CC BY
85
УДК 528:629.783 Джоел ван Кроненброк Leica Geosystems Швейцария
Email : joel. vancranenbroeck@leica-geosystems.com
УСПЕХИ В ОБЛАСТИ «ГНСС-КИНЕМАТИКА РЕАЛЬНОГО ВРЕМЕНИ» ДЛЯ ДЕФОРМАЦИОННОГО МОНИТОРИНГА СООРУЖЕНИЙ В РЕГИОНАХ ИЛИ В ПЕРИОДЫ ВЫСОКОЙ ИОНОСФЕРНОЙ АКТИВНОСТИ
Joël Van Cranenbroeck Leica Geosystems AG Switzerland
Email: joel.vancranenbroeck@leica-geosystems.com
ADVANCES IN GNSS-RTK FOR STRUCTURAL DEFORMATION MONITORING IN REGIONS OR PERIODS OF HIGH IONOSPHERIC ACTIVITY
Key words: GLONASS, GPS, Centralized Processing, RTK, GNSS Network corrections.
BACKGROUND
To achieve high GNSS-RTK accuracy in monitoring projects, the baselines between the reference station(s) and the monitoring receivers are kept as shorter as possible. However, there is always the challenge that the reference station(s) could by inadvertently installed in the deformation area itself (in the case of volcanoes, dam or building monitoring), or the signals tracked by the GNSS antennas are obstructed by the structure itself (as is the case for bridges and buildings).
In this paper, a new deformation monitoring concept that uses GNSS Network RTK principles will be described. Test results from sea wall monitoring in Hong Kong will be presented to demonstrate the superior results of this implementation of the GNSS Network RTK service, using single-frequency GNSS receivers, in a region where the ionospheric refraction effects are large and unpredictable. A GNSS Network RTK service has been installed some years ago by the Hong Kong Lands Department - consisting of a network of dual-frequency GPS continuously operating reference stations (CORS) and associated network processing software. An idea developed by Chris Rizos (University New South Wales, Sydney Australia) about ten years ago to combine a GPS CORS network with several low cost, single-frequency receivers (GPS L1-only) in Indonesia for volcano monitoring has encouraged the execution of a pilot project in Hong Kong. A combination of a sub-set of the Hong
Kong GNSS CORS network (a so-called “cluster”) and single-frequency GPS receivers was used to evaluate how the ionospheric delay errors could be mitigated in an application of GNSS Network RTK for sea wall structural monitoring.
The first results are indeed very promising and the comparison between singlebase GNSS-RTK and the one obtained by using the GNSS Network RTK corrections suggests a new and improved approach for GNSS-RTK monitoring.
Single-base GNSS-RTK for Structural Monitoring
The standard mode of precise differential positioning is for one reference receiver to be located at a reference station whose coordinates are known, while the second receiver's coordinates are determined relative to this reference receiver. The use of carrier phase data in real-time, single baseline mode (one reference station and one rover or user receiver’s coordinates to be determined in a relative sense) - also known as”single-base” mode - is now commonplace. These systems are also referred to as RTK systems (“real-time-kinematic”), and make feasible the use of GLONASS and GPS/GNSS-RTK for many time-critical applications such as engineering surveying, GLONASS and GPS/GNSS-guided earthworks/excavations, machine control and structural monitoring applications. Over the last decade and a half the use of GPS (and now GLONASS and GNSS) for structural monitoring, of dams, bridges, buildings and other civil structures, has grown considerably (see Ogaja et al., 2007, for a recent review), and nowadays the GNSS-RTK technique is widely used around the world. Such systems output continuous streams of coordinate results (or time series). The dynamics of the structure typically defines the nature of the coordinate analysis. For example, if a structure vibrates or deflects due to wind or surface loading the time series analysis is conducted in the frequency domain (see, e.g., Li et al., 2007), otherwise standard geodetic deformation monitoring techniques based on advanced network least squares analysis are used (Ogaja et al., 2007).
