ofAccording GNSS (i.e., to whetherthe data it sheetis an A provided-GPS by u-blox, the EVK- was improvedreceiver in terms or an Aof-GNSS TTFF receiver) when didusing not have assistance a significant4.2. 2 Acquisition impact.M8T receiver sensitivity has different acquisition sensitivities depending information for all experimental sites. In particular, in the on the system used for positioning and the use of assistance 4.2.2 Acquisition sensitivity In order to effectively analyze the influence of the use of assistance information received from the underpasses and lobby where the reception is very poor, theserver on acquisitiondata. The sensitivity, sensitivity the analysis values in accordingthis study was to focused each oncase the lobbyare and underpass, which In order to effectively analyze the influence of the use of assistance information received from the TTFF gains were 313 seconds and 61 seconds, respectively,are expectedsummarized to have relatively in Table low reception 5 (Ublox 2015). among The the minimum 5 experimental C/N sites.that server on acquisition sensitivity, the analysis in this studyAccording was focused to onthe thedata lobby sheet and provided underpass, by whichu-blox, the EVK-M8T receiver0 has different acquisition when using areassistance expected to information,have relatively low which reception meet theamongsensitivities the 5can experimentaldepending be received on sites. the based system on used the for 𝐶𝐶sensitivity /positioning𝑁𝑁0 values and the in use Table of assistance 5 can data. The sensitivity According to the data sheet provided by u-blox,values the according EVK-M8T to receivereach case has are different summarized acquisition in Table 5 (Ublox 2015). The minimum that can be 30-second requirement by the FCC (2015). Meanwhile, 𝐶𝐶/𝑁𝑁0 the be calculated as shown in Eq. (3). sensitivities depending on the system used for positioningreceived and based the onuse the of sensitivityassistance valuesdata. The in Table sensitivity 5 can be calculated as shown in Eq. (3). reason why thevalues TTFF according gain varies to each depending case are summarized on experimental in Table 5 (Ublox 2015). The minimum that can be 𝐶𝐶/𝑁𝑁0 site is that signalsreceived with based low on the sensitivity take a values relatively in Table longer 5 can be calculated as shown in Eq. (3). (3) C/N0 𝐶𝐶/𝑁𝑁0 amount of time to acquire ephemeris and signals compared 0 The thermal 𝐶𝐶 noise/𝑁𝑁 = power(𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 density 𝑝𝑝𝑝𝑝𝑝𝑝of the𝑝𝑝 𝑝𝑝 )M8T + ( (3)𝑡𝑡 receiverℎ𝑒𝑒𝑒𝑒 𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛is typically 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 174dBm) −(3)(𝑁𝑁 and𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 the 𝐹𝐹𝐹𝐹 𝐹𝐹𝐹𝐹noise𝐹𝐹𝐹𝐹) figure is to signals with high C/N . In conclusion, as shown in Fig.3~6 7 dB. The minimum receivable according to each case, calculated based on Table 5 and Eq. (3), The thermal0 𝐶𝐶 noise/𝑁𝑁0 = power(𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 density 𝑝𝑝𝑝𝑝𝑝𝑝of the𝑝𝑝𝑝𝑝 )M8T+ (𝑡𝑡 receiverℎ𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛is typically 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 174dBm) − (𝑁𝑁 and𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 the 𝐹𝐹𝐹𝐹 𝐹𝐹𝐹𝐹noise𝐹𝐹𝐹𝐹) figure is and Table 4, although the provision of assistance informationis as shown