Embed Size (px)
Citation preview
Dr.-Ing. Jochen BredemeyerHead of ResearchFlight Calibration Services GmbHHermann-Blenk-Str. 32AD-38108 BraunschweigGermany
E-mail: [email protected]
FLIGHT INSPECTION OF THE SECONDARY SURVEILLANCE RADARSIGNAL-IN-SPACE
ABSTRACT
The effectivness of the SSR as a main backboneof ATC is affected by the accumulation of signals inspace due to civil, military and also illegal transmis-sions on both SSR uplink and downlink channel. Con-ventional methods to inspect the performance of SSRas introduced in DOC8071 are not qualified to identifythe real contents within the bandwidth of those chan-nels in case of serious problems.
In contrast to this a proper investigation of the SSRshould include a view to the basic incoming pulseshape of telegrams aboard the inspecting aircraft. Anexperimental system to record the actual SSR Signal-in-Space and some algorithms to analyse the con-tents are introduced in this paper as well as variousmeasurements done with it during test flights: Heav-ily overcrowded air spaces induce a high radio loadtriggered by SSR ground stations and ACAS initiatedtransponder responses. The results derived from theanalyzed SSR video data allow a detailed separationof the involved telegram formats (Mode 1,2,3/A,C,S)so the system’s main benefit to be focussed on isthe numerical evaluation over time and motion. Fur-ther potential included is to detect interference apartfrom regular SSR telegram transmissions which blocktransponders or prevent receivers from getting a validaircraft response.
INTRODUCTION
Already during World War II U.S. and British Al-lied Forces developped so-called IFF (Identification
Friend or Foe) techniques which led to the first re-leseas of those systems. Developments continuedafter the war ended and in the meanwhile also civilaviation realized the benefit of active identification ofaircraft so the ICAO decided in 1954 to establish acivil derivative in compability to the military variant IFF”Mk X” [7]. It mainly bases on a airborne transponderbeing interrogated on the uplink channel 1030MHzand responding on the downlink channel 1090MHz.
The steady growth of air traffic in the past decadesin contrast to fixed frequency resources evokes thequestion in what way the high radio load on bothchannels diminishes air traffic control capabilities.The wide-area introduction of the newer, selectiveSSR variant Mode S would contribute to a strong re-duction of the radio load: Its main advantage againstthe conventional Modes 1,2,3/A,C is the capability totransmit selective calls to prevent the majority of air-borne transponders from responding. However, for astill unknown period the conventional modes must bemaintained because not yet all aircraft are equippedwith a Mode S transponder. Furthermore, europeancountries have currently no operable Mode S groundstations. In Germany i.e., there exists just one ex-perimental interrogator in Gotzenhain near Frankfurtwhose Mode S extensions are switched off most ofthe time.
Before the introduction of the Mode S based Air-borne Collision Avoiding System (ACAS) the uplinkradio load on 1030MHz could be estimated with aknown number of ground interrogators. In contrast tothis, an unknown number of non-selective respond-ing transponders led to a drastic radio load increaseon the downlink channel 1090MHz.
Band filter
Band filter
Uplink channel
Downlink channel
Pre Amp.
Pre Amp.
Low pass filter
Low pass filter
Log. Amplifier
Log. Amplifier
1030 MHz
1090 MHz
Baseband signal
Figure 1: Receiver for both SSR channels
To make things even more difficult, the growing ACASequipment extent of aircraft actually increases thetraffic on 1030MHz because ACAS instruments actlike ground stations and interrogate on the commonuplink channel.
In Germany like in most other countries there existscurrently no flight inspection of radio load on the SSRchannels. The following sections therefore describean experimental system and new analyze methods toperform a selective investigation on both uplink anddownlink channel.
RECEPTION OF SSR SIGNALS
Transient, short pulsed signals with a high dynamicrange require logarithmic receivers because a con-ventional automatic gain control (AGC) cannot quicklyenough adapt the current signal strength.