However, the limitation of single-base RTK is the distance between reference receiver and the rover receiver due to distance-dependent biases such as orbit error, and ionospheric and tropospheric signal refraction. This has restricted the interreceiver distance to 10km or less (depending upon the latitude). However in low latitude regions the ionospheric variability is so high that the use of single-base GNSS-RTK over short baselines may not even be possible in the local afternoon period. Yet in the case of short baselines there is a risk that the GNSS reference station could be located in the area that is subject to deformation. Furthermore, the object itself may cause signal refraction and in the worst case reduced signal availability impacting on the geometry (GDOP) of the solution. Some of these problems can be addressed using techniques where the GNSS reference station(s) are located as far away from the monitoring points as possible without being influenced too greatly by residual (double-differenced) atmospheric biases.
GNSS Network-RTK (N-RTK) is a centimetre-accuracy, real-time, carrier phase-based positioning technique capable of operating over inter-receiver distances
up to many tens of kilometres (the distance between a rover and the closest reference station receiver) with equivalent performance to single-base RTK systems (operating over much shorter baselines). The reference stations must be deployed in a dense enough pattern to model distance-dependent errors to such an accuracy that residual double-differenced carrier phase observable errors can be ignored in the context of rapid ambiguity resolution. N-RTK is therefore the logical outcome of the continuous search for a GPS/GNSS positioning technique that challenges the current constraints of single-base, cm-accuracy, high productivity, carrier phase-based positioning.
GNSS N-RTK for Structural Monitoring
All GNSS-based positioning techniques operate under a set of constraints (Rizos, 2002). These constraints may be baseline length, attainable accuracy, assured reliability, geometrical strength, signal availability, time-to-solution, instrumentation, operational modes, cost, and so on. GNSS product designers must develop systems (comprising hardware, software and field procedures) that are optimised for a certain target market, by addressing only those constraints that are crucial to the most common user scenarios. For example, single-base RTK systems are capable of high performance when measured in terms of accuracy, time-to-solution (i.e. speed of ambiguity resolution after signal interruption), utility (due to the generation of realtime solutions), flexibility (being able to be used in static and kinematic applications), ease-of-use, autonomy (operate their own reference station), and cost-effectiveness. However, the 10km baseline (or less in low latitude regions) constraint, the increasing availability of GNSS CORS networks (no need to operate their own), and the desire to use lower cost (i.e. single-frequency) user receiver hardware means that engineers are looking to alternative, more efficient GNSS-RTK based techniques for structural monitoring applications. See Rizos (2007) for discussion on the impact of the worldwide expansion of CORS networks on high accuracy GNSS users.
GNSS N-RTK is a technique that takes advantage of a network of permanently installed CORS streaming in real-time their raw observations to a central computing facility. Due to the more sophisticated modeling of residual spatially-correlated biases (due to atmospheric refraction of GNSS signals), the distances between CORS stations may be relaxed to many tens of kilometers (well beyond the baseline constraints of single-base GNSS-RTK). Hence the economics of operating CORS networks is significantly improved when N-RTK services are provided (Rizos & Cranenbroeck, 2006). There are a number of implementations of N-RTK - the commonest are VRS (Virtual Reference Station) and FKP (in German Flächenkorrekturparameter) - which involve processing of CORS network data in order to generate empirical ‘correction’ data (to principally account for the unmodelled residual double-differenced atmospheric biases) that are transmitted to users in RTCM-type messages (see, e.g. Heo et al., 2009). Alternative modes such as MAC (Master-Auxiliary Concept) place some of the burden of N-RTK processing on the user’s receiver (Janssen, 2009).
