UNAVCO Real-Time GNSS Positioning: High-Precision Static ... · 10−1 100 101 −60 −40 −20 0 20 40 60 log(mm 2 /Hz) Freq, Hz Stage RTX −60 −50 −40 −30 −20 −10 0

UNAVCO REAL-TIME RESOURCES:NETWORK STATUS:

http://www.unavco.org/instrumentation/networks/status/pbo/realtime

UNAVCO DATA POLICY:

http://www.unavco.org/community/policies_forms/data-policy/DataStreamingPolicy.pdf

NTRIP SOURCETABLE:

http://rtgpsout.unavco.org:2101

REQUEST REAL-TIME ACCESS:http://www.unavco.org/data/data-help/custom-data-request/custom-data-request.html

UNAVCO is funded by the National Science Foundation (NSF) and National Aeronautics and Space Administrations (NASA).

The GAGE facility, managed by UNAVCO, operates a real-time GNSS (RT-GNSS) network of ~450 GNSS stations. The network includes stations that are part of the EarthScope Plate Boundary Observatory (PBO), the Continuously Operating Caribbean GPS/GNSS Observational Network (COCONet), and the Trans-boundary, Land and Atmosphere Long-term Observational and Collaborative Network (TLALOCNet). Following community input from a real-time GNSS data products and formats meeting hosted by UNAVCO in Spring of 2011, UNAVCO now provides real-time PPP positions, and network solutions where practical, for all available stations using Trimble’s PIVOT RTX server software.

The UNAVCO real-time system has the potential to enhance our understanding of earthquakes, seismic wave propagation, volcanic eruptions, magmatic intrusions, movement of ice, landslides, and the dynamics of the atmosphere. Beyond the ever increasing applications in science and engineering, RT-GNSS has the potential to provide early warning of hazards to emergency managers, utilities, �rst responders and others. Recent upgrades to the network include eight Trimble NetR9 GNSS receivers with GLONASS and receiver-based RTX capabilities and ten new collocated MEMS based accelerometers. These new capabilities will allow integration of GNSS and strong motion data to produce broad-spectrum waveforms improving Earthquake Early Warning systems.

The UNAVCO Real-Time Network

A map of North America showing UNAVCO’s current real-time GNSS network of ~450 stations. The network includes stations that are part of the PBO, COCONet, and TLALOCNet projects.

−170˚

−160˚

−150˚−140˚

−130˚ −120˚ −110˚ −100˚−90˚

−80˚

−70˚

−60˚

10˚

20˚

30˚

40˚

50˚

60˚

70˚

There are approximately 129 Real-time GNSS sites with collocated meteorological instruments. The data from these instruments are made available in near-real-time via BINEX streams, and as RINEX �les downloadable from our archive.

−170˚

−160˚

−150˚−140˚

−130˚ −120˚ −110˚ −100˚−90˚

−80˚

−70˚

−60˚

30˚

40˚

50˚

60˚

70˚

−119˚00' −118˚30' −118˚00' −117˚30' −117˚00' −116˚30' −116˚00' −115˚30' −115˚00' −114˚30' −114˚00'

32˚30'

33˚00'

33˚30'

34˚00'

34˚30'

GLRS

P472

P482

P483

P484

P486

P494

P505

P506

SLMS

USGC

P797

A map of Southern California showing ten sites with recently installed Micro-Electro-Mechanical Systems (MEMS) based accelerometers.

Plot of the number of users requesting access to UNAVCO's real time data streams. Non-pro�t consortia include entities such as NOAA or UCAR.

Controlled outdoor kinematic and static experiments provide a useful method for evaluating and comparing real-time systems. At UNAVCO, we have developed a portable antenna actuator to characterize kinematic performance of receiver- and server-based real-time positioning algorithms and identify system limitations. We have performed initial tests using controlled 1-d antenna motions and present recent test results in this poster.

GNSS instruments are being used to record large displacements and accelerations from earthquakes that exceed M7. A few researchers have conducted tests to independently characterize GNSS instrument response over seismic frequencies [e.g. Ebinuma and Kato, 2012; Elosegui et al., 2006; Wang et al., 2012; Langbein et al., 2012]. Previous work has shown that receiver response is not frequency independent for signals greater than 1 Hz, and receiver response varies with receiver make and model. In addition, user selectable tracking parameters may a�ect a receiver’s response. In this paper, we show the response of Trimble’s receiver-based RTX real-time positioning service to frequencies from 0.2-6.0 Hz. UNAVCO will continue evaluations of new commercial and open-source real-time positioning services as they become available.

In addition to kinematic testing, long-term static testing of Trimble’s RTX service is ongoing at UNAVCO and will be used to characterize the stability of the position time-series produced by RTX.

