The Potential to Use Smartphone‐based GNSS Receivers for Surveying

The aim of the present study was to investigate the possibility of using smart‐ phone‐grade GNSS receivers for surveying purposes and to examine their ac‐ curacy in terms of positioning. The paper presents an analysis of the results of test measurements made using two smartphones: a Huawei P8 lite, pow‐ ered by Android 6.0 and an iPhone 6s, operating on IOS 10.1.1. Both devices were equipped with GPS and GLONASS modules and had the supporting sys‐ tem, A‐GPS. Measurements were made on two test parcels of different shapes and sizes. The test procedure encompassed three measurement series (I–III), each of which consisted of five determinations. The determinations were done at 5‐min time intervals and each new measuring session began after 90 min from the previous one. Measurements with the Huawei P8 lite were taken using the Locus Map app, and measurements with the iPhone 6s were performed with the use of Map‐o‐meter. The study consisted of determining the positioning accuracy of the test smartphones by comparing the results with reference coor‐ dinates obtained from RTN measurements using the Topcon HiPer V receiver with Magnet Tools software. Determinations of the reference coordinates were done with the help of the TPI NetPro network real‐time service corrections.


Introduction
The rapid technological development and miniaturization of GNSS receivers allowed their use in mobile phones. The first device of this type was the Benefon ESC! phone, released in 1999, equipped with a GPS positioning system operating on two GSM frequencies of 900 and 1800 MHz [1]. It had a fairly broad array of features such as displaying the user's position on a map, calculating the position, elevation above sea level and average speed, and setting the current position and destination waypoints. Following the announcement in 2011 of full operational capability of the GLONASS system [2], it became standard practice to use two navigation systems in smartphones. The first smartphone of this type was ZTE MTS 945 [3]. Today, smartphones use solutions supporting the operation of satellite navigation systems such as Assisted GPS (A-GPS), Enhanced GPS (E-GPS) [1], Mobile RTK (mRTK) [4,5], and Beidou, the Chinese satellite navigation system.
A-GPS augments GPS positioning in poor satellite signal conditions, e.g. in urban canyons and closed facilities [6]. Auxiliary information needed to determine the coordinates is supplied directly by the cellular network via GPRS, a packet-oriented mobile data service [1]. A-GPS shortens the time needed to download ephemeris data by supplying them directly from the mobile station, instead of the slower satellite link. Mobile RTK (mRTK) solution enable high-accuracy double difference carrier phase positioning using a smartphone [5]. It utilizes single-band assisted-GPS receivers, cellular link and inertial sensors [4]. Recently, it has been proven that positioning using CDGNSS (carrier-phase differential GNSS) with centimeter accuracy is possible, using low-quality, low-cost antennas mounted in smartphones [7]. However, miniaturized GPS chipsets still cannot compete with the sturdy high-accuracy antenna systems of precision GPS receivers [8].
Current technological development makes it relatively easy to create Web services for everyday use [9], which is why there are more and more applications available online that use GNSS services to perform measurements. The most frequently downloaded free applications include GPS Essentials, Locus Free Maps and GPS Measure Field Area for Android and iGeoTrans Lite, Map-o-meter, and Area -Calculate my Area for iOS.
The aim of the present study was to test the accuracy of determination of horizontal coordinates using smartphone-based GNSS receivers and to analyze the possibility of their use in surveying. The study also compared the results of determinations made using smartphones with those made using the RTN Topcon Hiper V set with the FC-2600 controller.

Method
The study examined the positioning accuracy of two popular one-frequency GNSS smartphones: the Huawei P8 lite powered by the Android 6.0 operating system and an iPhone 6s operating on IOS version 10.1.1. Both devices had GPS and GLONASS modules and the supporting system, A-GPS.
The tests were performed on November 2 nd , 2016 between 11.00 and 16.00 hours on two test parcels differing in shape and surface area. The parcels were established in the city of Lublin, in Józef Sowiński Street, and were marked with wooden stakes (Fig. 1). They were located in optimum field conditions with a very good view of the sky. The PDOP (position dilution of precision) coefficient during measurement ranged from 1.01 and 1.36 [10].

