HIGH-ACCURACY SERVICES WITHIN THE GALILEO CS: FEASIBILITY, DRIVERS AND EXPECTED PERFORMANCE

Size: px

Start display at page:

Download "HIGH-ACCURACY SERVICES WITHIN THE GALILEO CS: FEASIBILITY, DRIVERS AND EXPECTED PERFORMANCE"

Edgar Miles
5 years ago
Views:

1 GNSS TECHNOLOGY ADVANCES IN A MULTI- CONSTELLATION FRAMEWORK 26 September, 2014 Rome HIGH-ACCURACY SERVICES WITHIN THE GALILEO CS: FEASIBILITY, DRIVERS AND EXPECTED PERFORMANCE Juan R. Martín The contentofthe document, includingallthe information, data, communications, code, graphics, text, tables, images, photos, videos, music, drawings, sounds and in general all other information available in anyformand anymaterial and service presentisthe propertyofsogei and/or the authorsand/or of its licensees and assignors and is protected under the terms of legislation on copyright and intellectual property. Itisforbiddentouse, copy, alter, publishor distributethe documents, data and information and the associated images available on this document, without the written permission validly expressed bysogei and alwayssubjecttoanylegalrights. The copyright notes, the authorswhereindicatedor the source itself must in all cases be quoted in publications produced and distributed in any form. Property of GMV All rights reserved Good afternoon, ladies and gentlemen. Thank you Mr Chairman. It is for me a pleisure, and an honor, to be here today as one of the speakers in this workshop I would like to thank our host organization, SOGEI, for inviting me. As already introduced, the object of my talk will be the GNSS high accuracy services in the frame on the Galileo Commercial Service. 1

2 OUTLINE 1. Introduction to Precise Point Positioning (PPP): Fundamentals Benefits of a multi-constellation scenario State-of-the-art and trends 2. PPP within the Galileo Commercial Service (CS): Overview of the Galileo CS PPP feasibility within the Galileo CS CS demonstration plan Page 2 I have organized my presentation in two parts. In the first part, I will give an introduction to the Precise Point Positioning technology, which is the one that looks the most promising for the provision of high accuracy services within the Galileo CS. I will give a quick introduction on the PPP technique itself, the benefits a multi-constellation scenario can bring to it, with special consideration to Galileo, and the current state-of-the-art and trends. In the second part, I will give an introduction to the Galileo CS, including a summary of the analysis carried out to investigate the feasibility of PPP within it. I will conclude my talk with a presentation of the currently ongoing, and planned, demonstration activities. 2

4 USERS WISH LIST Page 4 Let s start the presentation by wondering ourselves what do users want? what is the ideal positioning system for users? If we ask them, the may answer us that the system they would like to have should be: -Accurate, as accurate as possible -Available everywhere -Instantaneous, we do not want to wait for the position to appear in the screen of our navigator -Cost affordable, and -Compatible and interoperable 4

GNSS POSITIONING ACCURACIES Standard Augmented High Accuracy 10 m 1 m

accuracy. What does accuracy actually means for users?

unaugmented services are sufficiently good.

developed augmentation systems such as SBAS or GBAS.

5 GNSS POSITIONING ACCURACIES Standard Augmented High Accuracy 10 m 1 m 0,10 m 0,01 m Page 5 Let s discuss now a bit about the first wish accuracy. What does accuracy actually means for users? We have users for which a few meters can be OK and the standard, unaugmented services are sufficiently good. Others need more accuracy, say around or beyond the meter, and who have developed augmentation systems such as SBAS or GBAS. And others need more accuracy, much more, say less than 10 cm and even 1 cm. My talk today will focus on this last group of users, the most demanding ones. 5

6 WHAT IS PPP? an advanced GNSS positioning technique which provides: Page 6 And what is PPP in the context of high accuracy GNSS? We could define PPP as and advanced high accuracy positioning technique that tries to satisfy the needs of the most demanding users, the ones who need centimetre level accuracy. If we would have to summarize what are the 3 main features of PPP, we could say that: -It is a high accuracy GNSS positioning technique, capable to deliver accuracies better than 10 cm. -It has global coverage, providing nearly the same performance everywhere. -It is an absolute positioning technique, this is, the estimated position is not relative to the position any reference station located nearby. 6

