Standards and specifications

1
Standards and specifications for GPS Control
Surveys
2
Overview

• Standards for Control Surveys
• Class, Order, Position Local Uncertainty (GPS)
• Geodetic Infrastructure
• Victorias Control Network
• Legal Traceability
• Best Practice
• GPS Data Quality and Integrity
• Example - RTK Control GPSnet Mildura

3
Introduction
What are standards? International Standards
Organisation ISO in the contexts of The US
National Standard for Spatial Data Accuracy

• Standards are documented agreements
• containing technical specifications or other
  precise criteria
• to be used consistently as rules, guidelines, or
  definitions of characteristics
• to ensure that materials, products, processes, or
  services are fit for their purposes.

4
Standards for Control Surveys in Australia

• Responsibility for standards is shared between
  Commonwealth and States
• High level cooperation between various agencies
  since 1950
• ICSM Inter-Government Committee on Surveying and
  Mapping facilitates action
• Geodesy Technical sub-committee activities
  include
• - Standards and Practices for Control
  Surveys (SP1)
• - Legal Traceability of GPS Measurements

5
Standards and Practices For Control Surveys (SP1)

• http//www.anzlic.org.au/icsm/publications/sp1/sp1
  .htm
• Ver 1.5 May 2002
• Incorporates concept of Positional and Local
  Uncertainty
• Part A - Standards of Accuracy
• Standards of Accuracy for Control Surveys - to
  achieve uniform standards for all national and
  state control networks
• Provides a useful reference for anyone to quote
  appropriate quality control standards
• Defines the concept of Class, Order, Positional
  Local Uncertainty

6
Standards and Practices For Control Surveys (SP1)

• http//www.anzlic.org.au/icsm/publications/sp1/sp1
  .htm
• Ver 1.5 May 2002
• Incorporates concept of Positional and Local
  Uncertainty
• Part B - Recommended Survey Reduction Practices
• Are a GUIDE only and are techniques to attain the
  precision needed for accuracy standards from Part
  A
• These practices are not exhaustive nor considered
  mandatory
• Choice of technique is a professional decision -
  match methodology to results needed to be
  achieved

7
Class

• Class of a survey is a means of categorising the
  internal quality or precision of a survey
• reflects suitability of network design, survey
  methods, instrumentation and computation
• verified by a minimally constrained adjustment
  (one station fixed)
• horizontal - relative error ellipses are compared
  to various categories of Class

8
Order

• Order of a survey is a means of categorising the
  external quality or accuracy of a survey
• Order is the conformity of new survey data within
  an existing network coordinate set
• verified by a fully constrained adjustment (fix
  the surrounding geodetic control)
• Includes the precision of any transformation from
  one datum to another

9
Order

• Good geodetic GPS results can be 1ppm
• even as good as 1ppb! (1 part per billion plate
  tectonics)
• Hence a very high quality GPS (high Class) survey
  may have to be distorted to fit existing
  control which may have been determined using
  lower class survey -
• therefore the resulting Order will have to match
  the lower Class of existing control

10
Positional Uncertainty

• Is the uncertainty of the horizontal coordinates
  or height of a point in metres at 95 confidence
  level with respect to the defined reference frame
  
• Can be thought of as Network or Absolute
  uncertainly
• Compatible with ISO standards

11
Local Uncertainty

• Is the average measure in metres at 95
  confidence level with respect to the adjacent
  points in the defined reference frame
• Can be thought of as relative uncertainly
• Compatible with ISO standards

12
Position and Local Uncertainty

• Are new and simple methods for classifying
  quality of positions
• replaces Order - phased out 2005
• Class will stay - continues to classify all
  aspects of a survey network
• Note Class and Order are at 1s
• Position and Local Uncertainty at 95 confidence

13
Determination of Class

• Determination of Class from the formula
• rc(d0.2)
• Where
• r maximum allowable semi-major axis in mm
• c is empirically derived factor by historical
  accepted precision
• d distance to any station in km
• Class is allocated by assessing the length of the
  maximum allowable semi-major axis r from a
  minimally constrained adjustment on the
  associated datum

14
Classification of Horizontal Surveys
Values of C assigned to various Classes of survey
rc(d0.2)
15
Error statistics
16
Assigning Order to a Horizontal Survey

