long range real time kinematic positioning service genesis stennis space centre 26 th – 28 th...

Long Range Real Time Kinematic Positioning Service

Genesis

Stennis Space Centre

 26th – 28th August 2002

John Hanley

Senior GPS Analyst – Norwich, UK

Outline of Presentation

Introduction. Overview - General Background & Applications. Basic Thales DGPS Infrastructure. Genesis Reference Station and Hub Processes. Genesis User Processes and Data Flow. Status and Operational & Trial examples.

Early Trials (Rig Moves and Ship & Ferry Trials). Recent operations.

Quality Control of High accuracy systems.

IntroductionThales GeoSolutions:

Supplier and a User [Commercial View] Built reputation in Surveying and Positioning

services offshore worldwide Genesis System

Why develop Genesis? How and Where? What for?

The Future – Refinement and Development New group with Thales Navigation [Ashtech GPS

Technology]

Overview of Thales LRTK Genesis

Code / Improved Code or Carrier Phase

Technical basis is to improve on DGPS by using the Carrier-Phase information

Carrier gives better repeatability and accuracy

Trade-Offs with improved code-based methods.

Initialisation Data transmission Filtering and QC

Combination of both approaches

LRTK - Genesis System

Increased level of accuracy using latest GPS technologies Regional high performance solution (North Sea Region) For specific applications requiring better accuracies than

standard DGPS Targeted at Offshore applications – 20-30cm QC information as per DGPS (Is more needed? Standards?) Delivered via satellite Long ranges and multiple stations

Genesis Applications

Offshore Positioning Vertical control Emphasis on the Z- component Vessel Passage with draught close to navigable depth High Accuracy Navigation AUV navigation Real-Time vertical use – Multi Beam High Accuracy 3D Control – Rig Moves

Genesis has been aimed specifically at the Offshore Market.Niche market viewed at the 20-30cm level

Conventional RTK Attributes

High accuracy systems (1-5 centimetres) Essentially Carrier phase based Operate over relatively short range Single baseline approach Maximum range varies from 10 to 40 km Dependant on a reliable radio link Require local geodetic point for installation “Black Box” systems with little QC Expensive 2-Receiver system

Genesis LRTK Attributes

Satellite delivery based LRTK system Accuracy of 20-30 cm Operational over long baselines PC based software with multi-station computations PC based software allows for added system QC Provides a highly accurate solution Refinement and Development – No Limitations

Thales GPS Infrastructure

GPSSatellites

VSAT Satellites

80+ Reference Stations• Dual-Frequency Stations• Single Frequency• Generate, process and transmit messages

2 MCC facilities • Aberdeen and Singapore• 2 remote MCC facilities at Perth and Reston

• Monitor and Control• Archive• Maintain and Plan• Manage external entities• Interface

LES facilities• Uplink SkyFix messages• Including SkyFix Premier Messages

19 inches rack

RIMS A

RIMS A

X25NETWORK

19 inches rack

RIMS A

High Power and

Low Power satellite links

Genesis & MFX3Genesis & MFX3

Thales Reference Station Configuration

LRTK Genesis

North Sea Coverage Region

7 stations in the North Sea Dual Frequency Enabled Choke Ring Antenna fitted Pre-processing carried out Data sent to MCS (Master

Control Station) in Aberdeen. Tromso & Hammerfest added

Reference Station Pre-Processing Raw data taken from Geodetic GPS Receivers

Clock Corrections Cycle Slip on L1 and L2 Code – Carrier Filtering Multipath Mapping SNR used to assess measurement Quality Observations Compressed Transmitted to MCS

Major Upgrade Process Underway [2001-2003]

MCS Hub Configuration - Current Data output from the pre-processing functions is compressed

prior to transmission (reduce the bandwidth requirement) Input Compressed Long Range Real-Time Kinematic messages Interface to Unit Database (Udb) for User Control information Uplink to the delivery satellite

Visual displays Operator configurable settings Simple error handling and printout facility Bandwidth and Need for compression ( 8 to 1 reduction )

Uplink Message

Genesis Reference Station Input

Genesis Hub

UDb

Genesis User Processes (1) Data Reconstruction

Proprietary Compressed data

received by Decoder

User Pre-Processing User Dual Frequency GPS receiver board Data pre-processed in similar manner to Reference Stations

Observation Combination/Differencing [@ User] The key issue in the use of carrier-phase ranges. High emphasis on

carrier data: reduces sensitivity to geometry (DOP chimneys) reduces sensitivity to code anomalies

Observation differencing (single or double) can be used to reduce the contribution of various error components.

Genesis User Data Flow

Data RX

Position Estimates &

Quality Control

Position Calculation

Repair & Filter Observations

Repair & Filter Observations

CombineSynchroniseTransform

Weights

Mobile Station Almanac/Time/Ephemeris

Phase/Code/Observables

Compressed Reference Station Data

Uplinked from MCS

Genesis User Processes (2) Network Approach

Network approach is very much at the centre of this system More than one reference station provides additional observations

and increases system availability and integrity

Position Determination and Quality Control The position computation is built around the use of double

differenced carrier-phase observations Use of ionospheric delay free data addresses ionospheric error Enhanced code and tropospheric weighting improves solution

robustness

- Combined Genesis and SkyFix Installation -

Clients NAV system

Position Outputs

Dual frequency DGPS data

Optional inputs:

•3rd party RTCM including Type 15’s

• DeltaFix Corrections

SkyFix RTCM (Type 1, 2, 3, 16, 55)

Hardware Installation Architecture

SkyFix decoder

Genesis decoder

GPS Receiver

Genesis & MultiFix PC

Development Test-BedStatic Trials 1999-2000

Initial Test Network- History TCP/IP delivery 3/4 Station Networks 20-30cm accuracy - Planning

Operational Performance Examples

Selection of operational examples. Rig-Moves and Survey Jobs/Trials. Slow dynamic and High dynamic applications Various baselines considered on different trials. ‘Truth’ required for performance comparison.