7 PPP: KEY PROCESSING FEATURES Page 7 What are the key processing features of PPP? -First, PPP employs highly accurate satellite ephemeris and clock models to minimize satellite errors, since the standard ephemeris and clock models transmitted by the satellites are not accurate enough. -Relies on very accurate carrier phase measurements and solves the for their ambiguities -Uses multi-frequency measurements to remove ionospheric propagation delays, taking advantage of the known property that the ionospheric delay is inversely proportional to the square of the carrier frequency. -Estimates the local troposphere delay, given that the existing models are not accurate enough. -Uses very detailed physical models and corrects for very small errors, irrelevant in standard applications, such as phase wind-up, antenna offsets, and displacements due to earth and oceanic tides. 7

8 PPP ARCHITECTURE Precise Orbit Determination and Time Synchro Raw measurements Corrections Raw measurements Page 8 And in order to do that, all PPP systems need a complex, processing system parallel to that of the core constellations, which includes: -A tracking network, comprising several tens of reference stations with global coverage -A high availability and redudant processing centre -A dissemination channel to transmit corrections to users, usually based on communication satellites and / or terrestrial systems 8

REAL TIME PPP PRODUCTS ACCURACY RT orbit accuracy ~ 3-5 cm (RMS) RT clock accuracy ~ 0.2-0.

9 REAL TIME PPP PRODUCTS ACCURACY RT orbit accuracy ~ 3-5 cm (RMS) RT clock accuracy ~ ns (RMS) Page 9 And this complex system is necessary to improve the accuracy of the satellite ephemeris and clock models down to a few centimeters, usually between 3 and 5 typically, and a fraction of the nanosecond, usually between 0,2 and 0,4. 9

PPP PERFORMANCE: ACCURACY GPS+GLONASS Real Time PPP Accuracy comparable to RTK with a single-baseline of around 10-15 Km Page 10 An employing this corrections and an advanced processing at user

10 PPP PERFORMANCE: ACCURACY GPS+GLONASS Real Time PPP Accuracy comparable to RTK with a single-baseline of around Km Page 10 An employing this corrections and an advanced processing at user level, we can achieve centimetric positioning as shown in this figure. This is a typical example of GPS+GLONASS PPP performance. You can see thath the RMS of the horizontal error is around 3 cm and the vertical error is around 5. 10

11 PPP PERFORMANCE: CONVERGENCE 3D Error GPS Only 0,6 m 0,4 m 0,2 m 10 min 20 min 30 min Page 11 But despite PPP accuracy can be quite comparable to that of RTK, RTK is still better than PPP in one aspect: Initialization time. In this plot you have a real example of PPP initialization process. The grey line shows the 3D error as a function of time and starting at the time the PPP is initialized. In this particular case, the PPP is processing only GPS measurements. What we usually see is that the typical error is: -Below 0,6 m say after 10 minutes -0,4 m after 20 min -0,2 m after 30 min And I say, usually, because the actual initialization time is a bit random and very sensitive to the number of satellites existing in a particular moment and the quality of their geometry. 11

QUICK START Very fast initialization (< 1 min) when receiver is placed in a known, precalibrated position stored in memory Work around for applications where user cannot afford

explored such as: -Advanced integer ambiguity tecniques, specially tailored to the PPP case -Transmission of precise tropospheric and ionospheric corrections, to speed up the estimation

We have seen very interesting talks on this subject in the ION-GNSS conference that took place two weeks ago in Florida -Integration with existing RTK networks Another technique which is