• Stations in a horizontal survey are assigned an
  Order commensurate with the Class of the survey
  and the conformity of the data within that data
  set
• Order is based upon the Class as well as the fit
  of the network to the existing coordinate system
  through a constrained adjustment
• The Order assigned following this adjustment may
• 1) NOT be higher than the Order of existing
  stations
• 2) NOT be higher than the Class assigned to that
  survey

17
Assigning Order to a Horizontal Survey
State Geodetic
Coordination
SMES
18
Class for Vertical Surveys

• Vertical control surveys are assigned a class
  according to the planned and achieved precision
  is a function of
• Network design
• Survey practices
• Equipment
• Processing techniques
• Datum for heights is AHD

19
Class for Vertical Surveys
Where r maximum allowable
error, in mm. c an empirically derived factor
for each particular CLASS of survey result. d
distance to any station in km. The values r
for GPS/trig heighting are considered to be
standard deviations. The values r for
differential levelling are considered to be
standard deviations on the
condition that at least one forerun and one
backrun agree within the c? d limits.
20
Class for Vertical Surveys
21
Order for Vertical Surveys
ORDER of heights from a survey is allocated on
the basis of the fit of that survey to existing
(constraining) heights. This technique is
identical to that employed in the determination
of CLASS and makes use of the same
formulae. Note for GPS heighting, Order is
assigned according to the accuracy of the
geoid-ellipsoid separation.
22
Positional Uncertainty

• Indicator of quality of a position in metres
• Uncertainty of a coordinate or height at 95
  confidence with respect to a defined
  reference frame
• The reference frame MUST described in the
  Metadata
• For horizontal is GDA94 and vertical is AHD
• Positional Uncertainty is reported as the total
  uncertainty propagated from the Zero Order
  Control AFN and AHD tide gauge bench marks

23
Positional Uncertainty

• Positional Uncertainty can report the quality of
  a position that is independent of the local
  control
• This is an important consideration when comparing
  positions from different sources and for
  positions not directly connected to the survey
  network
• For example
• 1) GPS processing results from Geoscience
  Australias on-line processing service AUSPOS
• 2) Wide area Differential GPS services Omnistar,
  Starfix (WGS84) and AMSA (Cape Schank)

24
Local Uncertainty

• Local Uncertainty is the average measure, at 95
  confidence, of the relative uncertainty of a
  point(s), with respect to adjacent points in the
  defined frame
• Is calculated between two points in question or
  from the point in question to adjacent points in
  the network
• Is similar to Order and replaces it (Order may
  still be used until 2005

25
Local Uncertainty

• Is the radius of a 95 circle calculated from the
  standard error ellipse produced by most least
  squares software
• or calculated from standard deviation in the
  case of height

C b/a
(Leenhouts, 1985). K q0 q1C
q2C2 q3C3 Radius aK Where a semi-major
axis of the standard error ellipse b semi-minor
axis of the standard error ellipse. q0
1.960790 q1 0.004071 q2 0.114276 q3
0.371625
26
Australian Regional GPS NetworkARGN
Geodetic Infrastructure
ARGN forms part of the International GPS Service
IGS Monitors Global Change (ITRF)
27
Australian Fiducial Network (AFN)Geoscience
Australia
Geodetic Infrastructure
AFN forms the framework for GDA Zero Order
Control
28
Australian National Network ANNANN is 78
stations at 500km spacings constrained to AFN
Accuracy is 0.1 ppm
Geodetic Infrastructure
29
Australian Height Datum AHD (1971)
Geodetic Infrastructure
Adopted from 97,230 kilometres of two way
levelling and fixing mean sea level of thirty
tide gauges across Australia
30
AUSGeoid AUSGeoid98 consists of approximately
3.6km grid of geoid-ellipsoid separations N
Values relative to the GRS80 ellipsoid,
Geodetic Infrastructure
AUSGeoid 98 has an absolute accuracy of 0.5m and
a relative accuracy of 1-5ppm
31
Victorias Horizontal Control Network 24,000
adjusted GDA
Geodetic Infrastructure
32
Level (Benchmark)network 40,000 3rd order
adjusted AHD
Geodetic Infrastructure
33
Combined 3D Network
Geodetic Infrastructure
Available from Survey Marks Enquiry Service SMES
34
Victorias Active Geodetic Network
GPSnet Base Stations 2nd Order Horizontal lt
0.01m 95 Local Uncertainty
35
Legal Traceability