Initially assessed against DGPS. Became clear that higher accuracy ‘truth’ was

required to assess performance and QC elements Algorithm improvement & ongoing trials from

2000-2002.

Aberdeen Rig Move Job - Slow Dynamics - 2001

Rig-Move greatly affected by the convoluted structural environment

Baselines 150km East - West High Repeatability can be clearly

seen Factors affecting positioning

Obstructions Constellation - Geometry of SV’s Constellation - Number of SV’s Loss of signal = Loss of Double Difference sets

Number of SV’s

Number of DD’s

ABZ Rig Move Job – Repeatability over DGPS

Multi Station DGPS 2 Station Genesis Solution

Early Dynamic Ship Trials - 2001 - Repeatability

Dynamic trials show repeatability Accuracy harder to assess due to the problem of

finding a suitable ‘truth’ system Post-Processing of raw data to obtain ‘truth’ will be

required Increase in Repeatability over DGPS Lower noise in LRTK solution

Multi Station DGPS 3 Station Genesis Solution

• Aberdeen 559 km

• Bergen 55 km

• Kristiansund 310 km

• Brønnøysund 647 km

• Sumburgh 344 km

Norwegian Ferry Dynamic Trial - 2001

LRTK Genesis

vs.

Post-Processed RTK solution

Norwegian Ferry Horizontal Performance

Norwegian Ferry – 3D Position Error

Delta East 0.06m (1- 68%)

Delta Height 0.10m (1- 68%)

Delta North 0.09m (1- 68%)

Norwegian Ferry – 3D Position Error

Norwegian Ferry – Height Comparison

Snøhvit Field Dynamic Job/Trial - 2002

3 Station networkHammerfest Tromso Kristiansund

Coincident Projects Aberdeen Pipe-lay Project Aberdeen Rig Move

Objective was to establish whether LRTK Genesis could provide height accuracy (in this high Latitude North Sea location) to determine a tide value for vessel.

Vessel reference position computed using numerous sensors:GPS antenna position (from Genesis)Pitch, Roll and Heave (from Motion Unit)Vessel draught sensor

Performance compared against short-range Thales Ashtech RTK and using Tide information logged at Hammerfest Tidal station.

Data currently being processed and evaluated.

Snøhvit Field Dynamic Job/Trial - 2002

Raw Antenna Height (JD175)

47.0

48.0

49.0

50.0

51.0

52.0

53.0

124005 134005 144005 154005 164005

GPS Time

Genesis (Hamm; Trom; Kris)

RTK

Genesis (Hamm; Trom)

-2

-1

0

1

2

RTK position changed to Differential

Lost RTK & Genesis Corrections

Snøhvit Field – Height Accuracy

Preliminary results on previous slide show that LRTK Genesis performance was comparable to short range RTK system.

Further processing by Thales Norway required.

Reception of corrections in high latitude areas is an issue. This is the operational reality !

Re-initialisation is therefore inevitable. Must not be excessive and as seamless as possible.

Quality Control process MUST be able to supply useful information to the user.

Snøhvit Field Trial – Preliminary Findings

LRTK SystemsImportance of QC – Issues and

Requirements

LRTK Systems Requirements

Continuous and high quality L1 & L2 GPS data Continuous reference station data / Corrections > 5 satellites good geometry = DOP

LRTK Systems Issues to Consider

Interruptions in local GPS data (masking / poor tracking) Loss of Correction link Latency of Reference station data Number of satellites drop DOP Holes

Convergence Time at Start-up (Initialisation) and on Re-Initialisation

UKOOA Guidelines Published by ‘United Kingdom Offshore Operators

Association (Surveying and Positioning Committee)’.

Installation and Operation of DGPS Equipment Quality Measures Minimum Training Standards GPS Receiver Outputs Data Exchange Format

Are similar standards suitable for LRTK

UKOOA Quality Measures Designed to produce a universal set of quality

measures for ‘DGPS’ positioning software

Unit Variance Marginal Detectable Error (MDE) Internal reliability External Reliability

F-test on Unit Variance W-test for Outliers

Statistical Measures

Unit Variance of the Position Fix Computation (Least Squares Weights)

F-Test on the Position Fix Unit Variance

W-Test on the Position Fix Residuals to detect Outliers

Still carried out on Least Squares Residuals, even within LRTK

LRTK Genesis - Statistical Measures

F-Test The F-Test is a test of the overall consistency of

the observations (double differences) and the resulting position solution.

Upper Test looks at poorer than expected measurements. [a priori = optimistic]

Lower Test looks at better than expected measurements [a priori = pessimistic]

W-Test The W-test is a statistical test applied to an

individual observation (double difference). The aim is to identify a faulty measurement.

The test is conducted using the residuals from the least squares position calculation.

High Accuracy Positioning - Recommendations

It is essential to assess the reliability and precision of each position fix to ensure the quality of GPS measurements

W-Test for outliers to be carried out for each fix F-Test for on Unit Variance to be carried out for each

fix. When no more outliers are identified in any fix,

precision and reliability measures are to be computed

Estimate of Precision – a posteriori error ellipse