12 QUICK START Very fast initialization (< 1 min) when receiver is placed in a known, precalibrated position stored in memory Work around for applications where user cannot afford convergence time in a cold start Page 12 Improvement of the PPP initialization time is one of the areas where more R&D efforts is currently being done, and different techniques are being explored such as: -Advanced integer ambiguity tecniques, specially tailored to the PPP case -Transmission of precise tropospheric and ionospheric corrections, to speed up the estimation filter convergence process. We have seen very interesting talks on this subject in the ION-GNSS conference that took place two weeks ago in Florida -Integration with existing RTK networks Another technique which is usually employed is what we call quick start, which consist in initializing the estimation filter at a known position. The known position can be precalibrated and stored in memory in our PPP terminal. This is a quite realistic possibility for the majority of land applications. Employing this technique, PPP can deliver full accuracy in less than a minute. 12

PPP PERFORMANCE: OUTAGES Page 13 Another interesting are to explore is the tolerance of PPP to correction data losses, which usually occur in real environments specially if we are using ground

13 PPP PERFORMANCE: OUTAGES Page 13 Another interesting are to explore is the tolerance of PPP to correction data losses, which usually occur in real environments specially if we are using ground communication channels. What we have seen in our filed trials, is that PPP is quite robust to data losses, I would say much more than RTK. The reason for this is that PPP is fundamented on precise models of reality, such as precise models of the orbit or clocks dynamics, and these models can compensate the loss of correction data up to several minutes as shown in this slide. 13

R&D EFFORTS IN PPP AND TRENDS Faster initialization: Advance ambiguity resolution techniques External aids: RTK, precise iono and tropo corrections Increased robustness: Gap-bridging techniques

are the key trends in PPP and areas where R&D efforts will concentrate: -The first area is faster initialization: As we have explained, there is a lot of room for improvement here.

14 R&D EFFORTS IN PPP AND TRENDS Faster initialization: Advance ambiguity resolution techniques External aids: RTK, precise iono and tropo corrections Increased robustness: Gap-bridging techniques Single frequency PPP External sensor / aids A multi-constellation scenario is beneficial for both Page 14 And to conclude this quick introduction to PPP, I will say just a few words on what I think are the key trends in PPP and areas where R&D efforts will concentrate: -The first area is faster initialization: As we have explained, there is a lot of room for improvement here. Extensive R&D is being performed in the area of advanced ambiguity resolution techniques and of external aids to improve the convergence process -A second area of effort would be that of increased robusstness, including gap-bridging techniques to avoid PPP reinitialization when raw data is lost and ambiguities need to be re-estimated. R&D efforts are being carried out for instance in the area of single-frequency PPP, and use of external sensors and aids. And in both cases, PPP can largerly benefit from a multi-constellation scenario as I will explain in the next part of my presentation. 14

16 MULTI-CONSTELLATION BENEFITS Quick exercise: Check satellites visibility at GMV Premises in Madrid. 8 GPS 18 Sept 2014 (1 week ago) 17:00 pm Plots generated with Trimble s planning tool: Page 16 16

17 MULTI-CONSTELLATION BENEFITS 8 GPS 8 GLO Plots generated with Trimble s planning tool: Page 17 17

18 MULTI-CONSTELLATION BENEFITS 8 GPS 8 GLO 4 GAL Plots generated with Trimble s planning tool: Page 18 18

19 MULTI-CONSTELLATION BENEFITS 8 GPS 8 GLO 4 GAL 2 BEI 22 SVsin Total Plots generated with Trimble s planning tool: Page 19 19

20 KEY BENEFITS FOR PPP How can PPP benefit from this multi-constellation scenario? Page 20 20

21 PPP PERFORMANCE: CONVERGENCE 3D Error GPS Only 0,6 m 0,4 m 0,2 m 10 min 20 min 30 min Page 21 But despite PPP accuracy can be quite comparable to that of RTK, RTK is still better than PPP in one aspect: Initialization time. In this plot you have a real example of PPP initialization process. The grey line shows the 3D error as a function of time and starting at the time the PPP is initialized. In this particular case, the PPP is processing only GPS measurements. What we usually see is that the typical error is: -Below 0,6 m say after 10 minutes -0,4 m after 20 min -0,2 m after 30 min And I say, usually, because the actual initialization time is a bit random and very sensitive to the number of satellites existing in a particular moment and the quality of their geometry. 21

