October Trial Report

Eurocontrol Experimental Centre Brétigny-sur-Orge, FRANCE Eurocontrol ADS Programme Technology Assessment Task 1999 ADS-B Trials Part I October Trial Report Abstract This report presents the first round of ADS-B trials organised in 1999 by the Eurocontrol ADS Programme in the context of the ADS Technology Assessment Task. The objectives of these trials were to measure performance and assess the maturity of three candidate ADS-B technologies, namely Mode S Extended Squitter, VHF Datalink Mode 4, and Universal Access Transceiver. The first trial round focused on air to ground performance. Two trial aircraft were equipped with all three technologies. Base stations were installed at the Experimental Centre. The 1 st round trial flights took place in the period 6-12/10/1999. This document describes the equipment configuration used, the flight scenarios, and presents analysis results of the data logs collected in the trial flights. EEC Ref: EEC/SUR6E1/AC/015 Version: 0.6 Issue date: 24-Dec-99 Authors: C. Tamvaclis G. Rambaud L. Rabeyrin

DOCUMENT CONTROL LOG SECTION DATE Version REASON FOR CHANGE OR REFERENCE TO CHANGE 25-Oct-99 0.1 1 st Draft 12-Nov-99 0.5 2 nd Draft 24-Dec-99 0.6 3 rd Draft Ref. EEC/SUR6E1/AC/015 2

Table of Contents REFERENCES... 5 ACRONYMS AND ABBREVIATIONS... 6 1 INTRODUCTION... 8 1.1 DOCUMENT OBJECTIVES... 8 1.2 BACKGROUND... 8 1.3 1999 TRIAL OBJECTIVES... 9 1.4 ADS-B PERFORMANCE MEASURES... 9 1.5 SCOPE... 9 1.6 DOCUMENT OVERVIEW... 10 1.7 ACKNOWLEDGEMENTS... 11 2 TRIAL CONFIGURATION... 12 2.1 AIRCRAFT... 12 2.2 GROUND VEHICLES... 13 2.3 BASE STATIONS... 13 3 TRIAL SCENARIOS... 15 3.1 FLIGHT REQUIREMENTS AND CONSTRAINTS... 15 3.2 FLIGHT PROFILES... 15 3.3 TEST PROCEDURES... 16 3.3.1 Pre-flight tests... 16 3.3.2 Session I... 16 3.3.3 Session II... 16 3.3.4 Session III... 16 4 EXECUTION OF THE TRIAL... 17 4.1 PREFLIGHT TESTS... 17 4.1.1 Cessna Installation... 17 4.1.2 Ilyushin Installation... 18 4.2 SESSION I... 18 4.3 SESSION II... 20 4.4 SESSION III... 21 5 ANALYSIS OF LOGGED DATA... 23 5.1 ANALYSIS METHOD... 23 5.2 SESSION I : 6/10/99 1090 EXT. SQUITTER AND VDL-4... 24 5.2.1 Aircraft Trajectories... 24 5.2.2 Message Success Rates... 29 5.3 SESSION II: 8/10/99 VDL-4 AND 1090 EXT. SQUITTER... 31 5.3.1 Aircraft Trajectory... 31 5.3.2 Message Success Rates... 33 5.4 SESSION III: 12/10/99 UAT AND VDL-4... 36 5.4.1 Aircraft Trajectories... 36 5.4.2 Message Success Rates... 39 6 CONCLUSIONS... 42 7 ANNEX A... 44 7.1 TECHNOLOGY CHARACTERISTICS... 44 7.2 INSTALLATIONS... 44 8 ANNEX B: 1090 EXT. SQUITTER TRIAL IN TOULOUSE... 49 Ref. EEC/SUR6E1/AC/015 3

8.1 TRIAL SCENARIO... 49 8.2 ANALYSIS OF LOGGED DATA... 49 8.2.1 Ext. Squitter Position Trajectories...49 8.2.2 Ext. Squitter Success Rate... 50 8.2.3 TCAS Acquisition Squitters... 52 Ref. EEC/SUR6E1/AC/015 4

REFERENCES 1. ADS Technology Assessment Task Specification, Version 1.0, Ref. DED.3/SUR/ADS.TSK.99.001, Eurocontrol ADS Programme, Sept. 1999 2. Eurocontrol ADS-B Trials 1999 Part II, Ref. EEC/SUR6E1/AC/021, ASTP, Jan. 2000 3. FREER-3 Trials ADS-B over VHF-STDMA Performance Analysis, Version 1.0, Ref. EEC/SUR6E1/WP/006, ASTP, Dec. 1999 4. Phase One Link Evaluation Report, SF21 Technical/Certification Subgroup ADS-B Link Evaluation Team, Nov. 1999 5. ADS-B MASPS, DO-242, RTCA, Jan. 1998 6. ARTAS 2 Specifications, EATCHIP Document, Nov. 1996 Ref. EEC/SUR6E1/AC/015 5

ACRONYMS and ABBREVIATIONS ADS Automatic Dependent Surveillance ADS-B Automatic Dependent Surveillance - Broadcast ARTAS ATC Radar Tracker And Server ASTP ADS Studies and Trials Project (EEC) ATC Air Traffic Control CAA Civil Aviation Authority CCI Co-Channel Interference CDTI Cockpit Display of Traffic Information CEV Centre d Essais en Vol (Brétigny-France) CPR Compact Position Reporting CRC Cyclic Redundancy Code DERA Defense and Evaluation Research Agency (UK) DGAC Direction Générale de l Aviation Civile (French CAA) EATMP Eurocontrol Air Traffic Management Programme ECAC European Civil Aviation Conference EEC Eurocontrol Experimental Centre EUROCAE European Organisation for Civil Aviation Electronics FAA Federal Aviation Administration (USA) FEC Forward Error Correction GFSK Gaussian Frequency Shift Keying GosNIIAS State Research Institute of Aviation Systems (Russia) GPS Global Positioning System LDPU Link Data Processing Unit (UPS/AT product) MASPS Minimum Aviation System Performance Standards MSR Message Success Rate MTL Minimum Trigger Level NLR National Aerospace Laboratory (Netherlands) nmi nautical miles PPM Pulse Position Modulation PSU Power Supply Unit RF Radio Frequency RTCA Radio Technical Commission for Aeronautics (USA) SCAA Swedish Civil Aviation Authority (also known as LFV) SF21 Safe Flight 21 SNR Signal to Noise Ratio SSR Secondary Surveillance Radar Ref. EEC/SUR6E1/AC/015 6

STDMA Self Organising Time Division Multiple Access STNA Service Technique de la Navigation Aérienne (part of DGAC) SV State Vector TCAS Traffic alert and Collision Avoidance System TMA Terminal Manoeuvre Area UAT Universal Access Transceiver UPS United Parcel Service UPS/AT UPS Aviation Technologies (formerly known as II Morrow) VDL-4 VHF Datalink Mode 4 VHF Very High Frequency VSWR Voltage Standing Wave Ratio Ref. EEC/SUR6E1/AC/015 7