Modern logarithmic amplifiers directly operate in theRF band with a low noise figure and due to theirimplicit rectification they deliver directly a basebandvideo signal. As shown in fig. 1 the expense tocontruct a complete double channel receiver can behighly reduced applying those types of amplifiers.Preselecting the RF signal with a steep band pass fil-ter and pre-amplifiying it enables to feed the log. amp.directly so we get the SSR video at the output. An
analog-digital converter can be connected after thesignal is fed through an anti-aliasing low pass filterand a further linear amplification.
SIGNAL RECORDING
In order to investigate the effectiveness of differentanalyzing methods a continous recording of the videosignal is required. The signal processing of raw datais done in post processing to test if the calculated so-lutions are reproducible.
The constant bandwidth of SSR signal contents is acontrast to the continously increasing speed of com-puters. The bandwidth of modern PC architecturesallow a recording speed of 20MByte/s or more with-out loss of data which corresponds to the selected8Bit wide A/D conversion at a sample rate of 20MHz.The recording system is the one as introduced in [1]and bases on an CompactPCI platform with LINUXas its operating system. The author suggests to re-place the old-fashioned method of recording and vi-sualizing SSR video data with an oscilloscope andcamera-based system described in [4, Att.A to App.F]by a modern variant as mentioned before.
In addition to the radar raw data the PC also recordsthe real time-synchronized flight path data from theFlight Inspection System (FIS) with a repetition rateof 0.1s.
IN-FLIGHT MEASUREMENTS
According to former investigations of FAA and DFS [6]the air space above Frankfurt is regarded as the onewith the highest radio load in the world on both up-link and downlink channel. During the ferry flight on17-Sep-2001 to a regular flight inspection in Switzer-land the Frankfurt/Main air space was passed by afew NM at 07:00 UTC. The crossing height was flightlevel 286 and both SSR channels were recorded for afew minutes as shown in the geographic scenario infig. 2. The labeled moments are referenced to thebeginning of the time synchronization with the FISwhilst the Airport Surveillace Radars (ASR) identifyFrankfurt Airport and Gotzenhain is the only radar inGermany with Mode S capability. To perform that testflight the experimental system was installed aboard ofD–CFMB and connected to a bottom L-band antennajust as to the FIS.
8.5 8.6 8.7 8.8 8.9 949.4
49.6
49.8
50
50.2
50.4
50.6
50.8
0s 50s
100s 150s
200s 250s
300s 350s
400s 450s
0s 50s
100s 150s
200s 250s
300s 350s
400s 450s
500s
WGS84 Coordinates
Longitude [°]
Latit
ude
[°]
ASR FFM North
ASR FFM South
Götzenhain
07:07 Uhr UTCMeasurement onuplink channel1030 MHz
06:58 Uhr UTCMeasurement ondownlink channel 1090 MHz
Figure 2: Flight path and moments of the measure-ments on 17-Sep-2001
CORRELATION OF SSR TELEGRAMS
To separat the received SSR telegram formatsin uplink and downlink one has to compare therecorded Signal-in-Space data with the known SSRtelegram patterns. The relevant formats are shown infig. 3 (refer to [2]).
Uplink 1030MHz
Suppr.
Reply
0 dB
-9 dB
0.8 ìs
1.6 ìs
2 ìs: 3 ì
ììì
s: 5 s: 8 s: 21 s
Mode:123/AC
2 ìs
max. 24.6 ìs
P1 P3 P4
P2
Mode 1,2,3/A,CIntermodeSuppression
0.8 ìs
3.5 ìs
4.75 ìs0.5 ìs
0.25 ìs
sync.phaserevers.