22 PPP PERFORMANCE: CONVERGENCE 3D Error GPS+GLONASS 0,4 m 0,25 m 0,15 m 10 min 20 min 30 min Page 22 In this slide, you can see how the initialization time improves when you add GLONASS satellites. This particular case shown here, I would say it is a bit better than the average. Typically what you can expect is: -0,40 m say after 10 minutes -0,25 m after 20 min -0,15 m after 30 min Though the addition of GLONASS measurements is very helpful for PPP and make it also more robust for instance to signal losses, is not a panacea, due to the need to estimate at user level the GLONASS inter-channel biases, which enter as new unknowns in the position estimation filter. Later in this presentation, I will show how different this can be with Galileo. 22

23 PPP INITIALIZATION Test executed on May 31 st 2013 during a 150 minute window with 4 IOV satellites over Madrid GPS+GLONASS+ 4 Galileo SVs 10 min, 20 cm Page 23 23

24 KEY BENEFITS FOR PPP Page 24 The second benefit we can anticipate is that derived from the new GNSS signals available, more robust and accurate than the legacy GPS and GLONASS signals. 24

signals blocked by buildings Page 25 In order to illustrate this second benefit, I will

25 SIGNAL BENEFITS Multi-GNSS PPP kinematictest in TresCantos executed on August 21 st 2014 Only 3 Galileo SVs available Sub-urban scenario, frequent trees foliage attenuation and signals blocked by buildings Page 25 In order to illustrate this second benefit, I will show the results of a real test we perfomed just one month ago, on the 21st of August. 25

26 SIGNAL BENEFITS GPS L2 tracking frequently lost Number of usable satellites usually dropping down to only 5 GPS+GLONASS SVs Consequence: PPP gaps Page 26 26

by the addition of just 3 satellites Promising

27 SIGNAL BENEFITS Notable PPP improvement by the addition of Galileo Overall robustness increased by the addition of just 3 satellites Promising results once the full constellation is deployed Page 27 27

29 GALILEO SERVICES Page 29 29

30 GALILEO COMMERCIAL SERVICE (CS) Galileo CS goal defined in EU Regulation 1285/2013 CS shall enable the development of applications for professional or commercial use by means of improved performanceand data with greater added value than those obtained CS through based on the commercial open service ranging and data, whose access shall be controllable...the CS shall be exploited through a revenue-sharing mechanism with the private sector. The CS signals shall be the Open Service ones, plus two encrypted signals in the E6-band. Page 30 30

band provides two channels: - The E6-B channel contains a data component, allowing the

31 CS FREQUENCIES E6-C channel: Transmits pilot signal component for ranging E6-B channel: Contains data component (448 bps) Both channels can be encrypted Page 31 The Galileo E6 band provides two channels: - The E6-B channel contains a data component, allowing the transmission of 448 bits per second - And the E6-C channel transmits a pilot ranging signal 31

32 WHICH SERVICES? Page 32 The main services foreseen to be part of the CS are authentication and high accuracy. Authentication is understood as the ability of the system to guarantee to the users that they are using signals from the Galileo satellites and not from any other source. In this context, spreading code encryption can be included as one of the authentication options for civil purposes and it also allows to control the access to the service. Another possibility is the used of navigation message authentication through asymmetrical or symmetrical architectures. The exact service performance and provision scheme is under analysis. High Accuracy is understood as the ability of the system to provide a positioning accuracy in the order of a few centimeters. The chosen technique for high accuracy is Precise Point Positioning (PPP). 32

Cadastral Off-shore Mining Construction Maritime

33 CS HA MARKET Sector Agriculture Application Tractor guidance Automatic steering Survey Land GIS/mapping Cadastral Off-shore Mining Construction Maritime Marine engineering Hydrographic survey Vessel docking Page 33 33

