Active Road Management Assisted by Satellite. ARMAS Phase II

Transcription

1 Active Road Management Assisted by Satellite ARMAS Phase II European Roundtable on Intelligent Roads Brussels, 26 January

2 2 Table of Contents Overview of ARMAS System Architecture Field Trials Conclusions

3 3 Table of Contents Overview of ARMAS System Architecture Field Trials Conclusions

4 Overview of ARMAS Technical Feasibility; Detailed analysis of key issues European Space Agency s (ESA) identified project Active in Phase I; Road Management Assisted by Satellite (ARMAS) aims at High-Level Architecture; transforming the transport infrastructure (roads, bridges, tunnels, urban areas) into safer and more customer-friendly Legal Issues (Privacy, Liability); the selected functionalities: environments, by: Rough Improving prototypes Safety; developed for main Increasing entities; Warnings Provision ; Dynamic Traffic Management capabilities; SOS Request ; Providing Electronic Fee Collection mechanisms; Development of a demonstrator for EFC based on Satellite Positioning ; ARMAS Phase I ARMAS Phase II Feasibility Study Test-bed development focused on EFC based in Satellite Positioning Time Frame Phase I Phase II Mar 03 Dec 03 Mar 04 Dec 05 4

5 Overview of ARMAS EFC based on Satellite Positioning Corridor Pricing/Passage Tolling; Congestion Zones/Cordon Pricing; Distance Based Pricing; Combination; Warnings Provision Provide information about hazards on the road ahead to the driver. SOS Request Electronic SOS request from inside the car: By the driver; Triggered automatically by car sensors. 5

6 6 Table of Contents Overview of ARMAS System Architecture Field Trials Conclusions

7 System Architecture DSRC GPS, EGNOS Galileo Road-Side and Central Equipment ARMAS Fixed-Part AFP GSM, GPRS UMTS CESARE II ISO MISTER E-Merge GTP On-Board Unit In-Car System ICS GPS SiS + EGNOS SiS + EGNOS over GPRS DSRC GSM Voice + GSM Data + GPRS Inertial Sensors (Gyros, Accel., Magneto.) Map Matching RAIM 7

8 System Architecture: Split of Functionality The definition of the architectural split between OBU and RSCE was made through a process of analysing different alternatives and quantifying those alternatives according to a set of specified attributes: Privacy; Social Accessibility; Cost; Scalability; Performance & Accuracy; Reliability & Maintenance; Interoperability; Fraud Robustness & Enforcement; Evidential Performance; Ability to Support Value-Added Services; Optional Functionality; Stakeholder Acceptance. Data Collection RSE Map Matching GNSS Position Enhancement Position Integrity Position Data Feature Recognition Segment Usage Route/ Topology Matching Road Usage Candidate Scoring Road Usage Charge Calculation Charge Record INS Sensors A B C1 C2 D 8

9 System Architecture: Split of Functionality 1. The attributes mentioned before were attributed a weight that will provide a relative importance between the attributes. 2. Later it was defined a rank for each architecture (A to D) and having in consideration the mentioned attributes. 3. Based on the weights and ranks obtained an overall was calculated. It should also be mentioned that this exercise was made according two visions: Private Cars and HGVs. Data Collection ICS AFP RSE Map Matching GNSS Position Enhancement Position Integrity Position Data Feature Recognition Segment Usage Route/ Topology Matching Road Usage Candidate Scoring Road Usage Charge Calculation Charge Record INS Sensors C1 9

10 System Architecture: Map Matching Gantries 10

11 11 Table of Contents Overview of ARMAS System Architecture Field Trials Conclusions

12 Field Trials Scenario 12

13 13 Field Trials: Dense Urban Scenario Congestion Charge Zone

14 14 Field Trials: Urban Scenario Congestion Charge Zone

15 Trial Run Example: Dense Urban Scenario 15

16 Trial Run Example: Urban Scenario 16

17 17 Geodetic GPS Truth Source Analysis Two different vehicles were used having the GNSS and roving GGPS antennas placed on their roofs. The position solutions derived from GGPS post-processing were taken as being the Truth Source. However, being GGPS as vulnerable to obstructions as a regular GPS receiver, the positions collected were much less than those collected from ICS, naturally with a larger impact in dense urban areas.

18 18 Geodetic GPS Truth Source: Accuracy ICS ACCURACY (COMPARED AGAINST GGPS) (metres) AVERAGE STD DEVIATION PERCENTILE 75 PERCENTILE 95 PERCENTILE 99 MAXIMUM MINIMUM 10,26 21,56 7,11 37,87 81,78 374,46 0,07

19 19 Road Centreline Truth Analysis Since Carrier Phase Differential availability in dense/built up urban environments can be intermittent and so cannot be relied upon exclusively to provide a truth source throughout the trial environments, additional analysis was performed using Linework Truth. Linework Truth provides a good coverage of reference positional information, for all of the test routes throughout all of the test environments. While the truth takes the form of a line and the nature of the correlation with test data will be nearest point, it provides a good reference to compare performance in many environmental conditions. To create the highly accurate linework required for truth data it was used the road centreline of Navteq maps.