1 Introduction 1.1 Document Objectives This document presents results from the first round of ADS-B flight trials organised in 1999 by the Eurocontrol Experimental Centre (EEC) ADS Studies and Trials Project (ASTP). These trials involved three ADS-Broadcast (ADS-B) technologies, namely VHF DataIink Mode 4 (VDL-4) operating at 136.975 MHz Mode S Extended Squitter (also known as 1090 Extended [or Long] Squitter) operating at 1090 MHz Universal Access Transceiver (UAT) operating at 966 MHz The trials were organised as part of the activities of the ADS technology Assessment Task of the Eurocontrol ADS Programme [1]. 1.2 Background The Eurocontrol Automatic Dependent Surveillance (ADS) Programme 1 is part of the Eurocontrol Air Traffic Management Programme (EATMP) and aims towards the harmonised implementation of an ADS infrastructure in ECAC (European Civil Aviation Conference) capable of supporting the full Gate-to-Gate concept. The ADS Technology Assessment Task [1] of the Eurocontrol ADS Programme will evaluate the various ADS datalink technologies and determine their characteristics. The information collected by this task will serve as input to the Business Case for ADS in ECAC in order to select the most appropriate ADS-B datalink technology (or combination of technologies). As part of the ADS technology assessment work in 1999, Eurocontrol decided to organise comparative flight trials of the three main candidate ADS-B technologies [listed in Sec. 1.1]. These trials consisted of two rounds. The first round focused on measuring the basic performance characteristics of the candidate ADS-B technologies for air-to-ground operation. The second round (done in December 1999) focused on air-to-air performance characteristics. The first round trial flights were originally scheduled for the 29 and 30/9/99 but finally took place on the 6, 8, and 12/10/99 for reasons explained in Sec. 4. Two aircraft and a ground vehicle participated in these trials. The following organisations contributed by providing equipment/services/technical support : Centre d Essais en Vol (Brétigny, France) Eurocontrol Mode S Programme FAA Tech. Centre (USA) GosNIIAS (Russia) Swedish CAA Melun Aerodrome Authority (France) NLR (Netherlands) Saab Celsius Transpondertech AB (Sweden) UPS Aviation Technologies (USA) The two aircraft were equipped with all three ADS-B technologies. Ground stations were deployed at the Eurocontrol Experimental Centre in Brétigny. The aircraft were supposed to fly in parallel in the air space around Paris and transmit ADS-B reports, which were to be logged on the aircraft and on the ground stations. Concurrently radar data were to be collected from the SSR radars located at Orly and Palaiseau. 1 More information can be found in the ADS Programme web page: http://www.eurocontrol.be/projects/eatmp/ads/default.htm Ref. EEC/SUR6E1/AC/015 8

In addition to the two ADS-B trial rounds, ASTP analysed in 1999 results from two other ADS-B related trials, namely the 1090 Ext. Squitter Trial in Toulouse and the FREER-3 Trial of VHF- STDMA for airborne separation assurance. Results of the Toulouse trial are described in Annex B. FREER-3 trial data analysis is described in Ref. [3]. The results of those preliminary trials served as benchmarks for comparison with the ADS-B trial results in Paris. ADS-B Trials have also been conducted in 1999 by FAA under the Safe Flight 21 (SF21) Initiative (SF21) in the Ohio Valley and Los Angeles, see Ref. [4]. The results of the Sf21 activities have been taken into account in the analysis of the Eurocontrol ADS-B trial data. 1.3 Trial Objectives The aims of the Eurocontrol 1999 ADS-B Trials were [1]: Measurement of the basic datalink performance characteristics for each technology in the airspace around Paris [this airspace is one of the most heavily loaded traffic environments in Europe]; Collection of data for the calibration of the models used in ADS-B simulations; Assessment and comparison of the ADS-B performances of the three technologies under the same environment and flight conditions; Collection of data for evaluation of the tracking performance of the ARTAS-2 prototype surveillance data processing system [6] (which is capable of using ADS, Mode S, and multiradar data) 1.4 ADS-B Performance Measures In line with the methodology described in ref. [1], datalink performance was measured in terms of the probability of successful message delivery (also known as success rate) and its variance with distance of the receiver from the transmitter. Success rate measurements were used to derive mathematically estimates of the effective ADS-B position update periods versus distance. The calculation method is explained in Sec. 5.1. The resulting update period versus distance curves were compared with application requirements stated in the ADS-B MASPS [5], to obtain an indication of ADS-B system range. ADS-B MASPS were selected as comparison baseline for reasons explained in Sec. 1.5. Message Success Rates were calculated as follows: Messages 2 are transmitted periodically, therefore the number of messages transmitted per time unit is known. Receiver logs can be used to determine the number of messages received in a time unit. This number divided by the number of messages transmitted provides an estimate of the average success rate within the time unit. The resulting success rate can be associated with the horizontal distance between transmitter and receiver, i.e. the great circle distance between transmitter and receiver at the middle of the time unit. 1.5 Scope The following issues must be taken into account in interpreting the results presented in this document a) ADS-B standards have yet to be agreed in ICAO. There is an ADS-B standard published by RTCA [5] but it has not yet been adopted by either ICAO or EUROCAE. Consequently both 2 The concept of a message is different in each ADS-B technology. The transmitted message units are called squitters in Mode S, bursts in VDL-4, and messages in UAT. Furthermore they carry different subsets of ADS-B data. All three technologies define multiple types of squitters or bursts or messages. An ADS-B position report update may require more than one message unit, depending on the technology. Success rate calculations in this document always refer to a message unit, (i.e. a squitter in Mode S, a burst in VDL- 4, and a message in UAT) of a specific type. Sec. 5.1 explains the assumptions made for the position update rate calculations Ref. EEC/SUR6E1/AC/015 9

the content of ADS-B messages and ADS-B system performance requirements have yet to be agreed in Europe or indeed internationally. b) ADS-B information to be included on VDL-4 messages is not yet finalised in the VDL-4 standards. The frequency of VDL-4 transmissions for ADS-B purposes is also not finalised. c) There are no UAT standards. d) The existing standards for Mode S ext. squitter define both the ADS-B information to be included in the various squitter types and the frequency of transmission of the latter. e) Most equipment used had prototype status. Furthermore antennas and wiring installations used in the trials may have not been optimal for the technologies under test. It is therefore possible that the results obtained in the trials are not indicating the best performance achievable by each technology. In any case, one of the trial aims was to assess the maturity of existing implementations and not to build new implementations. Every effort was made to verify the quality of the trial installations to ensure conformance with manufacturer specifications 3. Installations were performed by people and organisations with long experience in avionics installations. f) In the analysis presented in this report, RTCA ADS-B MASPS requirements [5] were used as baseline for comparisons and mathematical modeling, together with the parameter settings adopted in Safe Flight 21 [4]. This does not imply that Eurocontrol ADS Programme has adopted either the RTCA ADS-B MASPS requirements [5] or the link specifications of SF21. These documents were used in order to facilitate comparisons with the results of the ongoing link evaluation of the FAA. g) In practice ADS-B performance will be dependent on the number of participating ADS-B stations (e.g. aircraft, vehicles, and base stations) as well as the available frequency spectrum for transmission. The Eurocontrol 1999 ADS-B trials involved a very small number of stations hence the impact of input traffic load on performance could not be evaluated. Complex scenarios with multiple ADS-B terminals are best tested through simulation, and the results of the Eurocontrol ADS-B trial will serve to tune the simulation models to be used in the capacity analysis work of the ADS Technology Assessment.. 1.6 Document Overview The contents of this report are organised as follows: Sec. 2 describes the trial configuration, e.g. the types of aircraft and ADS-B equipment used. Sec. 3 presents the flight scenarios and the test procedures. Sec. 4 describes what happened during the execution of the trial flights. Sec. 5 presents the analysis of the data logged during the trials. Sec. 6 presents the conclusions drawn from the execution of the trials and the data analysis. Annex A provides details of equipment configurations and installation. Annex B: 1090 Ext. Squitter Trial in Toulouse presents the results of that trial. 3 Saab/Celsius and SCAA staff participated in the installation tests. It is regrettable that other equipment providers did not follow their example Ref. EEC/SUR6E1/AC/015 10

1.7 Acknowledgements The conduct of this trial was made possible thanks to the contribution of the organisations listed in Sec. 1.2. SAAB/Celsius and SCAA contributed not only equipment but also sent members of their staff to Brétigny to assist in the trial. The teams from the aircraft providers (NLR and Gos- NIIAS) made great efforts to ensure the success of the trials Besides the Eurocontrol ADS Programme Team, many EEC and Eurocontrol HQ people, too numerous to list, helped in the conduct of the trial. The EEC Mode S Team successfully handled all Mode S related installation and test issues. The organisation of the trial flights was made possible thanks to the huge contributions of Rene Camus (CEV), and Bernard Brunner (EEC). Finally, the authors of this report note that Michel Biot (EEC) performed most of the analysis of Mode S trial data. His help is gratefully acknowledged. Ref. EEC/SUR6E1/AC/015 11

2 Trial Configuration 2.1 Aircraft The following two aircraft were used in the 1 st trial round: Cessna 550 Citation II, provided by NLR, see Figure 2 Ilyushin 18D, provided by GosNIIAS, see Figure 1 Figure 1 GosNIIAS Ilyushin 18D at Melun Figure 2 NLR Cessna 550 Citation at Brétigny Ref. EEC/SUR6E1/AC/015 12