34 CS HA MARKET Page 34 34

35 CS HA MARKET 0,5 million devices In 2030 Tractor guidance Automatic steering Cadastral apps Page 35 35

37 CS DEFINITION STUDIES Two 1-year parallel CS definition studies kicked-off in Jan 2013: GALCS - GALileo Commercial Service definition CESAR - Galileo CommErcial Service, A Reality Objectives Page 37 Two parallel studies were launched to define the technical solutions and exploitation model of the service and also to analyse the needs and constraints of the market in the chosen services GALCS led by GMV with two partners, CGI and Helios consulting. And CESAR developed by FDC, Trimble, Fugro, Thales and Keynetics. Both studies concluded by mid-december 2013 obtaining interesting results 37

38 KEY SYSTEM CONSTRAINTS Analysis of PPP compatibility wrt Galileo system contraints: Bandwidth (448 bps) Existing Dissemination Network Data Transmission Latency Signal Characteristics Page 38 38

39 ACCURACY VS. BANDWIDTH Ideal situation: 1 Hz corrections Not compatible with CS bandwidth Page 39 39

40 ACCURACY VS. BANDWIDTH Corrections packed into 80 bps streams Capacity for up to 5 Providers Page 40 40

41 DISSEMINATION STRATEGY Number of CS Satellites? Ground infrastructure? Only satellites connected to ground can transmit CS data in real time Page 41 41

42 ULS LOCATION Page 42 42

43 DISSEMINATION STRATEGY Reference configuration: 5 Sites x 2 Antennas 73% Availability 5 Sites x 4 Antennas > 95% Availability Page 43 43

45 CS ROADMAP High-Accuracy COSMET Services Within the Galileo CS, 19 Feb 2014 Page Page

46 THE AALECS PROJECT AALECS = Authentic and Accurate Location Experimentation with the Commercial Service Quick Project Data KOM Jan 2014 Foreseen duration 2,5 years (Jan 2014 Jun 2016) Budget 4 M Consortium Overall Objective Phase#1 Phase#2 GMV (Prime), CGI, Qascom, Ifen, KU Leuven, Veripos Develop demonstrator platform to test CS with real SIS and support future service providers "Early Proof-Of-Concept (EPOC) Full demonstration Page 46 46

47 PHASE#1: EPOC EPOC Tests Scope: Generate and transmit sample CS data through the E6 Test the reception of E6 B/C signals both unencrypted and encrypted in laboratory and real user environments Initial performance characterization EPOC Experimentation activities Started last June Page 47 47

48 PREPARING FOR EPOC TESTS Users warned of E6 encryption activation Page 48 In fact, it is possible that some press release have reached you containing this information, or maybe you heard about the mysterious Galileo outages on July Yes, it was our fault ;) 48

49 EPOC TEST SET-UP Main constraint: No real-time transmission capability available yet in the system Total process latency: 2-3 days Page 49 49

STATIC TEST (JULY 22) Static test in a laboratory environment

(<50 cm RMS) based on orbit & clock corrections generated 3

GPS+GAL Page 50 The figure on the right shows the positioning

50 STATIC TEST (JULY 22) Static test in a laboratory environment Tracking of encrypted E6 signals Promising positioning accuracy (<50 cm RMS) based on orbit & clock corrections generated 3 days before! GPS+GAL Page 50 The figure on the right shows the positioning performance with a dataauthenticated PVT in a open-sky static scenario. The results obtained show that the RMS of the position error is always below 50 centimeters, which is a very promising result taking into account the age of the Cs Data. 50

KINEMATIC TEST (SEPT 03) Kinematic urban environment Tracking correctly almost all the time the encrypted E6 signals Generated,

test carried out the 3rd of September with unencrypted E6 signal.