Outdoor Kinematic Testing

Above: UNAVCO’s low-cost antenna actuator. The antenna carriage rides on plastic linear bearings. We employ a 500 mm double-action pneumatic cylinder with position feedback to drive the system. A servo pneumatic proportional control valve provides position control using a 0-10 V command signal. The carriage position is logged at 200 Hz using a Kinemetrics Basalt data recorder. Compressed air is provided by a portable consumer-grade air compressor.

Left: The magnitude of the di�erence between the RTX measured position and true stage po-sition peaks near 3 Hz. Right: The displacement spectrum of the stage (blue) and RTX (red).

10−1 100 101

−60

−40

−20

0

20

40

60

log(

mm

2 /Hz)

Freq, Hz

StageRTX

19:41:25 19:41:30 19:41:35 19:41:40 19:41:45 19:41:50−100

−50

0

50

100

Stage Position [200sps]

[mm

]

19:41:25 19:41:30 19:41:35 19:41:40 19:41:45 19:41:50−100

−50

0

50

100

RTX Position [50sps]

[mm

]

19:41:25 19:41:30 19:41:35 19:41:40 19:41:45 19:41:50−100

−50

0

50

100

Difference RTX and Stage Position

[mm

]

3.0 Hz Sinusoidal shaking (RTX)

2.0 Hz Sinusoidal shaking (RTX)

19:37:00 19:37:15−200

−100

0

100

200Stage Position [200sps]

[mm

]

19:37:00 19:37:15−200

−100

0

100

200RTX Position [50sps]

[mm

]

19:36:50 19:36:55 19:37:00 19:37:05 19:37:10 19:37:15 19:37:20−200

−100

0

100

200Difference RTX and Stage Position

[mm

]

Left: Increasing di�erence between the RTX measured position and true stage position at 2 Hz. Right: The displacement spectrum of the stage (blue) and RTX (red).

10−1 100 101

−60

−40

−20

0

20

40

60

log(

mm

2 /Hz)

Freq, Hz

StageRTX

1.0 Hz Sinusoidal shaking (RTX)

19:35:00 19:35:10 19:35:20−200

−100

0

100

200Stage Position [200sps]

[mm

]

19:35:00 19:35:10 19:35:20−200

−100

0

100

200RTX Position [50sps]

[mm

]

19:35:00 19:35:10 19:35:20−200

−100

0

100

200Difference RTX and Stage Position

[mm

]

Left: Small di�erences between RTX measured position and true stage position at 1 Hz. Right: The displacement spectrum of the stage (blue) and RTX (red).

10−1 100 101

−60

−40

−20

0

20

40

60

log(

mm

2 /Hz)

Freq, Hz

StageRTX

−60

−50

−40

−30

−20

−10

0

10

Phas

e, d

eg.

0.1 1.0 10Frequency, Hz

We also observe a frequency-dependent re-sponse in the phase of the estimated displace-ment of the antenna while undergoing sinusoi-dal shaking.

0.1 1.0 100

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Frequency, Hz

Gai

n

Similar to the post-processed results of Ebinuma and Kato (2012) and Langbien et al. (2012) we observe that RTX has a frequency-dependent response when the antenna undergoes sinusoi-dal motion.

RTX Testing Results:

Frequency [Hz]

Stage Amplitude [mm]

RTX Amplitude [mm]

Phase Difference [deg]

0.2 156.7 155.14 0.770.2 157.46 156.6 0.8970.5 157.51 158.81 1.2110.5 157.82 159.6 1.2661 159.41 162.24 2.852 104.37 119.6 11.3953 50.183 82.809 5.854 31.46 51.933 -7.555 41.631 64.086 -21.4236 39.22 58.455 -29.8

Test Con�guration:

Receiver Type: NetR9

Antenna Type: Trimble Zephyr

Position Solution: RTX receiver based PPP service

Receiver Motion: Kinematic (not �ltered)

RTK Engine Motion: Kinematic

Constellations: GPS + GLONASS

Sinusoidal Test Duration: 30 seconds

An example of carriage displacement error while reproducing a 1 Hz sinusoid. We will be adding an accumulation tank to the system to increase positioning accuracy.

19:35:00 19:35:05 19:35:10−200

−100

0

100

200

[mm

]

19:35:00 19:35:05 19:35:10−20

−10

0

10

20

[mm

]

Error

StageSine

Test Apparatus

UNAVCO Real-Time GNSS Positioning: High-Precision Static and Kinematic Testing of the Next Generation GNSS Network

AGU FALL MEETING 2014

Henry Berglund (berglund@unavco.org), Kathleen Hodgkinson, Frederick Blume, Dave Mencin, David Phillips, Chris Walls, Doerte Mann, Christine Puskas, Otina Fox, Chuck Meertens and Glen Mattioli

UNAVCO, Boulder, CO 80301-5553

Left: A schematic showing UNAVCO’s real-time data �ow. BINEX streams from telemetered stations are col-lected by primary and sec-ondary servers in the PIVOT system. At this point the BINEX is sent to the archive for permanent storage and to the primary and second-ary castors for streaming to the public. The NTRIP castors distribute RTCM, BINEX, and position streams to regis-tered users. RINEX is also generated and available for download from the archive via FTP.