The flight characteristics of the two aircraft were as follows: Aircraft Endurance, hours Range nmi Cruising Speed 4 (TAS) Ceiling ft Rate of Climb fpm Cessna 550 Citation II 3 :20 600-360 Knots 43000 2900 1600 Ilyushin 18D 8 3000 290 Knots 32000 1125 Each aircraft was meant to be equipped with the following ADS-B systems: Technology Equipment Antennas VDL Mode 4 one SAAB Celsius T4 Transceiver one VHF omnidirectional one SAAB Celsius WINLINCS data logger antenna (located at aircraft bottom) 1090 Ext. Squitter one Honeywell Transponder 5 one GPS receiver one Dassault 1090 ADS-B receiver 5 UAT one UPS/AT LDPU 7 one FAA Tech Centre CDTI Simulator 8 one GPS antenna two L-Band antennas (located top and bottom of the a/c) 6 one GPS antenna two L-Band antennas (located top and bottom of the a/c) 9 one GPS antenna The above equipment was loaned to Eurocontrol by the organisations indicated. Antennas were supplied by the aircraft providers, who also carried out all the wiring installations required. 2.2 Ground Vehicles A single 10 ground vehicle (Citroen Xantia, owned by the EEC) participated in the trial. The EEC car was fitted with a VDL-4 transceiver (same as the one used on the aircraft, see previous Section), a roof mounted VHF whip antenna, and a GPS antenna.. 2.3 Base Stations A single base station for each technology was installed at the Eurocontrol Experimental Centre, Brétigny, using the following equipment: 4 at around FL 300 5 supplied by the Eurocontrol Mode S programme 6 Two TX and two RX L-Band antennas were required for Mode S, but neither aircraft was able to provide four L-band antennas for Mode S as well as two L-Band antennas for UAT. 7 The UPS/AT LPDU contains a UAT transceiver and two Mode S Ext. squitter receivers. The latter were not used in this trial flights described in this report. They were used however in the 2 nd trial round (see [2]) 8 Software running on portable PC equipped with ARINC 429 interface for connection to the LPDU. 9 UAT Antenna installation at the bottom of the Cessna airframe was not feasible (see Sec. 4.1.1). 10 Effort was made to get a second vehicle from STNA, Toulouse, for use with Mode S. Unfortunately STNA could not send the car to Paris. Ref. EEC/SUR6E1/AC/015 13

Technology Equipment Antennas VDL Mode 4 one SAAB Celsius T4 Base Station one SAAB Celsius WINLINCS data logger 1090 Ext. Squitter one ERA Mode S Ground Station 5 one GPS receiver UAT one UPS/AT LDPU one FAA Tech. Centre CDTI simulator one VHF omni antenna 11 one GPS antenna one helical L-Band omni antenna 12 one GPS antenna one avionics L-Band blade antenna 13 one GPS antenna The EEC contributed a car which was equipped with a T4 transceiver (identical to the ones used in the aircraft) with one VHF omni antenna and one GPS antenna.. Annex A lists the characteristics of each technology as deployed in the trial and includes installation photographs. 11 KATHREIN model K512631, frequency range 116-152 MHz, 0 db Gain. It was located at EEC building top height 21 m. 12 ERA L-Band antenna, 5 db gain, with a 5 deg. cone of silence on the vertical plane and no cone of silence on the horizontal plane. All the ground station antennas were placed at the roof of the EEC building at a height of ~21 m. 13 UPS/AT Model AT-130 antenna tuned at 966 MHz Ref. EEC/SUR6E1/AC/015 14

3 Trial Scenarios 3.1 Flight requirements and constraints Originally it was planned that both aircraft use the Brétigny aerodrome (administered by the CEV). Due to CEV constraints, only the Cessna was able to use the Brétigny aerodrome. The Ilyushin had to use the Melun aerodrome, which is located 15 nmi east of Brétigny. Three flight sessions were planned. Each flight session were to last 2 hours and both aircraft would participate. The aircraft should follow a racetrack profile (described in Sec. 3.2) that would provide air to ground (i.e. to EEC) distances up to 200 nmi and air to air horizontal distances up to 100 nmi. In practice only the first session fulfilled this requirement (for reasons explained in Sec. 4). The aircraft should maintain a constant flight level near FL 300, except for ATC constraints. In practice, the aircraft had to fly at FL 270 due to ATC constraints. Vertical separation was kept within 3000 ft. 3.2 Flight Profiles Figure 3 shows the planned flight trajectories of the two aircraft. The first flight leg (~ 1 hour) would be in the direction of Brest (west of Paris) reaching a distance of 200 nmi from Brétigny (points B, B ). The two aircraft were to fly in parallel maintaining a lateral separation of about 50 nmi and a longitudinal. separation of <= 10 nmi. Then both aircraft would do a 180 deg. turn and the Cessna would gradually increase its lateral distance from the Ilyushin to about 100 nmi (points C, C ). The two aircraft would then gradually approach to about 10 nmi laterally (points D, D and stay within this lateral separation until approaching the airports. The longitudinal separation should stay within 10 nmi for most of the time. C N 50 nmi A B 50 nmi D Melun B 100 nmi 200 nmi C 10 nmi 80 nmi D ' Brétigny A1 A 20 nmi Figure 3 Planned profiles of 1 st Round ADS-B Trial flights Ref. EEC/SUR6E1/AC/015 15

3.3 Test procedures The flight profiles described in Sec. 3.2 were to be used in all three sessions. 3.3.1 Pre-flight tests 1. Aircraft installations were to be tested prior to the trial flights by Eurocontrol ADS and Mode S project personnel. These tests were meant to determine the quality of the antennas and wiring installed by the aircraft providers. 2. The VDL-4 installations (including the base station and the VDL-4 car) were also to be tested by SCAA and SAAB/Celsius personnel. 3.3.2 Session I 1. The NLR aircraft had to activate the 1090 Ext. Squitter transponder and also the VDL-4 transceiver. 2. The Ilyushin had to activate the 1090 Ext. Squitter transponder. 3. Both the 1090 Ext. Squitter and VDL-4 base stations were to be activated at the EEC. The VDL-4 equipped car (located at Melun) was also to be activated. 4. The received ADS-B messages were to be logged on both base stations and also in the VDL- 4 car for the duration of the session. 5. Radar plots were to be recorded from multiple SSR in the Paris area and also at Toulouse 14. 6. The VDL-4 equipped car was to be driven from Melun to Brétigny and back to Melun where it would wait for the arrival of the Ilyushin. 3.3.3 Session II 1. The NLR aircraft had to activate both the 1090 Ext. Squitter transponder and the VDL-4 transceiver plus a VDL-4 message logger (WINLINCS). 2. The Ilyushin had to activate the 1090 Ext. Squitter receiver 15, and also the VDL-4 transceiver and WINLINCS logger. 3. Both the 1090 Ext. Squitter and VDL-4 base stations were to be activated at the EEC. The VDL-4 equipped car (located at Melun) was also to be activated. 4. ADS-B message logs were to be logged on both base stations, the two aircraft and also the VDL-4 car for the duration of the session. 5. SSR plots were to be recorded from multiple radars in the Paris area and also at Toulouse. 6. The VDL-4 equipped car was to be driven from Melun to Brétigny and back to Melun where it would wait for the arrival of the Ilyushin. 3.3.4 Session III 1. The NLR aircraft had to activate the UAT transceiver (with LDPU recording enabled), and also the VDL-4 transceiver and WINLINCS logger. 2. The Ilyushin had to activate the UAT transceiver (with LDPU recording enabled), and also the VDL-4 transceiver and WINLINCS logger. 3. The VDL-4 base station was to be activated at the EEC. The VDL-4 station on the EEC car was also to be activated. 4. The received ADS-B messages were to be logged on the VDL-4 base station and the two aircraft as well as the VDL-4 car for the duration of the session. 5. SSR plots were to be recorded from multiple radar stations in the Paris area and also at Toulouse. 6. The VDL-4 equipped car was to be driven from Melun to Brétigny and back to Melun where it would wait for the arrival of the Ilyushin. 14 The Toulouse radar was necessary for achieving 100% SSR coverage of the flight paths. 15 It was not possible to activate both the 1090 Ext. Squitter transponder and receiver due to lack of sufficient Mode S antennas on the aircraft, see Sec. 4. Ref. EEC/SUR6E1/AC/015 16