19.75 ìs (short)33.75 ìs (long)
Mode S
1 2 3 4
Bit:P1 P2 P6
Downlink 1090MHz
1.45 ìs
20.3 ìs
0 1 2 3 4 5 6 7 8 9
P1 P2 P3 P4
0.5 ìs
F1 F2C1 A1 C2 A2 C4 A4
20 ìs
1 ìs 1 ìs
3.5 ìs
8 ìs
Preambel
0 1 2 3 4 5 6 7 8 9 10 11 12 13 ìs
1 2 3 4 5
Data Bits
Bit:
1 ìs
56 oder 112 ìs
Mode 1,2,3/A,C
Mode S
Data Bits
1 0 0 0 0 01 1 1 1 1
0.45 ìs
Figure 3: SSR telegram formats
The conventional uplink formats differ in their distancebetween P1 and P3 while the amplitude of pulse P2indicates a side lobe transmission and prevents thetransponder from replying. In case of Mode S thereis a preambel consisting of P1 and P2 followed by theDPSK (Differential Phase Shift Keying)-coded datapulse P6. Mode S SLS (Side Lobe Suppression) isrealized by a pulse P5 which jams the Sync PhaseReversal so the transponder does not react either.Furthermore, Mode S capable radars can add a fol-lowing pulse P4 to the telegram which only triggersconventional transponders. This feature is especiallyused by ACAS which recognizes also conventionallyequipped aircraft by transmitting Whisper-Shout se-quences [2, Att. A 3.1.2].
The content of a reply to all conventional interroga-tions is surrounded by two frame pulses F1 and F2within a constant interval whereas a Mode S reply isled by a 4 pulse preambel.
Generally the existence of those relevant pulses hasto be checked in order to recognize a valid interroga-tion or reply. However, i.e. if many transponders replyin a highly overcrowded air space there may occurSynchronous Garbling so the radar signal processioncannot anymore detect single reply formats.
In contrast to this, the recorded video raw data of theexperimental system allows to test and to optimizesoftware algorithms which offer a better performance.A well-known method of pattern recognition is the dis-crete autocorrelation function (ACF) [5]:
φSS(m) =N�1
∑n=0
s(n)s(n+m)
It comprises a stepwise multiplication and summa-tion of the SSR model sequence s(n) with the rawdata with a total length of N. In case of m = N itreaches maxmimum compliance and therefore indi-cates a valid telegram at a clear time of arrival.
Selected correlation patterns s(n) like the GPS GoldCodes cause steep maximums and a large main-to-side-lobe ratio but SSR telegrams are not trimmed toreach maximum correlation performance. Thereforeit is necessary to include further security checks if amaximum is found.
The application of the ACF with SSR signals is illus-trated in fig. 4. In the upper diagrams two recordedsequences are displayed: A military Mode 2 and aMode S interrogation.
The model patterns below include different valuationsduring pulses and gaps which leads to a better maxi-mum detection performance: Passing by single stepsn<N during its run the ACF reaches negative valuesif one or more recorded pulses are shifted into pat-tern gaps as shown in the lower diagrams whereas
the main maximum is achieved if a total congruencyof model pattern and recorded data is given. Due tothe fact that a conventional Mode 2 interrogation mayinclude a P2 in case of SLS the valuation within thispulse duration is set to zero in order not to falsify theACF result. As a result of adapting different modelpattern valuations it was possible to correlate tele-gram formats down to the receiver base noise level–90dBm.
Things are getting more complicated if the downlinkchannel is concerned where it becomes clear that theevaluation of the ACF maximum cannot be the onlycriterion to be taken into consideration. Fig. 5 showsthe corresponding diagrams of two downlink telegramformats. Within the Mode 1,2,3/A,C (left) reply framepulse 4 octal numbers are encoded so 4 �3= 12 gapsor pulses are statistically distributed. The model pat-tern regards those positions as neutral and thereforeits portions of the total ACF value are zero. Only thetwo frame pulses F1 and F2 effect a positive ACFvalue which leads to a fairly unusable main-to-side-lobe ratio comparing the results with the uplink chan-nel. Consequently, another aspect must be the ampli-tude of the frame pulses at their known positions. Thelinear difference of their A/D converter values can beexpressed in dB due to the calibration curve of theLogAmp so a logarithmic threshold is applicable.
Focussing on the Mode S downlink format (right)things are easy as on the uplink channel: A rela-tively long preambel preceding the information pulsescauses a unique positive maximum without ambiguity.