51 KINEMATIC TEST (SEPT 03) Kinematic urban environment Tracking correctly almost all the time the encrypted E6 signals Generated, received and lost pages Tracking, losses and dummy pages 1 2 Page 51 In this slide, you can see the results obtained in a kinematic test carried out the 3rd of September with unencrypted E6 signal. The figure on your right shows the followed path during the test, it is mainly located in an expansion zone of Tres Cantos where there are many buildings under construction. During the test all the available Galileo satellites were in view and at mid-high elevations (>25º) On your left, we can see again that the tracking is correct almost all the time, with only a few pages lost, which demonstrates the robustness of the E6 signal not only in static and open sky environments. 51

52 KINEMATIC TEST (SEPT 03) Accuracy (RMS) after 15 min: North 0.81 m East 0.49 m Up 0.24 m magicppp Solution GPS+GAL 3 days latency in PPP products transmission! Page 52 The PVT solution obtained in the kinematic test. It is a bit worse than the one shown in the static test. Here we can see a long convergence period, it is mainly due to the quality of the products. Nevertheless, after removing the first 15 minutes, the RMS statistics are always below 1 meter. 52

CS DEMONSTRATOR ARCHITECTURE CSDEMO Elements: CS Receiver( RXP) Provider Test Bed (PTB) Datagens Signal Generator (SG) GNSS Service Center System Emulator (EMU) User Terminal Monitor Station Real RXP

53 CS DEMONSTRATOR ARCHITECTURE CSDEMO Elements: CS Receiver( RXP) Provider Test Bed (PTB) Datagens Signal Generator (SG) GNSS Service Center System Emulator (EMU) User Terminal Monitor Station Real RXP Data Simulated SIS DATAGEN Signal Generator Real-Time CS Data Link EMU GSC V1 Page 53 Now I am going to present an overview of the CS Demonstrator architecture The CS receiver is an operational receiver able to process Galileo E6 signal, to log data [Click 1] and to communicate in real time with the PTB. The receiver will be used as a monitor station to feed the CS data generation algorithms, the datagens. The PTB is the core of the CS Demonstrator. It is a platform able to receive data from external CS Providers and also to generate CS data using files or with the real SIS. [Click 2] The PTB will be able to process the data from the monitor stations and external stations both in real time or from recorded files. [Click 3] The PTB will be able to store CS data in files to feed a Signal Simulator, or to send the data to the RT interface of the GSC V1. The Signal Generator and Threat Simulator is a platform able to generate 53

54 NEXT STEPS Jan 14 June 14 Oct 14 Jan 15 Jun 15 Jan 16 Jun 16 KO SIS-TRR CSDEMO-AR SIS-TRR T0 T0+6 T0+10 T0+12 T0+15 T0+18 T0+21 T0+24 T0+30 Phase#1: EPOC EPOC tests Potential extension of EPOC tests Phase#2: Lab tests Phase#2: RT Tests Integration with GSC V1 Phase#2: RT Tests with external CSPs Page 54 54

56 CONCLUDING REMARKS PPP is a reasonably mature technology, though still subject to important evolutions and R&D efforts. Galileo is expected to bring important benefits to PPP: Better accuracy Shorter initialization time Increased robustness PPP feasibility for the provision of high accuracy services within the Galileo Commercial Service has been analysed. Available data bandwidth in the CS data channel seems to be enough for PPP: Capability for up to 5 Service Providers (TBC). Achievable data latency: Few seconds (TBC). Page 56 56

57 CONCLUSIONS Strong value proposition for key user communities Key technical advantages: 1 single integrated receiver Dissemination trough MEOs at high latitudes and in difficult environments. Key challenges and open points: Dissemination strategy Overall security design, access control, key dissemination to users SLAs (contents/kpis & actors) Full demonstration of the Galileo Commercial Service will be carried out over the next 2 years. Page 57 57

ION GNSS 2011 FILLING IN THE GAPS OF RTK WITH REGIONAL PPP

ION GNSS 2011 FILLING IN THE GAPS OF RTK WITH REGIONAL PPP SEPTEMBER 22 th, 2011 ION GNSS 2011. PORTLAND, OREGON, USA SESSION F3: PRECISE POSITIONING AND RTK FOR CIVIL APPLICATION C. García A. Mozo P.