4 Execution of the Trial Originally the three trial sessions were scheduled for the 29 and 30/9/99. Due to delays experienced by the aircraft providers in acquiring connectors and racks for the UAT LDPU, the sessions had to be rescheduled for the 6 and 7/10/99. 4.1 Preflight tests 4.1.1 Cessna Installation The NLR aircraft arrived at Brétigny on the 1/10/99 for pre-flight testing. The installation checks showed the following: (see Annex A for pictures of the installed equipment and installation diagrams) VDL-4: A single VHF antenna was used for VDL-4 which was located under the fuselage near the left wing of the aircraft. It would have been preferable to install it under the middle of fuselage but this was not feasible. VHF Antenna VSWR was measured by SAAB/Celsius at 1.33 (the T4 manufacturer specifies a VSWR < 1.50). A GPS antenna was installed in a antenna box on top of the aircraft fuselage and connected to the T4 transceiver. GPS SNR was measured by SAAB and found to be adequate for T4 operation. Operation was tested with the ground station (distance ~ 1 nmi) and the car (distance ~ 50 m) and it was found to be correct. 1090 Ext. Squitter: Two L-Band antennas had been installed in an antenna box on top of the aircraft fuselage. It was not feasible to install additional antennas under the fuselage of this aircraft. Antenna and wiring installation was found to be according to spec, and passed the tests prescribed by the transponder manufacturer. A GPS receiver had been connected via ARINC 429 to the Honeywell transponder to supply position information. The GPS antenna was installed in an antenna box at the top of the aircraft fuselage. An ADC system was connected via ARINC 429 to the Honeywell transponder to supply barometric altitude data. Operation was tested with the ground station (distance ~ 1 nmi) and it was found to be correct. UAT : Two L-Band antennas had been installed in an antenna box on top of the aircraft and they were connected to the LDPU. Normally one antenna should be installed under the fuselage but this was not feasible. Neither antenna was tuned to 966 MHz. Their DC resistance exceeded 300 Ohm while the LDPU manual requires a DC resistance of less than 10 Ohm. Eurocontrol loaned to NLR two tested 966 MHz avionics antennas 16 which were then installed on the same positions on top of the aircraft 17. The antenna wiring installation was tested again on the 6/10/99 and was found to be correct. A GPS antenna had been installed in an antenna box on top of the aircraft fuselage and connected to the LDPU. 16 UPS/AT model AT-130 L-Band blade antenna, tuned to 966 MHz. 17 UPS/AT recommended against the use of both antennas, since they were located on top of the aircraft. Ref. EEC/SUR6E1/AC/015 17

The LDPU had not been connected to a barometric altimeter, and no air/ground switch had been installed. 4.1.2 Ilyushin Installation The GosNIIAS aircraft arrived at Melun on the 4/10/99. Installation checks showed that: VDL-4: A single VHF antenna 18 was used for VDL-4 placed on the rear bottom of the aircraft. SAAB Celsius measured the VHF antenna VSWR and found it to be 2.05, which exceeds the specified maximum (=1.5). A GPS antenna 19 was installed on the top of the aircraft fuselage and connected to the T4 with a HF-cable. SAAB measured the GPS SNR and found it to be adequate for T4 operation. The T4 transceiver was powered by a +27 V DC source with an autonomous connector. The T4 manufacturer specifies +28V DC. VDL-4 system operation on the aircraft was tested with the car (distance ~ 50 m) and was found to be correct. 1090 Ext. Squitter: Two L-band antennas 20 had been placed at the rear top and bottom of the aircraft fuselage and were connected to the Honeywell transponder via HF L-Band cables; A GPS receiver (TOPSTAR 100 receiver by SEXTANT, France) was connected to the Honeywell transponder via ARINC 429; The Honeywell transponder had been connected via ARINC 429 to baroaltimeter simulator driven by the GPS altitude output; The EEC supplied a 115V 400Hz power supply unit (psu) for use with the Honeywell transponder; the psu was unearthed (isolated from the body bus); Operation was tested with a transportable receiver (distance ~ 50m) and it was found to be correct. UAT: The same two L-band antennas 20 and HF L-Band cables used for Mode S were also to be used for UAT; The measured antenna DC resistance impedance was < 1 Ohm, and hence conformant to LDPU manufacturer specs; The same type of GPS antenna and cable was used as for VDL-4; The LDPU was not connected to a barometric altimeter, and no air/ground switch had been installed. The LDPU was driven by a +27V DC supply (the LDPU manufacturer specifies +28V DC) 4.2 Session I This session took place as planned on the 6/10/99 with both aircraft participating. On the Cessna both Mode S Ext. Squitter and VDL-4 transceiver were activated. On the Ilyushin only the Mode S ext. squitter transponder was activated 21. 18 Russian-made АШС model. 19 AeroAntenna Technology Inc., USA Model AT575-9 S/N: 10429 20 Sensor Systems, Model SD65-5366-7L, bandwidth 960-1220 MHz 21 It was later discovered that the Ilyushin crew had also activated a VHF-STDMA station operating at 136.95 MHz, which was not foreseen in the trial scenario. In any case this did not appear to affect in any significant way VDL-4 operation on 136.975 MHz (25 KHz separation). Ref. EEC/SUR6E1/AC/015 18

The Brétigny VDL-4 and Mode S Ext. Squitter stations were active during the session. A WINLINCS terminal was used to monitor VDL-4 operation on the base station, while the ERA station provides its own logging and display facilities. Onboard the two aircraft Dassault 1090 receivers were used to monitor Mode S transponder transmissions. The following problems occurred concerning the ADS-B equipment 1090 Ext. Squitter: The ERA Base Station froze repeatedly during the session. The extended squitter recordings of the ERA station were found to be corrupted. ERA suggests that the failure was due to overloading of the receiver by SSR/TCAS traffic. However the station had been working correctly in preflight tests. An earlier version of the base station software was finally used and it worked but in the meantime the larger part of the session had been lost; On the Ilyushin, the Honeywell transponder was switched off for 30 minutes. This was done on pilot request at takeoff, because of supposed interference problems with VHF radio voice. It transpired that the Honeywell transponder had nothing to do with this interference; The Honeywell transponder on the Ilyushin transmitted pseudobarometric (GPS) altitude on Mode C replies. This caused problems with ATC and the transponder had to be reset a number of times during the flight.. VDL Mode 4 The aircraft position as received on the ground station became incorrect when the NLR aircraft crossed the Greenwich meridian. SAAB/Celsius reported that this was due to some error in the transceiver software concerning the encoding of longitude sign. Saab/Celsius restarted the ground station (by manual switching) when the aircraft was approaching maximum distance from Brétigny. This was due to the above problem with the passage of the Greenwich meridian. About 35 min of flight time were lost. Radar recordings were collected during the sessions and the observed radar tracks are shown in Figure 4. The two aircraft flew closely to the prescribed profile (compare with Figure 2). In the beginning of the flight the Cessna had to keep a holding pattern waiting for the Ilyushin to climb to the allocated flight level (270). On the return leg, the Ilyushin had to keep a holding pattern waiting for the Cessna to approach at the 10 nmi distance. The maximum distance from Brétigny exceeded 200 nmi, going beyond radar cover as Figure 4 indicates. Ref. EEC/SUR6E1/AC/015 19

Brétigny Melun Cessna Ilyushin Figure 4 Trial Session I: Aircraft radar tracks 4.3 Session II This session was planned for the morning of the 7/10/99. On the afternoon of 6/10/99 French ATC decided to block the trial flights because they were busy with the operational introduction of 8.33 KHz VHF channels. After negotiation, a new time slot was allocated for the afternoon of 8/10/99. Then on the 8/10/99, the Ilyushin engines failed to start at the allocated slot start time. Consequently, only the NLR aircraft was able to participate in this session, and its flight profile was simplified to follow the trajectory originally intended for the Ilyushin. Both VDL-4 and Mode S Ext. Squitter were activated on the Cessna. Both the Mode S Ext. Squitter and VDL-4 Base Stations were active at Brétigny. WINLINCS terminals were used to monitor VDL-4 operation on the aircraft and on the base station. For Mode S Ext. Squitter, the ERA position display was used. The EEC car equipped with VDL-4 participated in the session. A WINLINCS terminal onboard the car was used to monitor VDL-4 reception. Radar recordings were collected during the sessions and the observed radar track of the Cessna is shown in Figure 5. The maximum distance from Brétigny was about 160 nmi. Ref. EEC/SUR6E1/AC/015 20