In the case of structural monitoring the CORS need to be located on very stable sites and the coordinates of the antenna phase centre for each CORS are determined with a relative accuracy of few millimeters. If monitoring over a long period of time, due regard must also be taken of reference frame stability -something beyond the scope of this paper. Another distinguishing characteristic of most structural monitoring applications is that the continuous streams of 3D coordinates are needed at a monitoring centre, not at the monitoring sites’ receivers. Hence other modes of N-RTK, such as reverse or server side N-RTK may be more appropriate (Lim & Rizos, 2008; Rizos, 2007). In the pilot project described in this paper the monitoring receivers deployed over the deformation area stream their observations to a PC server running the GNSS N-RTK modeling software, to generate N-RTK corrections in real-time, and integrated as Master Auxiliary or Virtual Reference stations in the processing.
Because of the advantages that a GNSS N-RTK approach to structural monitoring can offer, there is increasing worldwide interest in positioning infrastructure - national and regional CORS networks. Then the focus of the monitoring project becomes the deployment of a sufficient number of receivers at monitoring points to derive detailed enough deformation signals to help structural engineers determine whether the structure is responding to loads within design specifications, or whether the structure has suffered serious damage. While the use of installed CORS infrastructure makes GNSS more attractive for structural monitoring, the high cost of dual-frequency receivers is still a constraint to a massive expansion in the number of monitoring receivers on a project. How can the cost of GNSS monitoring be driven down even further? The use of low cost GPS L1-only monitoring receivers was proposed by several of the first author’s (Chris Rizos) graduate students almost ten years ago, see, e.g., Chen (2001), Roberts (2002).
Mixed-mode GNSS for Structural Monitoring
Data from single-frequency GNSS receivers cannot be corrected for ionospheric delay, as is the case with dual-frequency data. Therefore a combination of single- and dual-frequency instrumentation in a mixed-mode network could, in principle, ensure high accuracy coordinate results using a large number of receivers deployed across a region experiencing deformation, while keeping GNSS hardware costs as low as possible. Such an approach was used to develop a monitoring system for Indonesian volcanoes, as described in Janssen & Rizos (2003), and papers cited therein. This is possible by augmenting the single-frequency receivers with a small number of dualfrequency receivers surrounding the zone of deformation. The primary function of this outer network is to generate empirical ‘correction’ terms to the doubledifferenced phase observables within the deformation monitoring network. This mixed-mode methodology can in fact be implemented in real-time using a cluster of CORS - that have their raw observations processed using GNSS N-RTK software to generate 2D spatial models for residual tropospheric and ionospheric biases. These models then can be used to correct the monitoring receivers’ single-frequency data, either by software in the monitoring receiver itself (standard N-RTK approach) or at
the monitor centre if the reverse N-RTK approach is used (Rizos, 2007; Rizos & Cranenbroeck, 2006).
PILOT PROJECT IN HONG KONG
For demonstrating the benefits of the GNSS technology described in section 1 for monitoring sea walls a pilot project in Hong Kong was setup by an engineering company (Fig. 1). They used equipment and software developed and delivered by Leica Geosystems.
Figure 1: GPS monitoring antenna installed on a sea wall in Hong Kong.
The Leica GPS GMX902 dual-frequency monitoring receiver and the Leica GNSS AX1202 antenna have been installed with power supply and the communication equipment in an all-weather instrument cabinet (Fig. 2).
Figure 2: GPS receiver Leica GMX902 with power supply and communication
interface.
The CORS data were provided as a service by the Hong Kong Lands Department, by their Hong Kong Satellite Positioning Reference Station Network (Fig. 3).
Figure 3: The Hong Kong Satellite Positioning Reference Station Network.
The processing of the real-time data was carried out by the centralised RTK processing software, Leica GNSS Spider Positioning. Initially this was performed using the single-base GNSS-RTK approach (Fig. 4).
Figure 4: GNSS single-base RTK system architecture
It appeared, however, quite clearly that in the local afternoon period, but also randomly sometimes throughout the day, sudden jumps occurred in the coordinate
time series (Easting, Northing and Height), making reliable structural monitoring difficult or even impossible. It should be noted that the standard noise-like variation in GNSS-RTK time series are something users have to cope with via some form of smoothing and filtering. However, in this case the outliers were due to large biases resulting from extreme and highly variable ionospheric conditions.