• Legal Traceability is a formal process whereby
  all measurements must be related back to National
  Standards
• This process has been well defined for EDM etc,
  through the National Measurement Act (1960)
• The objective of this Act is to ensure that
  measurements are what they purport to be
• National Standards Commission have adopted the
  positions in the Australian Fiducial Network
  (AFN) as a recognised value standard
• The AFN is the mechanism by which GDA is
  connected to the International Terrestrial Frame
  (ITRF)

36
Legal Traceability

• In Victoria GPS surveys are traceable to the AFN
  via the Australian National Network (ANN)
• By 25,000 adjusted GDA marks in SMES
• Connection to GPSnet
• Ensuring GPS surveying measurements are what
  they purport to be is a combination of Best
  Practice and Traceability
• GPS should not be used as the sole method for
  measuring length in legal surveys within
  Australia

37
Best Practice Guides

• Concentrate on GPS surveying
• GPS guide in SP1 refers to relative GPS
  positioning two or more GPS receivers,
  observing Carrier Phase observations (CA code
  L1 phase)
• Specifically for GPS hardware and software
  designed for geodetic surveying applications
  operating in differential mode
• Many agencies recommend SP1 Guides -
  comprehensive and includes
• Geodetic Datums and Geoid Separations
• Equipment Validation
• Planning - Network Design and Geometry
• General Requirements for GPS Observations
• Specific Requirements - Classic, RTK and baseline
  processing

38
Geodetic Datums and Geoid Separations

• GDA (Geocentric Datum of Australia 1994) WGS 84
  is closely aligned to ITRF in 2001 differs to GDA
  94 by 0.5 metres due to tectonic movements
• All adjustments of GPS data should be 3D on the
  GRS80 ellipsoid - for practical purposes WGS84
  ellipsoid is equal to GRS80
• Horizontal measurements should form a closed
  figure connected to a minimum of 2 stations with
  a class/order appropriate to the survey
• Vertical datum is AHD
• GPS connections should be observed to bench marks
  (3 rd order) these with the geoid model AUSGeoid
  98 enable fitting to the vertical datum

39
Equipment Validation

• Zero baseline where two GPS receivers are
  connected to the same antenna - positions should
  show noise in the system
• Quimby - more about that later
• Field verification procedures - over high quality
  geodetic marks
• look at RTK verification in Mildura
• NSW - Cadastral Surveyors should be concerned if
  field differences are more than (against Order
  1, L3 Control)
• 15mm /- 3ppm for Coordinates
• 35mm /- 8ppm for Height

40
Planning - Network Design and Geometry

• Appropriate technique for required precision -
  good GPS should meet Class A Order 1 horizontal
  and L3 heights
• Location and distribution of points depend more
  on accessibility and redundancy than distribution
• All GPS surveys should be connected to state
  control (some manufacturers recommend users
  operate in absolute mode RTK eg for site surveys
  - everything is relative to that job only!)
• Planning to sufficiently reduce error budget
• Redundancy of observations - Class A network
• 20 stations are reoccupied 3
  times
• 100 stations are reoccupied 2 times

41
GPS Methods for different Class
Good GPS practice should achieve Class A Order
1 coordinates and L3 for Height
42
Planning - Network Design and Geometry
Closest control points are not directly observed
due to a natural feature
43
Planning - Independent baselines

• An independent baseline is achieved when the
  data used is not a simply different combinations
  of the same data - trivial triangles
• With n receivers the total number of base number
  of base lines computed is n(n-1)/2. However only
  n-1 of those baselines are independent
• Dependant baselines are highly correlated
• Trivial baselines introduce false redundancy in
  degrees of freedom

44
GPSnet Data Quality Integrity
45
GPSnet Quality and Integrity Conditions
GPS network quality conditions includes

• Scrutinising data quality, integrity and
  performance of each
• GPS base station site

• Ensuring data reliability of the GPSnet system
  and
• timely availability of data

• Monitoring the network shape and determining a
  the conformance
• level for each station within the network

46
Quimby(Quality integrity monitoring of base
stations)By Neil Brown Department of
GeomaticsUniversity of Melbourne
Quimby reports system performance (similar to
TEQC University Navstar Consortium)

• Data completeness (all available, L1, L2, C1, P2,
  D1, D2)

