GPS Data Quality: Integrity and Positional Accuracy

1
GPS Data Quality Integrity and Positional
Accuracy

• Candidate Neil Brown
• Supervisors Dr Allison Kealy
• Dr Phil Collier

2
Overview

• Research problem, progress and outcomes
• VSIS recommendations

3
Positional Quality and SDI
4
GPS Data Quality

• One or more reference stations are required to
  achieve sub-metre, centimetre or millimetre level
  accuracies with GPS
• GPS quality concerns can be divided into
• Integrity
• Positional accuracy

5
GPSnet Integrity

• Concerns (general)
• Data arrived at central server/latency
• Data is uncorrupted (correct RINEX format)
• Concerns (receiver)
• All channels are working
• No missing data
• Low noise, cycle slips, bad observations
• Low multipath error

6
GPSnet Integrity Software

• Wrote software to perform automatic quality
  checking
• Computes quality indicators
• Compiles reports and graphs
• Sends email error messages to network managers

7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Problems Identified by Software

• Malfunctioning receiver
• Interference from nearby RF antenna
• Communication network problems, base station
  power and hard disk failure (missing data)
• Relative performance of the difference models of
  receiver

11
Satellite Orbit
Satellite clock, antenna and transmitter
Receiver clock, antenna and tracking loops
Datum, ocean loading
12
Challenges with GPS

• Eliminate/model errors so that more accurate
  positions may be obtained
• Estimate more meaningful estimates of the quality
  of the results obtained

13
Data Collection
Atmospheric Model

• Data Screening
• bad observations
• - cycle slip detection and repair

Coordinates plus Quality Estimates
Error Modelling
Functional Model
Stochastic Model

• Pre-processing
• clock estimates
• differencing
• linear combinations

Ambiguity resolution and least-squares Adjustment

• Data Screening
• bad observations
• - unpaired observations

Results
Pre-processing
Parameter Estimation
14
Parameter Estimation

• Functional (mathematical) model
• Relates parameters to measurements
• Modifies measurements to remove error
• Stochastic model
• Used to weight the measurements in the least
  squares adjustment
• Models the magnitude and relationship between
  residual errors in the measurements

15
Ionospheric Error Handling

• In order to estimate L1 and L2 ambiguities
  directly ionospheric error must be removed
• Mathematical technique of pre-elimination is
  used
• A stochastic model is required to describe the
  relationship of errors between satellites and
  through time to make pre-elimination effective

16
Tropospheric Error Handling

• Estimation of the tropospheric error is essential
  for high accuracy long baseline positioning
• A stochastic model is required to describe the
  relationship of errors between satellites and
  through time to make estimation effective

17
Results to Date

• A modest improvement in ambiguity resolution has
  been seen through a modified ionospheric
  stochastic model
• Some improvement in fidelity of quality estimates
  has been achieved through multipath modelling

18
VSIS Issues to be addressed

• GDA94
• Full, universal adoption
• No longer support old datums
• Improve GPSnet
• Higher order
• Inclusion in national adjustments
• Improve local network/definition of GDA

19
Local vs National Control
20
VSIS Issues to be addressed

• Move to real time GPSnet
• Whole of government adoption of GPSnet
• Vicroads
• Emergency services
• Vic Channels

21
Questions?