After a meeting with the engineering company, who reported similar phenomena in other monitoring projects in Hong Kong, the authors approached the Hong Kong Lands Department and requested their assistance in delivering the real-time data streams of several CORS located in and around the monitoring project. Leica GNSS SPIDER NET software was installed with the necessary options to process the CORS network cluster and to redirect the RTK network corrections as observations for one of the closest reference station located nearby the monitoring receivers (see Fig. 5). That reference station acted as a MAX station (RTCM v3.n Master-Auxiliary Concept, via Master-Auxiliary Corrections, or MAX - Heo et al., 2009) in the Leica Spider site server. It should be emphasized however that any other reference station participating to that cluster could have been selected without affecting the results.
Figure 5: GNSS Network-RTK system architecture
Comparison was made between the “standard” single-base GNSS-RTK solution (Figure 4) and the GNSS N-RTK solution (Figure 5), in 2D and in height, using the Leica GNSS QC software. The plots (Figures 6 & 7) show clearly that the GNSS N-RTK corrections dramatically improved the results. Note that the 2D and height plot scales are different - each horizontal line is 1cm and 5cm respectively. Apart from
occasional outliers, the variability is less than 2cm for horizontal components and less than 5 cm for the height component. The results presented here are not filtered or smoothed, as they were just the output in the NMEA format of the baselines solutions computed by the Leica GNSS Spider Site server - positioning option. (A recursive low band pass filter - Exponential Weighted Moving Average - could now be applied on these unbiased N-RTK generated results in order to deliver a few millimeters accuracy, and in real-time.)
At the same time, in order to verify whether single-frequency GPS L1-only receivers could also benefit from the GNSS N-RTK, it was decided to simultaneously process the different baselines using only the GPS L1 frequency observations. The “Quasi-Static” method of initialization was used to perform fast L1 RTK. The results were even more impressive in terms of initialization (ambiguity resolution) and accuracy (Figures 8 & 9). Note that the 2D and height plot scales are different - each horizontal line is 1cm and 5cm respectively. Apart from occasional outliers, the quality of the time series is very similar to that in Figure 6 & 7, computed using higher cost dual-frequency monitoring receivers, i.e. the variability is less than 2cm for horizontal components and less than 5cm for the height component.
Figure 6: Comparison of RTK positioning 2D results -the blue line is the single-base RTK solution and the yellow line is the N-RTK solution (each horizontal line is 1cm).
Figure 7: Comparison of RTK positioning height results - the blue line is the singlebase RTK solution and the yellow line is the N-RTK solution (each horizontal line is
5cm).
Figure 8: Comparison of RTK positioning 2D results - the green line is the singlebase RTK L1-only solution and the red line is the N-RTK L1-only solution (each
horizontal line is 1cm).
Figure 9: Comparison of RTK positioning height results - the green line is the singlebase RTK L1-only solution and the red line is the N-RTK L1-only solution (each
horizontal line is 5cm).
CONCLUDING REMARKS
The results presented here of a sea wall monitoring project in Hong Kong demonstrate that the combination of GPS Network RTK resources (CORS network and Network RTK modeling software) delivers outstanding advantages, such as maximum (unbiased) accuracy and reliability. The following summary comments can be made with regard to the GNSS N-RTK based monitoring technique:
- Better control over the operations and the results by taking advantage of installed CORS infrastructure.
- Reliable time series solutions for projects located in low latitude regions where the ionospheric turbulences severely affect signal and data processing.
- The possibility to mix dual-frequency receivers (GNSS CORS) with affordable single-frequency receivers for slow deformation motion monitoring.
- No need for subsequent networked baselines adjustment.
- No need to establish single CORS in urban areas (obstructions) for high rise building or long bridge monitoring projects.