Right: P484 in Southern California. A MEMS accelerometer is clamped to the vertical leg of the monument. The instrument is connected to the Geodetic Module (below) which is mounted inside the GPS enclosure. The Geodetic Module combines GPS and seismic data on location.

References:

Ebinuma, T. and T. Kato, Characteristics of Very-High-Rate GPS for Seismology, Earth Planets and Space, 64, 2012.

Elósegui, P., J. L. Davis, D. Oberlander, R. Baena, and G. Ekström. "Accuracy of high‐rate GPS for seismology." Geophysi-cal research letters 33, no. 11 (2006).

Langbein, J. O., J. R. Evans, F. Blume, and I. A. Johanson. "Tests of GNSS receivers for dynamic, high sample rate response using controlled sources of displacement, velocity, and acceleration." In AGU Fall Meeting Abstracts, vol. 1, p. 1122. 2012.

Wang, G., F. Blume, C. Meertens, P. Ibanez, and M. Schulze. "Performance of high-rate kinematic GPS during strong shaking: observations from shake table tests and the 2010 Chile earthquake." Journal of Geodetic Science 2, no. 1 (2012): 15-30.

Summary:Our preliminary testing indicates that the re-sponse of RTX is not frequency independent and care must be taken with interpreting ob-servations that contain frequencies greater than 1Hz. Our future testing will compare RTX performance with and without GLONASS. As new commercial and open-source real-time positioning services become available, they will be included in our testing. In addtion we will evaluate the response of receivers to shak-ing, as all testing to date has been conducted with only the antenna undergoing motion.

Latency and completeness of the MEMS sites in Anza. The �gure shows the latency at 1 hour interval for the 1-sps RTGPS data. The table below the plot summarizes thecompleteness of the data at 1 hour intervals and the latencies using both the median and the mean.

Summary of the latencies across the network. 1-hour latencies were extracted for all PBO MEMs sites between 11/1/2014 and 11/24/2014 from the RTGPS database. The sites were sorted by telemetry type. O�ine sites were excluded. There are two side peaks on the CDMA trace, one at ~75 ms and another at 500 ms. This may indicate a regional network issue.

Power spectral density plots for the MEMs sites. This analysis was done with data col-lected from Nov. 1-2. The vertical component is much noisier, as expected. All the site have similar noise characteristics. However, P494 and P505 are quieter in the 0.1 to 0.01 Hz band. The calculations were made using a 1 hour, 50% overlapping window, with a Welch estimator.

0

100

200

300

A J O J A J O J A J O J A J O J A J O2010 2011 2012 2013 2014

Num

ber O

f Reg

iste

red

Use

rs

Government and Non-Profit ConsortiaCommercial

TotalAcademia

Number of Users Registered for UNAVCO RTGPS Data Streams

UNAVCO Real-Time GNSS Network Real-Time Meteorological Data

Real-Time Sites With MEMSWho’s Using The Data?

Examples Of Network Performance

Real time-EnabledGPS + Met Network

PIVOT SYSTEM

HIGH-RATE DATA STREAM

HIGH-RATE DATA FILE

DataInput

DataOutput

BINEX

Primary CasterMars

Secondary CasterTheia

Primary StreamControl + Data

Processing

Secondary StreamControl + Data

Processing

BINEX

Load BalancerPermanent

GPSArchive

RINEX

SurveyUsers

ResearchUsers

RTCMBINEX,

Position

0

5000

10000

15000

0 200 400 600 800 1000

CDMA 243 173

DIRECT CONNECTION 29 107

DSL WISP 4 66

RADIO 122 239

VSAT 9 866

Communications Sites Median (ms)

Num

ber O

f Obs

erva

tions

Latency versus Communications

Latency (milliseconds)

0

10

20

30

0 200 400 600 800 1000Latency (milliseconds)

Perc

enta

ge

B) Latency Normalized by number of Observations

A) Number of Observations versus Latency

1

10

100

1000

0.001 0.01 0. 1

East

P484P494P505P506P797

Frequency (Hz)

1

10

100

1000

0.001 0.01 0. 1

North

P484P494P505P506P797

Frequency (Hz)

1

10

100

1000

10000

0.001 0.01 0. 1

Up

P484P494P505P506P797

mm

2 /Hz

Frequency (Hz)

mm

2 /Hz

mm

2 /Hz

P484 100 99.9 118 176.0

P494 100 99.2 5 20.4

P505 100 95.3 46 264.7

P506 100 98.4 1 49.4

P797 100 99.8 70 81.4

50

100

150

200

20 27 03 10 17

Completeness and Latency of RTGPS + MEMs Sites

Completeness(%) Latency (ms)GPS Median Mean Median Mean

Com

plet

enes

s %

0

October 2014 November 2014