Brétigny Figure 5 Trial Session II : Cessna radar track There were no ADS-B or equipment related incidents during this flight, except for the known problem of incorrect longitude sign on VDL-4 (see previous Sec.). The absence of the Ilyushin prevented the logging of air to air messages on 1090 Ext. Squitter. 4.4 Session III This session was originally planned for the afternoon of the 7/10/99 but it had to be rescheduled for the 11/10/99 due to the French ATC problems mentioned in the previous section. However, by that date the NLR aircraft was no longer available, so only the Ilyushin could participate. The latter became ready to fly only at the evening of the 11/10 and so the third session was eventually carried out on the 12/10/99. In this flight, both UAT and VDL-4 were activated on the aircraft. VDL-4 used the Russian GPS antenna (certificate IEMA.464656.001 PC)) of the VHF-STDMA system A UAT pseudo-aircraft station was installed at Brétigny using material recovered from the Cessna. The FAA Tech Centre CDTI was used to monitor LDPU performance on the aircraft and on the ground. Similarly a WINLINCS terminal was used to monitor VDL-4 operation on the aircraft and on the ground. The EEC car equipped with VDL-4 participated in the session. A WINLINCS terminal onboard the car was used to monitor VDL-4 operation. At the beginning of the flight the aircraft crew by mistake switched off the power supply of the UAT and VDL-4 systems, consequently there was no recording of the first few minutes of the flight. The radar track of the Ilyushin is shown in Figure 6. Due to ATC constraints the flight was directed to the south and the return path was shifted to the east. The aircraft arrived at a maximum distance of about 160 nmi from Brétigny. Ref. EEC/SUR6E1/AC/015 21

Brétigny Figure 6 Trial Session III: Ilyushin radar track Ref. EEC/SUR6E1/AC/015 22

5 Analysis of logged data 5.1 Analysis Method The collected data consist of the messages received on each mobile and base station timestamped with GPS UTC time. The logged messages were decoded using, for VDL-4 the WINLINCS software, for Mode S ERA and Dassault software, and for UAT software supplied by UPS/AT. The resulting lat./long. info was used to plot the 2xD aircraft trajectories (e.g. positions along lat. and long. axis). determine great circle distance from the ground station For each mobile station two lat./long. trajectories have been derived: one from its own ADS-B station (which corresponds to own GPS log and shows any equipment operation interruptions) and one from the ground station. The former trajectory represents the information input to the ADS-B system while the latter represents the information delivered by the ADS-B system. The timestamp information is used to calculate the success rate (probability of successful delivery of a datalink message) per minute = number of received messages per minute divided by the number of theoretically transmitted messages in one minute the update period = time elapsed between successive messages The calculated success rates were correlated with distances (minute average) to produce plots of message success rate variation with distance. These plots were then be used to estimate the variation of ADS-B report update period with distance, which were compared with the requirements for state vector update periods stated in the RTCA ADS-B MASPS. The ADS-B state vector update period was not measured directly in the trials because none of the three systems operated in its standard configuration 22. For example the trial VDL-4 equipment transmitted one burst (=single slot message) per second, but the future VDL-4 system is expected to transmit at a variable rate ranging from 1 burst/sec to 1 burst/10 sec depending on the operational conditions. Therefore it would not be fair to compare measured update periods under the trial configurations. For this reason update periods were estimated from the measured success rates assuming that the message transmission rates were those that would be applied in an operational ADS-B system. The update periods depends also on the relationship of datalink messages to ADS-B state vectors. In the case of Mode S Ext. Squitter it was assumed that the critical part of the state vector requires reception of both a position long squitter and a velocity long squitter. Both squitter types are supposed to be transmitted twice per second 23. In the case of VDL-4, it was assumed that! a single 1-slot VDL-4 message (=burst) provides sufficient information for the critical part of the ADS-B state vector. 22 e.g. the configuration projected for the future Operational ADS-B system. In fact, there are no formally agreed system descriptions for any of the three technologies. The SF21 Link Descriptions [4] have been used as reference for the present analysis. 23 Note that the trial Mode S transponders transmitted only position long squitters. ADS-B state vector updates require also velocity long squitters Ref. EEC/SUR6E1/AC/015 23

! These bursts are transmitted at a variable frequency 24, namely 1 burst/sec at the airport, 1 burst /5 sec at TMA, and 1 burst/10 sec en route. To facilitate update period calculations it was assumed that the airport rate would apply at distances up to 3 nmi from the base station, the TMA rate at distances up to 50 nmi from the base station, and the en route rate for distances beyond 50 nmi. All these numbers correspond to assumptions made in VDL-4 simulations up to now. In the case of UAT, the following assumptions were made based on the UAT system description in Safe Flight/21:! A single UAT message conveys the information needed for a state vector report! The UAT message is transmitted once per second Based on the above assumptions, the state vector update rates were calculated using the following equations: VDL-4 and UAT: Mode S: P= 1-(1-p) Tc/T => Tc = T * ln(1-p)/ln(1-p) P= ((1-(1-p) Tc/T ) 2 => Tc = T * ln(1- P)/ln(1-p) Where P = required percentile of update period, Tc = update period at percentile P, T= datalink message period, p = (measured in the trial) success rate. 5.2 Session I : 6/10/99 1090 Ext. Squitter and VDL-4 Both the Cessna and the Ilyushin participated in this session (see Sec. 4.2). The Cessna had both Mode S Ext. Squitter and VDL-4 stations active. The Ilyushin was supposed to have only the Mode Ext. Squitter station active (see Sec. 4.2). A Mode S Ext. Squitter (Dassault) receiver was also activated on each aircraft to capture broadcasted squitters (e.g. aircraft own positions) 25. 5.2.1 Aircraft Trajectories Figure 7 shows a two dimensional plot of the Cessna lat./long. positions transmitted by the Honeywell transponder and recorded by the on-board Dassault receiver. These positions correspond to the GPS input to the ADS-B system. It can be seen that the Honeywell transponder and the GPS receiver were operating for almost the complete duration of the flight. Unfortunately there is no equivalent recording from the Ilyushin. 24 Note that in the trial the VDL-4 transceiver transmitted one burst per second. 25 The Dassault receiver was not connected to an external antenna hence it could not capture incoming squitters from the other aircraft. Two additional Mode S antennas would have been required, but they were not available on either aircraft. Ref. EEC/SUR6E1/AC/015 24

49.2 49 Cessna Mode S Ext. Squitter Transmitted Positions latiitude, deg 48.8 48.6 48.4 48.2 48 47.8 47.6 47.4-4 -3-2 -1 0 1 2 3 longitude, deg Bretigny Figure 7 Mode S Ext. Squitter positions broadcasted by the Cessna 6/10/99 Figure 8 plots the received position squitters on the Mode S 1090 Ground Station (ERA) at Brétigny. Due to the base station problems mentioned in Sec. 4.2, only the last 20 minutes of the flight were recorded successfully. Consequently, the trajectories depicted in Figure 8 contain only the return approach of the two aircraft to the Brétigny and Melun airports. The trajectory of the Ilyushin was lost well before landing, but this is due to the low altitude of the aircraft, which made the Ilyushin invisible (loss of line of sight) from Brétigny. The trajectory of the Ilyushin contains many more gaps than the Cessna trajectory. The better quality reception from the Cessna was surprising because the Cessna did not have a bottom Mode S antenna (see Sec. 4.2) and hence it was expected to perform worse than the Ilyushin. This poor Ilyushin Mode S performance can be partly attributed to multiple power resets that occurred during the flight while trying to fix the pseudo baro altitude problem described in Sec 4.2. Ref. EEC/SUR6E1/AC/015 25