inThe Table thermal 6. Using noiseassistance power information density through of the Table M8T 6 results receiver in a gain of at least 3 3~6 dB. The minimum receivable according todB each-Hz andcase up, calculated to 11 dB- Hz.based on Table𝐶𝐶/𝑁𝑁0 5 and Eq. (3), leads to a largeis asTTFF shown performance in Table 6. Using difference, assistance whether information the through Fig. 8isTable shows typically 6 resultsthe minimum 174dBm in a andgain in ofthe theat leastnoiselobby 3 andfigure underpass. is 3~6 WhendB. The using𝐶𝐶 / 𝑁𝑁a 0standalone-GNSS dB-Hz and up to 11 dB-Hz. 𝐶𝐶/𝑁𝑁0 assistance information provided was from a large numberreceiver inminimum the lobby, the receivable average minimum0 according of satellitesto each usedcase, at calculated the TTFF point is 32 dB-Hz, and 𝐶𝐶/𝑁𝑁 C/N0 0 Fig. 8 shows the minimum in the lobbywith and an underpass.A-GNSS r eceiverWhen usingusing assistancea𝐶𝐶 standalone/𝑁𝑁 information,-GNSS the average minimum is 20 dB-Hz, which of GNSS (i.e., receiverwhether in itthe is lobby,an A-GPS the average receiver minimum or an A-GNSS of satellites based used at on the Table TTFF 5point and is Eq. 32 dB(3),-Hz, is 0asand shown in Table 6. Using 0 results in a gain of about 12 dB-Hz. When𝐶𝐶/𝑁𝑁 using a standalone-GNSS receiver in the underpass, the with an A-GNSS receiver using assistance𝐶𝐶/𝑁𝑁 information, the average minimum is 20 dB-Hz, which 0 receiver) did not have a significant impact. 0 average minimumassistance information of satellites used through at the TTFFTable point 6 results is 27 dBin- aHz, C/N and0 gainthe𝐶𝐶 /average𝑁𝑁 minimum results in a gain of about 12 dB-Hz. When𝐶𝐶/𝑁𝑁 using a standalone0-GNSS receiver in the underpass, the when using 𝐶𝐶an/ 𝑁𝑁A- GNSS receiver0 is 17 dB-Hz, which results in a gain of about 10 dB-Hz. average minimum of satellites used at the TTFF point is of27 atdB least-Hz, and30 dB-Hz the𝐶𝐶 /average𝑁𝑁 and upminimum to 11 dB-Hz. 0 0 𝐶𝐶/𝑁𝑁 𝐶𝐶/𝑁𝑁 4.2.2 Acquisitionwhen sensitivity using 𝐶𝐶an/ 𝑁𝑁A- GNSS receiver is 17 dB-Hz, which results in a gain of about 10 dB-Hz. 0 4.2.3 GDOP Fig. 8 shows the minimum C/N0 in the𝐶𝐶/𝑁𝑁 lobby and 𝐶𝐶/𝑁𝑁0 𝐶𝐶/𝑁𝑁0 underpass.0 When using a standalone-GNSS receiver in the 4.2.3 GDOP 𝐶𝐶/𝑁𝑁 Fig. 9 and Table 7 show the GDOPs according to the use of a standalone-GNSS receiver, A-GPS In order to effectively analyze the influence of the use lobby, the average minimum C/N of satellites used at the receiver or A-GNSS receiver for all experimental sites.0 Through evaluating the GDOP value in the lobby of assistance informationFig. 9 and Table received 7 show fromthe GDOPs the serveraccording basedon to the on useTableTTFF of 2,a pointitstandalone was isfound 32-GNSS dB-Hz,that the receiver, positioningand with A-GPS resultsan A-GNSS of the standalone receiver -usingGNSS receiver could only be receiver or A-GNSS receiver for all experimental sites. Through evaluating the GDOP value in the lobby acquisition sensitivity,based on Table the analysis2, it was foundin this that study the positioning was focused results of theassistance standalone information,-GNSS receiver the could average only be minimum C/N0 is 20 dB- on the lobby and underpass, which are expected to have Hz, which results in a C/N0 gain of about 12 dB-Hz. When https://doi.org/10.11003/JPNT.2018.7.4.205 Kahee Han et al. A-GNSS Performance Test in Various Urban Environments 213