L1, L2 Phase measurements on L1 and L2 C1
Pseudorange using C/A-Code on L1 P1, P2
Pseudorange using P-Code on L1,L2 D1, D2 Doppler
frequency on L1 and L2

• Cycle Slips
• Data Gaps
• Multipath

47
Quimby Plots
Complete Observations Melbourne June 2002 lt 85
Complete gives an automatic e-mail warning to
Operations Manager
48
Quimby Plots
Cycle Slips Melbourne June 2002 What should it
be? Cycle slips for 1000/observations are less
than 5 for 50 NGS CORS and less than 10 for 75
of stations
49
Quimby Plots
Multipath on L1 (MP1) Melbourne June 2002 What
should it be? 50 of IGS stations have RMS MP1
under 0.4m, and 75 have less than 0.5m.
50
Quimby Plots
Multipath on L2 (MP2) Melbourne June 2002 What
should it be? RMS MP2 is less than 0.6m for 50
of IGS stations and less than 0.75m for 75 of
IGS stations
51
Quimby Plots
Additional Information Latency GDOP Epochs
Satellites
52
RTK Control Surveys
Case study GPS for Cadastral Control
Survey Standards and Guidelines for Cadastral
Surveys using GPS (http//www.blm.gov/nhp/efoia/wo
/fy01/im2001-186attach1.pdf) Local Accuracy
Standards at 95 Confidence Circle Less than
0.05m - Cadastral Project Control (static) Less
than 0.1m - Cadastral Measurements (RTK) Network
Accuracy Standards at 95 Confidence Circle Less
than 0.1m - Cadastral Project Control Link to
CORS Less than 0.2m - Cadastral Measurements
53
RTK Control Surveys

• Dual frequency or single frequency geodetic
  receivers can be used
• Dual frequency increases reliability of
  ambiguity resolution and range (cost more!)
• Typical range for RTK is 20km - But accuracy
  criteria limits this to 10km (0.1m at 95) or 5km
  (0.05 at 95)
• Typical precision claimed by manufactures is
  0.01 /- 2ppm
• (1 sigma horizontal)
• Ambiguities must be correctly resolved for all
  occupations
• Two independent occupations of all new stations
  are recommended minimum - occupy some stations
  three times or use a conventional check (EDM)
• Multipath is a significant error source for
  short occupations - reoccupy using a different
  constellation - minimum 45m after

54
GPSnet Mildura - RTK
RTK range tests Repeater ready sites
55
GPSnet Mildura - RTK
Baselines
Radio Repeater
New Control
56
Classic
GPSnet
Existing Control
New Control
57
RTK
GPSnet
Existing Control
New Control
58
RTK
GPSnet
New Base
Existing Control
New Control
New Base
59
RTK Control Recommended Procedure

• Two step process
• 1) Validate or create Project Control
• 2) Undertake RTK survey - re-occupy and do checks
• Validate against existing control -3rd Order GDA
  or better
• Measure to validation and new control points at
  least twice some three times
• Each measurement should be for at lease greater
  than 30 seconds (180 is a default ) DOPlt6 15 deg
  mask lt5 sats
• Force a re-initialise between measurements using
  different antenna heights
• This checks initialisation's for correct
  ambiguity resolution and randomises some errors
• Post- process real-time data (a specification
  for some large exploration companies) has
  advantages e.g., no delay - process forward and
  backward

60
RTK Control

• Require a Local Uncertainty of 0.05m and a
  Position Uncertainty of 0.1m (GDA connection to
  GPSnet)
• Rule of thumb take the inverse between the two
  measurements this should be good to better than
  0.02m
• Point tolerance can be set at 0.02m for
  recording measurements
• This achieves 0.05 (2.450.02 at 95 for East,
  North)
• Fit to control should be good to better than
  0.05 horizontal position vector
• Assuming standard error propagation the maximum
  difference for RTK measurements should not be
  bigger than 0.086m to achieve 0.1m at 95
  confidence

61
References

• GPSnet Information
• www.Land.vic.gov.au/geodesy
• SP1
• www.anzlic.org.au/icsm/publications/sp1/sp1.htm
• Standards and Guidelines for Cadastral Surveys
  using GPS
• www.blm.gov/nhp/efoia/wo/fy01/im2001-186attach1.pd
  f