Although implemented for a trial in Hong Kong, the authors believe that with the return of high solar cycle activity the proposed mixed-mode solution strategy could find application in many other places than only those currently exposed to severe ionospheric disturbances (i.e. low latitude regions).
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to Simon Kwok, director of the Geodetic Department of the Hong Kong Lands Department Administration, for his support during this trial.
REFERENCES
CHEN, H.Y. (2001), A Study on Real-Time Medium-Range Carrier Phase-Based GPS Multiple Reference Stations, UNISURV S-64, School of Surveying & Spatial Information Systems, The University of New South Wales, Sydney, Australia, 182pp, ISBN 0 7334 1872 2.
ROBERTS, C.A. (2002), A Continuous Low-Cost GPS-Based Volcano Deformation Monitoring System in Indonesia, UNISURV S-73, School of Surveying & Spatial Information Systems, The University of New South Wales, Sydney, Australia, 271pp, ISBN 0 7334 1976 3.
HEO, Y., YAN, T., LIM, S., & RIZOS, C. (2009), International standard GNSS real-time data formats and protocols, IGNSS Symp. 2009, Gold Coast, Australia, 1-3 December, CD-ROM procs.
JANSSEN, V. (2009), New ways to Network RTK, Position, 41, June-July issue. JANSSEN, V., & RIZOS, C. (2003), Processing mixed-mode GPS network processing approach for deformation monitoring applications, Survey Review, 37(287), 2-19.
LI, X., RIZOS, C., GE, L., & AMBIKAIRAJAH, E. (2007), Application of 3D time-frequency analysis in monitoring full-scale structural response, Journal of Geospatial Engineering, 8(1-2), 41-51.
LIM, S., & RIZOS, C. (2008), A conceptual framework for server-based GNSS operations, Journal of GPS, 7(2), 35-42.
OGAJA, C., LI, X., & RIZOS, C. (2007), Advances in structural monitoring with Global Positioning System technology: 1997-2006, Journal of Applied Geodesy, 1(3), 171-179.
RIZOS, C. (2002), Network RTK research and implementation: A geodetic perspective, Journal of Global Positioning Systems, 1(2), 144-150.
RIZOS, C. (2007), Alternatives to current GpS-RTK services & some implications for CORS infrastructure and operations, GPS Solutions, 11(3), 151-158.
RIZOS, C., & CRANENBROECK, J.van (2006), Alternatives to current GPS-RTK services, 19th Int. Tech. Meeting of the Satellite Division of the U.S. Inst. of Navigation, Fort Worth, Texas, 26-29 September, 1219-1225.
BIOGRAPHICAL NOTES
Joël van Cranenbroeck is currently International Project Director for Geodetic Monitoring Services at Leica Geosystems AG, Heerbrugg, Switzerland. He has led the development of hardware and software solutions for GNSS Network-RTK since 2001 and has made significant contributions in geodetic monitoring development and applications such as the method for aligning high rise structures (such as the Burj Dubai). Joel is Chair of Working Group 6.2 in FIG Commission 6, was awarded in 2009 the title of Honored Lecturer of the Siberian State Academy of Geodesy in
Novosibirsk, and is senior scientist consultant for two universities in Belgium. He has designed numerous projects for structural monitoring applications such as bridges, dams, tunnels, etc. He worked at the Belgian Cadastre organization, at the Geodetic Department of the Belgian National Geographical Institute and in Star Informatic - a GIS software based Belgian company - before becoming the Leica Geosystems representative in Belgium in 1993.
CONTACTS
Joël van Cranenbroeck
International Project Director for Geodetic Monitoring Services Leica Geosystems AG Switzerland Heinrich Wild Strasse,
CH-9435 Heerbrugg SWITZERLAND Tel. +32-474-98 61 93
Email: joel.vancranenbroeck@leica-geosystems.com Web site: www.leica-geosystems.com
© J. Van Cranenbroeck, 2011