Fig. 10. Comparison of skyplot for lobby environment; Standalone-GNSS (left) vs. A-GPS (center) vs. A-GNSS (right). using a standalone-GNSS receiver in the underpass, the also be used to estimate location by using the assistance average minimum C/N0 of satellites used at the TTFF point information. In other words, the use of assistance data for all is 27 dB-Hz, and the average minimum C/N0 when using an experimental sites resulted in a lower GDOP value. Therefore,

A-GNSS receiver is 17 dB-Hz, which results in a C/N0 gain of it can be confirmed that A-GPS and A-GNSS receivers using about 10 dB-Hz. assistance information have better positioning accuracy than standalone-GNSS receivers. 4.2.3 GDOP

Fig. 9 and Table 7 show the GDOPs according to the use 5. CONCLUSION of a standalone-GNSS receiver, A-GPS receiver or A-GNSS receiver for all experimental sites. Through evaluating the This paper examined the reception sensitivity, GDOP, GDOP value in the lobby based on Table 2, it was found and TTFF, in order to analyze the performance of A-GNSS that the positioning results of the standalone-GNSS receiver receivers using assistance information in Korea according could only be used as approximate estimates with low to the rapidly changing various GNSS environments. confidence levels, while the A-GPS and A-GNSS receivers Independent from the one-sided receiver performance could be used as estimates with good location accuracy. specifications announced by receiver manufacturers, this GDOP performance improvements were also observed in study tested the performance of the A-GNSS service for the underpass, apartment complex, and residential area. multiple GNSS signals currently available in the sky of Korea. However, the rooftop where sufficient LOS signals can be When the receiver status was set to cold start mode, the obtained showed small differences among standalone-GNSS A-GNSS receiver, compared to standalone-GNSS receiver, receiver, A-GPS receiver, and A-GNSS receiver. resulted in a reduction in the minimum C/N0 of available Fig. 10 presents the skyplot obtained by the receiver satellite signals by about 12 dB-Hz in the lobby and by about from 20 seconds after the receiver started to operate in the 10 dB-Hz in the underpass in a building. This showed that lobby. In the skyplot, 'G' refers to GPS, 'E' to Galileo, 'B' to the use of assistance information enables the estimation BeiDou, and 'Q' to QZSS. The green satellites are the satellites of locations using satellite signals even in a very poor currently used for estimating location, blue ones indicate reception environment. As satellite signals of low C/N0 were satellites in the tracking stage after acquiring signals, and also used, the number of satellites available to the receiver red ones are satellites that have not yet acquired signals. increased, thereby reducing the GDOP. In particular, a Since A-GPS receivers and A-GNSS receivers receive the stand-alone GNSS receiver operating in the cold start mode ephemeris of the satellites from the assistant server, even if has a ‘Fair’ GDOP performance value with a maximum of the ephemeris cannot be extracted from the received signal, 11 depending on location, but when using A-GNSS, the the corresponding satellite is displayed on the skyplot. That GDOP has an ‘Excellent’ value of less than 3 regardless of is, the number of satellites used to estimate location also location. Therefore, the positioning accuracy is expected to increased after the same period of time when using assistance increase when using assistance information. The TTFF also information. The increase in the number of satellites used decreases as navigation signals with low C/N0 are used to is possible because satellite signals with low C/N0 can estimate location. The results of the experiments confirmed

http://www.ipnt.or.kr 214 JPNT 7(4), 205-215 (2018)