48.7 Mode S Ext. Squitter Trajectory 48.6 48.5 48.4 latiitude, deg 48.3 48.2 48.1 48.0 47.9 Cessna Ily ushin Bretigny Melun 47.8 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 longitude, deg Figure 8 Mode S Ext. Squitter recording on the ERA ground station 6/10/99 Figure 9 plots the position reports as received on the VDL-4 Ground Station at Brétigny. Since only the Cessna VDL-4 station had been activated, only the Cessna trajectory appears. It can be seen that: VDL-4 covered 3/4 of the flight in air-ground mode. The missing part of the flight corresponds to the period where the a/c was furthest from Brétigny. During that period the ground station was reset (see Sec. 4.2) and hence no positions were received. There are discontinuities in the reported positions around longitude 0. In fact the recorded positions presented negative longitude signs as positive. We have corrected this error in Figure 9, but even so there remain some discontinuities around longitude 0 as well as some misplaced positions between longitude 1 and 0.5. SAAB reported that the longitude sign problems are due to some transceiver software error in burst encoding/decoding. There are also a few intermediate gaps within the reported trajectory. The reasons are not clear. Ref. EEC/SUR6E1/AC/015 26

49.2 Cessna Trajectory - 6/10/99 49 48.8 Cessna EEC ground station 48.6 Latitude,deg 48.4 48.2 48 47.8 47.6 47.4-2 -1.5-1 -0.5 0 0.5 1 1.5 2 2.5 3 Longitude, deg Figure 9 VDL-4 Recording on Ground Station 6/10/99 Figure 10 and Figure 11 zoom on the take-off and landing trajectories for the Cessna as received on the VDL-4 Base Station. For comparison Figure 12 zooms on the Cessna landing trajectory as received on the Mode S Ext. Squitter Base Station. Both appear to be of good quality but mathematical analysis in the following section will show whether they would meet the MASPS requirements for ADS-B state vector update period. VDL-4: Cessna Take Off 48.612 Latitude, deg 48.607 48.602 EEC 48.597 2.32 2.325 2.33 2.335 2.34 2.345 2.35 2.355 2.36 2.365 Longitude, deg Figure 10 Cessna take off as recorded by VDL-4 base station 6/10/99 Ref. EEC/SUR6E1/AC/015 27

48.605 Cessna Landing : VDL-4 EEC 48.595 Latitude, deg 48.585 48.575 48.565 2.29 2.3 2.31 2.32 2.33 2.34 2.35 Longitude, deg Figure 11 Cessna landing as recorded on the VDL-4 Base Station 6/10/99 48.605 Cessna Landing: Mode S Ext. Squitter EEC 48.595 Latitude 48.585 48.575 48.565 2.29 2.30 2.31 2.32 2.33 2.34 2.35 Longitude Figure 12 Cessna landing as recorded on the Mode S Ext. Squitter Base Station 6/10/99 Ref. EEC/SUR6E1/AC/015 28

5.2.2 Message Success Rates Figure 13 shows the variation of the one-minute average 1090 Ext. Squitter success rate versus distance from the EEC base station. Success rate has been plotted separately for each aircraft. As expected from the trajectory graphs of the previous section, the Ilyushin had a much lower success rate than the Cessna. The distance covered goes only up to 40 nmi due to the base station logging problems encountered during the session Figure 14 shows the corresponding plot of Cessna VDL-4 message success rate versus distance from the EEC VDL-4 base station. The success rate achieved is in the order of 90-100% up to around 130 nmi and then drops off. There are a few intermediate drops below 90% corresponding to the holes observed in the VDL-4 trajectory plotted in Figure 9. The success rates achieved by VDL_4 are much higher than those achieved by Mode S ext. squitter. However, one should remember that VDL-4 operated in a channel free of co-channel interference (CCI), since only two VDL-4 transmitters were active. Mode S had to cope with CCI from SSR, TCAS and military systems, but then, it will always have to face CCI from such systems The observed lower success rates for Mode S are not necessarily an indication of poorer performance, because Mode S transmits at a higher rate (twice per second) than VDL-4. Figure 15 provides a more meaningful performance quality comparison for the two technologies. It plots the 95th percentile of state vector update period against distance for the two technologies. It also shows the MASPS requirement for the state vector update period. The update period values have been calculated from the measured success rates according to the method described in Sec. 5.1. As Figure 15 shows, the observed trial VDL-4 performance would meet MASPS requirements for the SV update period up to a range of 130 nmi except possibly for distances below 3 nmi and also a gap at 50-70 nmi. The observed trial Mode S Ext. Squitter performance would meet MASPS requirements up to the range of 30 nmi except for the gap around 15-20 nmi. For both technologies, the intermediate performance gaps might be due to equipment problems rather than some inherent technology failure. Ref. EEC/SUR6E1/AC/015 29

100% Mode S Ext. Squitter Success Rate Success Rate, % 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cessna Ilyushin 0 5 10 15 20 25 30 35 40 45 distance, nmi Figure 13 1090 Ext. Squitter Success Rate: Reception on ground station - 6/10/99 VDL-4 Message Success Rate 110% 100% 90% 80% Success Rate, % 70% 60% 50% 40% 30% 20% 10% 0% 0 20 40 60 80 100 120 140 160 180 Distance, nmi Figure 14 VDL-4 Message Success Rate: Reception on ground station 6/10/99 Ref. EEC/SUR6E1/AC/015 30

Update Period, sec 45 40 35 30 25 20 15 Cessna SV Update Period - 95% confidence VDL-4 Mode S MASPS 10 5 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 Distance, nmi Figure 15 Comparison of ADS-B state vector update periods 6/10/99 5.3 Session II: 8/10/99 VDL-4 and 1090 Ext. Squitter Only the Cessna participated in the second session (see Sec. 4.3), and it had both Mode S Ext. Squitter and VDL-4 stations active. A VDL-4 equipped car was also active during that session. 5.3.1 Aircraft Trajectory Figure 16 shows a two dimensional plot of the lat./long. positions recorded on the Cessna VDL-4 station. Own positions reflect the input from the GPS receiver contained within the station, and hence describe the GPS input to the ADS-B system. It can be seen that the T4 transponder and the attached GPS receiver were operating for most of the flight, with one notable gap on the return leg at longitudes 1.8-2.0 degrees. The reason for this gap is an automatic re-boot of the transponder (power reset). Note also that the VDL-4 car was detected on the aircraft. Figure 17 shows a two dimensional plot of the lat./long. positions recorded on the VDL-4 ground station that was installed at the EEC. We have corrected the wrong longitude signs that the T4 actually transmitted when the aircraft was in negative latitudes. There remain Cessna position deviations around the Greenwich Meridian similar to those observed in the first session (see Figure 9). In any case the ground station managed to capture the greatest part of the transmitted Cessna trajectory. Figure 18 shows a two dimensional plot of the lat./long. positions recorded on the Mode S Ext. Squitter Base Station(ERA) that was also installed at the EEC. The range of the captured Cessna trajectory is considerably better than what Mode S delivered in the first trial session (see Figure 8), where Mode S operation was plagued by equipment problems. It is clear however that a) the Mode S trajectory becomes progressively sparser as the distance from the base station becomes larger. This behaviour is similar to what was seen in the Toulouse trial of June 1999 (see Annex B8.2.1, Figure 36). b) The Mode S Base station captured a smaller part of the aircraft trajectory than VDL-4. Ref. EEC/SUR6E1/AC/015 31

48.7 Cessna VDL-4 Recording 48.6 48.5 NLR aircraft EEC car EEC ground station Latitude, deg 48.4 48.3 48.2 48.1 48-2.5-2 -1.5-1 -0.5 0 0.5 1 1.5 2 2.5 3 Longitude, deg. Figure 16 Cessna VDL-4 Station Log - 8/10/99 48.7 VDL-4 Base Station recording 48.6 48.5 NLR aircraft EEC car EEC ground station Latitude, deg 48.4 48.3 48.2 48.1 48-2 -1.5-1 -0.5 0 0.5 1 1.5 2 2.5 3 Longitude, deg Figure 17 EEC VDL-4 Base Station log - 8/10/99 Ref. EEC/SUR6E1/AC/015 32

