Local monitoring by low cost devices and free and open sources softwares Abstract Ludovico Biagi, Florin-Catalin Grec, Marco Negretti, Maria Grazia Visconti Politecnico di Milano, DICA@ComoCampus The purpose of local monitoring is to closely model displacements, evolution and deformations of civil engineering structures and model local phenomena like landslides. For certain applications, latency of the monitoring results is not a crucial point. In other cases a small latency is fundamental: for example, a suspected Alpine landslide that insists on villages. In these cases, contiunous quasi real time monitoring by permanent stations is required. Here, a typical monitoring configuration is given by a GNSS reference permanent station, outside the monitored area, in a stable site and several permanent stations in the monitored area. Data can be processed and the results can be monitored in a quasi continuous way by automated procedures. In these situations, the use of low cost GNSS receivers could be interesting because it can significantly increase the number of monitored points without notably increasing the network cost. This investigation is the scope of our work and several experiments have been set up and performed to assess them. To reach our goals, geodetic and low cost receivers are teamed up: Leica GR10 as reference station and several single frequency u-blox 7P receivers. Once the aquisition phase is completed data are processed with commercial (Leica GeoOffice), academic (Bernese) and free and open source softwares (gogps and RTKLIB). The results are compared in the final step of the project, accuracy and reliability of solutions are discussed. Their results will be described in the presentation and the full paper. Keywords Low cost receivers, monitoring, deformations, landslides 1 Introduction Several experiments are planned and will be performed in the next months. At the present, a first experiment is under analysis in order to assess whether or not is it possible to use u-blox receiver for precise geodetic applications. On the roof of the Politecnico of MILANO building a week of u-blox static data has been collected, from the GPS day 074 to GPS day 079 for the year 2014 (from 15 to 20 of March 2014). The place chosen for the experiment is about 60 m distant from Milano GNSS permanent station. The data have been acquired at 1 second sampling rate. 431
2 Results Geomatics Workbooks n 12 "FOSS4G Europe Como 2015" Milano PS coordinates have been estimated in ITRF2008. The reference coordinates for u-blox receiver have been calculated with LGO software in post processing with respect to MILANO permanent station. The following processing parameters have been adopted: 10 degrees cut off angle, Klobuchar ionospheric model, Saastamoinen tropospheric model and broadcast ephemerides. Reference UTM coordinates for u-blox station are the following: E = 517 880.331 m; N = 5 036 253.753 m; h = 193.380 m Then, u-blox data have been processed in LGO by hourly static sessions to assess their accuracy. Therefore, 168 results have been obtained. First of all, during the processing 4 blunders, whose results differ more than 5 cm from reference solution, were identified for the whole session. For these solutions, ambiguities have not been resolved and errors' magnitude is of about 20 cm, the maximum reaching 1 m for East component. Table 1 provides a more detail review of the outliers influence in the final 3D position of u-blox station. Session ds [cm] 34 37.1 58 12.3 106 129.9 160 38.9 Table 1: Vectors of residuals for outliers. At a second step, the same procedure was carried in Bernese. Bernese processing parameter are set up according to guidelines for single frequency, short baselines: 10 degrees cut off angle, dry Neill tropospheric model, IGS ionosphere model, precise ephemerides double difference on phase (L1 carrier), Sigma method for ambiguities handling. At the time of this contribution only preliminary results are ready. The results are quite similar with those from LGO but 5 blunders are present, probably caused by problems in solving of cycle slips: sessions 22, 24, 30, 40 and 142. By averaging the hourly solutions, except those affected by blunders, Bernese gives as final solution the exact reference coordinates. Table 2 reports the results of the nor bludered sessions, both for LGO and Bernese. Leica GeoOffice Statistics deast [m] dnorth [m] Bernese dh [m] deast [m] dnorth [m] dh [m] Mean 0.0002-0.0009 0.0019 0.0009-0.0015 0.0008 RMS 0.0018 0.0025 0.0044 0.0021 0.0049 0.0045 Max 0.0041 0.0033 0.0145 0.016 0.047 0.0143 Min -0.004-0.0089-0.008-0.0031-0.010-0.010 Table 2: Comparison between statistics of residuals from LGO and Bernese. 432