that the TTFF decreased when using assistance information eCall functionality in an IVS, Application Card, Ver. for all experimental sites, and the TTFF became faster by 01.00, [internet], cited 2018 Nov 1, available from: a minimum of 25.45 seconds and a maximum of 313.524 https://cdn.rohde-schwarz.com/pws/dl_downloads/ seconds depending on location. In particular, when using dl_application/pdfs/Verification_of_Next_Generation_ assistance information for all sites, the TTFF was less than eCall_functionality_of_an_IVS_ac_5215-9683_v0100. 30 seconds, which was the TTFF requirement of e-Call. pdf Therefore, it is expected that assistance information will be Rohde & Schwarz 2018b, Test your eCall and ERA-Glonass very helpful for systems such as emergency rescue services system modules, Application Card, Ver. 04.00, [internet], where fast location estimation speed is important. cited 2018 Nov 1, https://cdn.rohde-schwarz.com/ pws/dl_downloads/dl_application/pdfs/SMBV100A_ CMW500_Test_ac_en_5214-5532-92_v0400.pdf ACKNOWLEDGMENTS Singh, S. 2006, Comparison of assisted GPS (AGPS) performance using simulator and field tests, In This research was supported by a grant (18NSIP-B082188 Proceedings of the ION GNSS Conference, 26-29 -05) from National Land Space Information Research September, Fort Worth, Texas, USA, pp.1-12. Program funded by Ministry of Land, Infrastructure and The European Parliament and the Council of the European Transport of Korean government and Korea Agency for Union 2015, Regulation (EU) 2015/758 of the European Infrastructure Technology Advancement. Parliament and of the Council concerning type- approval requirements for the deployment of the eCall in-vehicle system based on the 112 service and REFERENCES amending directive 2007/46/EC, Official Journal of the European Union, 123, 77-89. ECC 2016, Establishing Criteria for the Accuracy and TTA 2015, Positioning system for Public Safety Stage 2: Reliability of the Caller Location Information in support Architecture, TTAK.KO-06.0401-Part2, TTA, Seongnam, of Emergency Services, Report 225. Korea (in Korean with English abstract) FCC 2015, In the Matter of Wireless E911 Location Accuracy TTA 2016a, Positioning system for Public Safety Stage 3: Requirements, PS Docket No. 07-114, Washington D.C. Interfaces, TTAK.KO-06.0434, TTA, Seongnam, Korea Kaplan, E. D. & Hegarty, C. J. 2006, Understanding GPS: (in Korean with English abstract) Principles and Applications, 2nd ed. (Boston: Artech TTA 2016b, Positioning system for Public Safety Stage 4: House Inc.) Test Requirements, TTAK.KO-06.0435, TTA, Seongnam, Karunanayake, M. D., Cannon, M. E., & Lachapelle, G. 2007, Korea (in Korean with English abstract) Analysis of assistance data on AGPS performance, TTA 2017a, Positioning system for Public Safety Stage Measurement Science and Technology, 18, 1908-1916. 1: Requirement, TTAK.KO-06.0401-Part1/R1, TTA, Korea Transportation Safety Authority 2016, Emergency Seongnam, Korea (in Korean with English abstract) Rescue System e-Call, TS magazine, October 2016, TTA 2017b, Positioning system for Public Safety Stage 5: Test [internet], cited 2018 Nov 1, available from: http:// Procedure, TTAK.KO-06.0401-Part5, TTA, Seongnam, ts.ecatalog.kr/src/viewer/main.php?host=main&si Korea (in Korean with English abstract) te=20161002_142301_3 Ublox 2015, NEO/LEA-M8T u-blox M8 concurrent GNSS López, J. M. L., Aguilar, F. L., & Abascal, J. J. C. 2012, A-GPS timing modules data sheet, UBX-14006196-R06. https:// performance in urban areas, Mobile Multimedia www.u-blox.com/sites/default/files/NEO-LEA-M8T_ Communications: 6th International ICST Conference, DataSheet_(UBX-14006196).pdf MOBIMEDIA 2010, Lisbon, Portugal, September U.S. Coast Guard 1996, NAVSTAR GPS User equipment 6-8, 2010. Revised Selected Papers, eds. Rodriguez, introduction, Public Release Version. J., Tafazolli, R., Verikoukis, C. (Berlin, Heidelberg: van Diggelen, F. S. T. 2009, A-GPS: Assisted GPS, GNSS, and Springer), pp.413-422. SBAS (Boston: Artech House) Person, J. 2008, Writing your own gps applications: Part 2, Weil, L. R. 2011, Difference Between Signal Acquisition and Developer Fusion, [internet], cited 2018 Nov 1, available Tracking, Inside GNSS, 6, 22-27. from: https://www.developerfusion.com/article/4652/ Won, J.-H., Eissfeller, B., Lankl, B., Schmitz-Peiffer, A., writing-your-own-gps-applications-part-2/2/. & Colzi, E. 2008, C-Band User Terminal Concepts Rohde & Schwarz 2018a, Verification of next generation and Acquisition Performance Analysis for European https://doi.org/10.11003/JPNT.2018.7.4.205 Kahee Han et al. A-GNSS Performance Test in Various Urban Environments 215

GNSS Evolution Programme, Proceedings of the 21st International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2008), Savannah, GA, Sep 16-19 2008, pp.940-951.

Kahee Han is a M.S. student in the Depart- ment of Electrical Engineering at Inha Uni- versity, South Korea. She received B.S. degree in Electronic Engineering from Inha Uni- versity in 2017. Her research interests include GNSS signal design and Assisted-GNSS.

Jung-Hoon Lee graduated Electric Engin- eering at Inha University, Incheon, South Korea in 2017, and currently is in the master's degree course in Autonomous Navigation Laboratory at the same university.

Ji-Ung Im graduated Computer Science Engineering at Inha University, Incheon, South Korea in 2017, and currently is in the master's degree course in Autonomous Navi- gation Laboratory at the same university. His research interests include self-driving cars.

Jong-Hoon Won received the Ph.D degree in the Department of Control Engineering from Ajou university, South Korea, in 2005. After then, he had worked with the Institute of Space Technology and Space Applications at University Federal Armed Forces (UFAF) Munich, Germany. He was nominated as Head of GNSS Laboratory in 2011 at the same institute and involved in lectures on advanced receiver technology at Technical University of Munich (TUM) since 2009. He is currently an assistant professor of Electrical Engineering of Inha University, South Korea. His research interests include GNSS signal design, receiver, navigation, target tracking systems and self-driving cars.

http://www.ipnt.or.kr