Mode S Ext. Squitter Base Station Recording 48.7 48.6 EEC latitude, deg. 48.5 48.4 48.3 48.2 48.1 48.0-1.0-0.5 0.0 0.5 1.0 1.5 2.0 2.5 longitude, deg. Figure 18 Mode S Ext. Squitter Base Station Log - 8/10/99 5.3.2 Message Success Rates Figure 19 shows the variation of the one-minute average success rate of Cessna Mode S long squitters versus distance from the Mode S ground station. This session provided data up to 180 nmi, unlike the previous session (see Figure 13), but still the quality of the trajectory deteriorates rapidly as the aircraft is moving away from the base station. In comparison with the Toulouse trial results (see Annex B.8.2.2, Figure 39), the success rates of Figure 19 are significantly lower. For example at 40 nmi the Toulouse success rate ranged from 60 to 70% while in Figure 19 it ranged from 5 to 35%. The Toulouse environment has a lower Mode S fruit rate than Paris, and this factor might contribute to the observed performance difference. However, the Toulouse trial permitted also success rate measurements in Paris while the aircraft was en route from and to the UK (see Annex B.8.2.2, Figure 38). In that case the measured success rate at 70 nmi ranged from 5 to 35% while in Figure 19 it is below 10%. The ground station and antenna were the same in both cases, but the Toulouse trial aircraft (BAC 1-11) had two L-Band antennas located top and bottom of the fuselage (unlike the Cessna). Figure 20 shows the corresponding plot for the Cessna VDL-4 message success rate versus distance from the EEC Base station. The plot is similar to what was obtained in the previous session (see Figure 14). Success rate is in the order of 90-100% up to around 130 nmi then dropping off, and there are a few drops below 90% in intermediate distances. It should be noted that VDL-4 success rate calculations were somewhat perturbed by a problem with UTC reception timestamps in the WINLINCS recordings. The UTC reception timestamp is sometimes frozen for a period of a few minutes or it can jump backwards by up to 10-20 sec. The reason for this problem is not known. It is mitigated by the fact that the log includes the UTC transmission timestamp. The latter may also suffer from similar problems but usually at different times. This problem caused the loss of a few data and may have introduced some additional variation in the success rate estimates 26. Figure 21 compares the performance of the two technologies in terms of the 95 th percentile of the state vector update period. The ADS-B MASPS requirement for the 95th percentile of the state vector update period is also shown. The update period values have been calculated from the measured success rates according to the method described in Sec. 5.1. Figure 21 indicates that the observed trial VDL-4 performance would meet MASPS requirements for the SV update period up to a range of 130 nmi. In the case of Mode S Ext. Squitter, the up- 26 The same problem happened also on the other two sessions but it was less noticeable Ref. EEC/SUR6E1/AC/015 33

date period would meet MASPS requirements up to the range of 40 nmi, although there are two update period peaks exceeding the MASPS requirement around 15 and 25 nmi. 100% M ode S Ex t. Squitte r Succe ss Ra te 90% 80% 70% Success Rate, % 60% 50% 40% 30% 20% 10% 0% 0 20 40 60 80 100 120 140 distance, nmi Figure 19 Variation of Mode S Extended Squitter Success Rate versus distance 8/10/99 Ref. EEC/SUR6E1/AC/015 34

VDL-4 Message Success Rate 110% 100% 90% 80% Success Rate, % 70% 60% 50% 40% 30% 20% 10% 0% 0 20 40 60 80 100 120 140 160 180 Distance, nmi Figure 20 Variation of VDL-4 Message Success Rate versus distance 8/10/99 50 SV Update Period at 95% confidence 45 update period, sec 40 35 30 25 20 15 VDL-4 Mode S Ext. Squitter "MASPS" 10 5 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 distance, nmi Figure 21 Cessna State Vector Update Period at 95% confidence 8/10/99 Ref. EEC/SUR6E1/AC/015 35

5.4 Session III: 12/10/99 UAT and VDL-4 Only the Ilyushin participated in the third session (see Sec. 4.4), and it had both UAT and VDL-4 stations active. The second UAT station was installed at the EEC. A VDL-4 equipped car was also active during that session. 5.4.1 Aircraft Trajectories Figure 22 shows a two dimensional plot of the Ilyushin lat./long. positions as recorded on its onboard UAT station (LDPU). Own positions reflect the input from the GPS receiver contained within the LDPU, and hence represent the input to the ADS-B system. It can be seen that the LDPU and the attached GPS receiver were operating for most of the flight, with two notable gaps on the first flight leg at longitudes 2.3-2.4 degrees. In fact, the LDPU log indicated that five resets occurred during the flight. Figure 23 shows the corresponding two dimensional plot of lat./long. positions recorded on the Ilyushin VDL-4 station. This log indicates the information transmitted by the onboard T4 and also shows that the EEC car and the EEC VDL-4 base station were detected. The recorded aircraft trajectory matches the trajectory recorded on the LDPU since they both result from (different) onboard GPS receivers. There were two major T4 operation interruptions and they occurred in the same period as in the UAT case but they lasted longer. Figure 24 shows a two dimensional plot of lat./long. positions recorded on the UAT LDPU that was installed at the EEC. Clearly, only a small part of the trajectory transmitted was captured, and there were many more disruptions than in the transmitted trajectory. Figure 25 shows the corresponding plot of lat./long. positions recorded on the VDL-4 ground station installed at the EEC. Both the aircraft and the car were detected. The range of the captured Ilyushin trajectory is wider than that of UAT but the quality seems equally poor, especially if compared to the VDL-4 trajectories produced in the previous sessions. Ref. EEC/SUR6E1/AC/015 36

49 Ilyushin own UAT Track Melun 48 latitude, deg. 47 46 2 2.5 3 3.5 4 4.5 5 longitude, deg Figure 22 Ilyushin UAT LDPU own position log 12/10/99 Ilyushin VDL-4 recording 49 48.5 Ilyushin EEC car EEC VDL-4 base station 48 Latitude, deg. 47.5 47 46.5 46 2 2.5 3 3.5 4 4.5 5 Longitude, deg. Figure 23 Ilyushin VDL-4 station log - 12/10/99 Ref. EEC/SUR6E1/AC/015 37

48.7 UAT reception log 48.6 48.5 48.4 latitude, deg. 48.3 48.2 48.1 48 47.9 Ilyushin Bretigny Melun 47.8 47.7 47.6 47.5 47.4 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 longitude, deg. Figure 24 UAT ground station reception log, 12-10-99 Ref. EEC/SUR6E1/AC/015 38

49 VDL-4 Ground Station Log 48.5 Ilyushin EEC car EEC ground station 48 Latitude, deg. 47.5 47 46.5 46 2 2.5 3 3.5 4 4.5 5 Longitude, deg. Figure 25 EEC VDL-4 Ground Station Log 12/10/99 5.4.2 Message Success Rates Figure 26 shows the variation of the one-minute average success rate of Ilyushin VDL-4 messages versus distance from the EEC ground station. VDL-4 messages were received up to 150-160 nmi away from the base station just like in the previous sessions, but the measured success rates are well below what was seen then. Figure 27 shows the corresponding plot for the Ilyushin UAT message success rate versus distance from the EEC. UAT messages have been received from up to 70 nmi away but the success rate varied widely. This was somewhat surprising since UAT operated in an environment free of co-channel interference, unlike Mode S. Figure 28 compares the quality of ADS-B performance of the two technologies in terms of the 95 th percentile of the state vector update period that would result from the measured success rates. The ADS-B MASPS requirement for the 95th percentile of the state vector update period is also shown. The update period values have been calculated from the measured success rates according to the method described in Sec. 5.1. The observed VDL-4 performance is widely off the MASPS requirements. This contradicts the results of the previous two sessions, suggesting that there must have been a serious problem in the Ilyushin VDL-4 installation. The only problem identified in the pre-flight tests was a poor VHF antenna VSWR. For UAT, Figure 28 suggests that MASPS requirements could be met up to 60-70 nmi, although there are ten peaks exceeding the MASPS upper limit for the SV update period. These peaks are Ref. EEC/SUR6E1/AC/015 39

linked due to the resets noted in the LDPU log. UAT performance may also have suffered by the non optimised L-Band antennas used on the aircraft. Ilyushin VDL-4 Message Success Rate versus distance 45% 40% 35% Success Rate, % 30% 25% 20% 15% 10% 5% 0% 0 20 40 60 80 100 120 140 160 Distance, nmi Figure 26 VDL-4 message success rate at the EEC ground station 12/10/99 90% UAT Sucess Rate versus Distance 80% 70% Success Rate, % 60% 50% 40% 30% 20% 10% 0% 0 10 20 30 40 50 60 70 80 Distance, nmi Figure 27 UAT Message Success Rate air to ground 12/10/99 Ref. EEC/SUR6E1/AC/015 40