20 Road Centreline Truth Source 20

21 21 Road Centreline Truth Source: Accuracy ICS ACCURACY (COMPARED AGAINST ROAD CENTRELINE) (metres) Global Geographic Scenario Dense Urban Open City Rural Day Time of Day Twilight Night AVERAGE 8,62 17,25 6,43 4,07 10,61 6,09 6,17 STD DEVIATION 22,14 38,20 7,93 5,18 27,70 9,47 13,31 PERCENTILE 75 8,12 17,27 8,50 5,28 9,73 6,80 6,38 PERCENTILE 95 25,70 55,81 18,71 10,17 34,97 19,78 18,42 PERCENTILE 99 81,99 194,24 32,76 15,34 111,62 46,54 56,39 MAXIMUM 820,37 820,37 147,62 109,86 820,37 200,20 289,29 MINIMUM 0,00 0,00 0,00 0,00 0,00 0,00 0,00

22 22 Road Centreline Truth Source: Availability ICS AVAILABILITY (COMPARED AGAINST ROAD CENTRELINE) Global Geographic Scenario Dense Urban Open City Rural Day Time of Day Twilight Night OVERALL 99,20% 98,53% 98,82% 99,94% 99,30% 99,98% 98,20% Error < 3m 37,48% 23,66% 37,19% 47,35% 33,74% 39,53% 44,76% Error < 8m 74,49% 52,86% 72,26% 91,13% 68,99% 80,74% 81,96% Error < 20m 92,36% 79,29% 95,46% 99,40% 89,95% 95,05% 95,69% Error < 50m 97,85% 93,79% 99,40% 99,64% 96,97% 99,07% 98,82%

23 23 GPS only versus GPS and EGNOS Analysis With the objective of comparing the performance of the positioning using GPS and EGNOS with the positioning using GPS only, we used the PEGASUS (version ) tool developed for EUROCONTROL and made available to us by ESA. The major reason for the use of this tool is that we were able to use EGNOS archived data and apply it to all the raw GPS positions available, bypassing the problem of EGNOS SiS availability. Using the road centreline truth source we then performed an RNPlike analysis on the produced data.

24 GPS only versus GPS and EGNOS 24

25 GPS only versus GPS and EGNOS 25

26 26 GPS only versus GPS and EGNOS : Accuracy Accuracy by GEOGRAPHICAL SCENARIO (metres) Type Geo Scenario Average Error (m) Error STD Deviation Percentile 75 Percentile 95 Percentile 99 Min. Max. Dense Urban GPS + EGNOS Urban Rural Dense Urban GPS Only Urban Rural

27 27 Table of Contents Overview of ARMAS System Architecture Field Trials Conclusions

28 Conclusions As expected, the GPS+EGNOS+INS positioning solution is good enough for the tested tolling situations. Accurate charging was produced although an extensive trial with a large number of vehicles is necessary for statistical validity. The only environment where the results are not optimal is the Dense Urban where, as expected, the performance of the system becomes heavily dependant on the performance of the INS, Map- Matching Algorithms and if applicable, tolling scenario (e.g. the definition of the Charging Zone boundaries). The other two system functionalities, SOS Request and Warnings Provision were easily implemented using the base tolling platform which proves that the approach followed is valid for providing more service and business oriented functionalities to the end customer. Besides being the basis of the Integrity concept of the system, EGNOS improved also the accuracy of the positioning. 28

29 Conclusions GPRS proved to be enough for uploading the charge-related data gathered in the OBU. However, for the dissemination of EGNOS based on SISNeT, the GPRS communication channel showed some problems: High latency; Low bandwidth; Coverage problems; We concluded that in order to be able to have a reliable and complete EGNOS feed the ideal solution would be transmission based on radio broadcast such as RDS or DAB although UMTS should already bring a very noticeable improvement. 29

30 30 Project Achievements ICS prototypes + AFP prototype (with COTS hardware) made available for further testing and trials; Platform able to test several diferent scenarios and applications; Hands-on knowledge of the problems/benefits of the use of ISO and MISTER; Field Trial data that raised several issues but overall confirmed the validity of the project choices; Comprehensive body of knowledge about all the technical issues related with EFC based in GNSS; Enabling the consortia members to engage with key industry stakeholders;

31 Questions? Paulo Alexandre Gomes Avenida Conselheiro Fernando Sousa, nº 19 12º Lisboa, Portugal Tel: Fax: Project Web Site: 31