Below, the graphical representation of the residuals of LGO in the 3 components is provided without blunders to help visual interpretation of results. They prove that u-blox receivers are suitable for applications with cmlevel accuracy: for the 2D position the errors are well below 1 cm in most of the cases. However, for height, as expected, the residuals are slightly higher but still without moving away from the cm-level accuracy (around 1.5 cm the maximum residual). Figure 1: Time series for residuals (LGO processing), static method, 1 hour sessions East (blue), North (yellow) and h (red). These results were confirmed by Bernese processing. Again, for the horizontal components the residuals are slightly lower than for the vertical component and the number of blunders is smaller. Float solutions are graphically compared with the fix ones. From doing so, it can be easily seen the impact of ambiguities on the final results. Once this step was fulfilled, in good sessions the residuals shrunk from cm level to mm level. On the contrary, for problematic sessions, fixing worsenes the solution because the residuals increase with respect to float solution. These problems are presented in Table 3. 433
Session dn [cm] dh [cm] 21 - From -18.1 to -22.4 23 From 37.7 to 39.7 From -15.0 to -15.4 40 - From -2.47 to -5.9 142 From 9.5 to 10.1 - Table 3: Change in residuals magnitude after fixing ambiguity for problematic sessions. Figure 2: Comparison between residuals for float and fix residuals (Bernese processing), static method, 1hour sessions. At the moment of writing the abstract, the data are also being analysed with two free and open source softwares, gogps and RTKLIB, and the results will be compared with LGO reference results. gogps is a software package developed at the Geomatics Laboratory of Politecnico di Milano designed to improve the positioning accuracy of low-cost (single-frequency) GPS devices by RTK technique. Currently it requires raw data (observations) at 1 Hz in input. RTKLIB is an open source program package for standard and precise positioning with GNSS (global navigation satellite system). It supports standard and precise positioning algorithms with: GPS, GLONASS, GALILEO, QZSS, BEIDOU and SBAS. 434
It supports various positioning modes with GNSS for both real-time and postprocessing: Single, DGPS/DGNSS, Kinematic, Static, Moving-Baseline, Fixed, PPP-Kinematic, PPP-Staticand PPP-Fixed. The results will be presented in the final paper. 3 Conclusions In general, low cost receivers like navigation type or chip sets use the phase only to smooth the code observations but some manufactures like u-blox offer the option to extract the phase data. This first experiment has been performed as a preliminary step just to test the accuracy and reliability of u-blox receivers for monitoring purposes. The investigations reveal that up to now the post-processing of phase observations meet the required accuracy for monitoring of local phenomena. Therefore, in the next months experiments more appropriate to landslides monitoring are going to take place. References Dach, R., Hugentobler,U., Friedz, P., Meindl, M. (2007, January). Bernese GPS software. Astronomical Institute, University of Bern Hongki, J., & Sung-Han, S., & Tatakowski, A., & Spencer, B.F., & Nelson, M. (2012). Feasability of displacement monitoring using low-cost GPS receivers. Structural Control Healt Monitoring, doi: 10.1002/stc.1532 Realini, E., & Reguzzoni, M. (2013). gogps: open source software for enhancing the accuracy for of low-cost receivers by single-frequency relative kinematic positioning. Measurement Science and Technology, 24 (11), 115010 Schwieger, V. (2009, September). High-sensitivity GPS: an availabilty, reliability and accuracy test. INCT Bulletin des Sciences Geographiques (No 23), Institut National de Cartographie e de Teledetection (INCT), Algier, Algerien. Takasu, T., & Yasuda, A. (2009). Development of the Low-cost RTK-GPS Receiver with an Open Source Program Package RTKLIB, Proceedings of International Symposium on GPS/GNSS. 435
436