100 SV Update Period at 95% confidence Update period, sec 80 60 40 UAT VDL-4 MASPS 20 0 0 20 40 60 80 100 120 Distance, nmi Figure 28 ADS-B State Vector Update period (at 95% confidence) versus distance 12/10/99 Ref. EEC/SUR6E1/AC/015 41

6 Conclusions 6.1 Performance/Range The VDL-4 implementation (T4 transceiver) produced the better performance among the three ADS-B implementations tested. The T4 implementation achieved air to ground ranges 27 in the order of 120-130 nmi. This performance is similar to results reported from previous T4 trials in Sweden. Longer air to ground ranges have been reported in VHF-STDMA trials, where higher transmission powers have been used (10 W versus 5 W in the T4). The SF21 LET specification for VDL-4 indicates that VDL-4 transmitters might use 10-50 W TX power. Therefore such transmitters might well achieve ranges beyond 150 and even 200 nmi but this has yet to be tested. There is also the untested option of using two VHF antennas for VDL-4 on the aircraft which might have a positive impact on air to ground performance 28. The Mode S Ext. Squitter implementation (Honeywell Transponder, ERA receiver) achieved the lowest air to ground range (~ 40 nmi) of the three ADS-B trial systems. This is somewhat lower than what was measured in the previous Eurocontrol trials in Toulouse and Paris (see Annex B) with the same equipment (base station, transponder, antenna) but a different aircraft (DERA BAC-1-11). In fact TCAS acquisition squitter measurements in Toulouse suggested that the air to ground range might reach 100 nmi (see Annex B8.2.3), but the Toulouse airspace has a much lower Mode S fruit rate than Paris. However SAA/SF21 trial results [4] from Los Angeles Basin and the Ohio Valley have also Mode S ext. Squitter ranges in excess of 100 nmi [using the same type of transponder from Honeywell]. These performance differences might be attributed to the following factors:! The Cessna did not have a Mode S antenna at the bottom of the fuselage (but the BAC 1-11 and the Ilyushin did have such antennas)! The ERA station had a lower sensitivity (higher MTL) than the equipment used in the SF21 trials. However, on paper the difference does not seem significant (-86 versus 87 dbm).! The ERA station did not use the advanced decoding techniques that were implemented in the Mode S receivers of the SF21 trials. Indeed the SF21 Link Evaluation report [4] states that improved 1090 MHz receivers (relative to existing TCAS receivers) will be needed to meet all ADS-B MASPS range and integrity requirements.! Neither the ground nor the aircraft Mode S antennas used pre-amplification (used in the SF21 trials)! The ground Mode S station omni antenna was less efficient than the (DME and 6-sector) antennas used in the SF/21 trials particularly regarding its sensitivity to multipath. The UAT trial implementation (UPS/AT equipment) achieved an air to ground range of about 70 nmi. This is somewhat less than the ranges reported in the SF/21 Ohio Valley trials [4]. The transceiver was identical to what was used in the SF/21 trials but performance must have been penalised by the generic L-Band antennas used in the aircraft and also the avionics L-Band antennas used on the ground station. 6.2 Equipment maturity Only the Mode S Honeywell transponder was a commercial and certified product. All the other equipment used were uncertified prototypes and not available commercially. Eurocontrol could not find any alternative sources of equipment for any of the three technologies. 27 In this discussion range is measured as the maximal distance at which MASPS requirements [5] for the 95 th percentile of the state vector update period are met. 28 The Eurocontrol trials as well as all previous T4 and VHF-STDMA trials have used a single VHF antenna on the aircraft. Ref. EEC/SUR6E1/AC/015 42

Equipment related problems were encountered with all three technologies, and they involved both hardware and software. These problems were relatively minor but they did cause some loss of data and in some cases may have impacted negatively on performance. In particular all airborne receivers suffered from random albeit infrequent resets in both trial aircraft. These resets caused some data loss. Their cause is not known (the aircraft providers assert that their power supplies are very stable). The VDL-4 implementation [T4] suffered from position encoding and UTC timestamp errors (software problems?). The latest versions of the Mode S Ext. Squitter receiver {ERA} did not work. There were striking variances between the results obtained on the two aircraft despite the fact that identical equipment was used. Antenna and wiring implementation differences appear to have played a critical role for all three technologies. Yet, no manufacturer had defined installation test procedures that would permit reliable measurement of installation quality prior to the flights. The variances observed suggest also that the equipment was not robust enough to deal with variations in wiring, connections, and antennas that are usually encountered in aircraft installations. 6.3 Next steps The results obtained in the 1 st round did not permit to define with confidence the capabilities of the three candidate technologies. The reasons were explained in the previous two subsections. A number of potential improvements were listed, and these will be tried in the 2 nd round of the Eurocontrol ADS-B trials for 1999. This 2 nd round will focus on air to air performance, since unfortunate circumstances prevented the collection of air to air measurements in the 1 st Round. ADS Technology Assessment will be continued in the year 2000 and will address any issues arising from these trials by conducting studies, simulations, and further trials. The case of ADS-B for surface movement in airports will be one of the items to investigate in 2000. Progress in the definition of ADS-B operational requirements should allow also further refinement of performance analysis to enable more reliable range measurements and technology comparisons. If new ADS- B products appear in the market, they will also be tested. Eurocontrol seeks partners for the above technology assessment activities. Interested parties are invited to contact Dr. C. Tamvaclis (EEC) or Mr. P. Van Der Kraan (Eurocontrol HQ) to discuss the possibilities for collaboration. Ref. EEC/SUR6E1/AC/015 43

7 Annex A 7.1 Implementation Characteristics The following table lists the characteristics of the ADS-B equipment used in the trial. Characteristic 1090 ext. squitter VDL-4 UAT Frequency, MHz 1090 136.975 966 TX Rate 1 Mbps 19.2 Kbps 1 Mbps Modulation PPM GFSK 2400 KHz GFSK 312 KHz Synchronisation 4 pulse preamble first 24 bits first 36 bits Message length, bits 112 192 (after sync) 246-372 Parity, bits 24 16 48 FEC and 24 CRC Address, bits 24 27 25 Lat/Long, bits 17 unspecified 24 TX power 29, W 500 5 50 Polarisation vertical vertical vertical TX Rate, msg/sec 2 1 1 RX MTL, dbm -74 30, -86 31 unspecified -92 Notes! The Mode S equipment implemented CPR encoding for position coordinates as specified in the Mode S standards.! The VDL-4 transceiver did not implement CPR encoding, although it is required in the latest versions of the VDL-4 standards. 29 At the output of the transmitter 30 Dassault Ext. Squitter Receiver 31 ERA Mode S Ground Station Ref. EEC/SUR6E1/AC/015 44

7.2 Installations Figure 29 Honeywell Transponder and Control Panel installed on Cessna Figure 29 shows the installation of the 1090 Ext. Squitter transponder and its control panel on the Cessna. The portable PC in the background is the Dassault Ext. Squitter receiver. Figure 30 shows the installation of the UPS/AT LDPU on the Cessna. The LDPU was used only for UAT. The LDPU was connected to a control panel and a portable PC running CDTI software (from the FAA Tech. Centre) as shown in that Figure. Similarly Figure 31 shows the Saab/Celsius T4 transceiver installed on the Cessna. The transceiver was connected to a WINLNCS terminal (portable PC) not shown in the Figure. Figure 32 and Figure 33 show the corresponding Honeywell, LDPU, and T4 installations on the Ilyushin. Figure 34 shows that T4 transceiver and WINLINCS terminal installed on the car. Figure 35 shows the antenna installation on the car. The aircraft in the background of this Figure is the Ilyushin at Melun. Ref. EEC/SUR6E1/AC/015 45

Figure 30 LDPU with Control Panel and CDTI installed on Cessna Figure 31 VDL-4 T4 Transceiver installed on Cessna Ref. EEC/SUR6E1/AC/015 46

Figure 32 Honeywell Transponder with Control Panel and LDPU installed on Ilyushin Figure 33 VDL-4 T4 Transceiver and WINLINCS terminal installed on Ilyushin Ref. EEC/SUR6E1/AC/015 47

Figure 34 VDL-4 Transceiver and WINLINCS terminal installed on car Figure 35 VHF and GPS antennas installed on car Ref. EEC/SUR6E1/AC/015 48