Fig. 1. Test parcel 1 and test parcel 2
Source: [11] Since there is no ISO standard for accuracy testing of GNSS measurements performed with smartphones, we proposed a new procedure based partly on ISO 17123-8 GNSS Field Measurement Systems in Real-Time Kinematic (RTK) as described by [12]. This procedure encompassed three measurement series (I-III), each of which consisted of five determinations (A, B, C, D, E). The determinations were done at 5-min time intervals and each new measuring session began after 90 min from the previous one. This was done to minimize the impact of changes in the configuration of satellites, deviation of their orbits, as well as ionospheric, tropospheric and environmental conditions of the receiver. The results of the determinations were compared with a Real Time Network (RTN) measurement performed using the Topcon HiPer V receiver with Magnet Field software using TPI NETpro VRS corrections. This receiver can be used to determine the horizontal position to an accuracy of 1 cm [13].
Because the test smartphones had different operating systems, the tests were done using different measurement applications. Determinations with the Huawei P8 lite were done using Locus Map, and measurements with the iPhone 6s were performed with the use of Map-o-meter. Both applications allow the setting of individual waypoints and route-tracking.

Results and Discussion
To visualize the results, the observation data obtained in KML format were imported to the QGIS geographic information system application. The results of determinations of the vertices of the test parcels are presented in Figures 2 and 3. One notices immediately that the coordinates measured using the iPhone 6s are more precise and are more closely clustered around the results of the RTN measurement (marked in aquamarine) than those measured with Huawei P8 lite.
As a next step, standard deviations of the results were calculated using Microsoft Excel. Figure 4 shows a comparison of standard deviations of the horizontal coordinates of the vertices of test parcels 1 and 2. An analysis of the data in Figure 4 shows that the largest standard deviations were obtained when positioning was done with the Huawei P8 lite. They fell within the range of 1.32 to 4.15 m for the X coordinate and 1.63 to 2.87 m for the Y coordinate. The standard deviations for the iPhone 6s were in the range of 0.53 to 2.42 m for X and 0.92 to 1.88 m for Y. An analysis of Figure 5, presenting the standard deviations of a single point position determination, demonstrates that the measurement taken with iPhone 6s exhibited a much smaller error than that performed using Huawei P8 lite. The mean error was 1.75 m for the horizontal coordinates. To compare, the mean error for Huawei P8 lite was more than twice as large at 3.67 m.

Conclusions
Comparing the results of the present study with the results of [1], it can be stated that the differences between measurements taken with an RTN precision receiver and a smartphone have decreased fivefold over the last 6 years. The current rapid development of mobile phones and GNSS chipsets, which are their inseparable components, shows that the measurement accuracy of smartphones may greatly improve in a short time.
Studies have shown that smartphones can be used for quick and easy measurement of horizontal coordinates but are not very precise instruments. The mean standard deviations of determinations of horizontal coordinates were in the range from 1.27 to 2.93 m for the iPhone 6s and from 2.84 to 5.04 m for the Huawei P8. The accuracy of point determinations was 1.60 m for the Huawei P8 lite and 2.00 m for the iPhone 6s. The average confidence interval for the mean for the iPhone 6s was between −1.44 and −0.12 m for dX and −2.11 and −0.67 m for dY and for the Huawei P8 lite between −1.89 and 1.17 m for dX and −1.37 and 1.40 m for dY.
The accuracy of the measurements taken with the smartphones was within the range of accuracies provided by the ASG-EUPOS NAWGIS service (1-3 m) which assists satellite navigation and the field control of topographic maps and GIS systems [14]. They can also be used for mapping-survey measurements performed to create or update a database for group II and III field details (3 m -group II, 5.0 mgroup III, §29.2. [15]). However, the much worse accuracy of the measurements indicates that the variation of the obtained coordinates may be large. Additional studies of the accuracy of smartphone measurements in less favorable conditions, for example, in the vicinity of densely built-up areas, should be carried out to establish whether such an application is practicable.