Summing up those aspects we point out the followingnecessities to detect sucessfully SSR telegram for-mats applying the ACF algorithm:
� Maximum value of ACF as a mark of identifica-tion
� Signed valuation of model telegram patterns
� Ratio of significant pulse amplitudes
Interrogation Mode 2 Interrogation Mode S
−105
−100
−95
−90
−85
−80
−75
0 5 10 15 20 25 30 35 40 45 50
Zeit [µs]
Baseband signal
Leve
l[dB
m]
−105
−100
−95
−90
−85
−80
−75
−70
−65
0 5 10 15 20 25 30 35
Zeit [µs]
Baseband signal
Leve
l[dB
m]
0
10
20
30
40
50
60
70
80
90
100
0 1 2 3 4 5 6 7 8 9 10-2
-1.5
-1
-0.5
0
0.5
1
Zeit [µs]
Congruency: maximum of ACF
A/D
conv
erte
rva
lue
Val
uatio
n
0
20
40
60
80
100
120
140
160
0 1 2 3 4 5 6 7 8-2
-1.5
-1
-0.5
0
0.5
1
Zeit [µs]
Congruency: maximum of ACFA
/Dco
nver
ter
valu
e
Val
uatio
n
−7000
−6000
−5000
−4000
−3000
−2000
−1000
0
1000
2000
3000
0 1 2 3 4 5 6 7 8 9 10
Zeit [µs]
Passage of ACF to maximum
φ SS(
m
)
−6000
−4000
−2000
0
2000
4000
6000
8000
0 1 2 3 4 5 6 7 8
Zeit [µs]
Passage of ACF to maximum
φ SS(
m
)
Figure 4: Autocorrelation function in the Uplink channel
Reply Mode 1,2,3/A,C Reply Mode S
−100
−95
−90
−85
−80
−75
0 5 10 15 20 25 30 35 40 45 50
Zeit [µs]
Baseband signal
Leve
l[dB
m]
−100
−95
−90
−85
−80
−75
−70
−65
0 10 20 30 40 50 60 70 80
Zeit [µs]
Baseband signal
Leve
l[dB
m]
0
10
20
30
40
50
60
70
80
0 5 10 15 20 25-1
-0.5
0
0.5
1
Zeit [µs]
Congruency: maximum of ACF
A/D
conv
erte
rva
lue
Val
uatio
n
0
20
40
60
80
100
120
0 1 2 3 4 5 6 7 8 9-2
-1.5
-1
-0.5
0
0.5
1
Zeit [µs]
Congruency: maximum of ACF
A/D
conv
erte
rva
lue
Val
uatio
n
−6000
−5000
−4000
−3000
−2000
−1000
0
1000
0 5 10 15 20 25
Zeit [µs]
Passage of ACF to maximum
φ SS(
m
)
−12000
−10000
−8000
−6000
−4000
−2000
0
2000
0 1 2 3 4 5 6 7 8 9
Zeit [µs]
Passage of ACF to maximum
φ SS(
m
)
Figure 5: Autocorrelation function in the Downlink channel
PRESENTATION OF RADIO LOAD CALCULATIONS
By applying the decribed techniques it becomes pos-sible to perform a refined analysis of the collectedSSR raw data. Within the FIS time slot of 100ms alltelegrams are summed up and displayed over a timeaxis which corresponds to the flight path as depictedin fig. 2.
0 100 200 300 400 5000
200
400
600
800
1000
1200
1400
1600
1800Downlink Mode 1,2,3/A,C
time [sec]
tele
gram
m r
ate
[10
s−1 ]
total number thereof > −80dBm
0 100 200 300 400 5000
10
20
30
40
50
60
70Downlink Mode S
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −80dBm
Figure 6: Downlink telegram rates
Evaluating the downlink channel first fig. 6 depicts thetotal number of correlated replies including all repliesdown to the limit of the receiver’s sensitivity. Morethan 15000 conventional telegrams/s (above) statis-tically significate that every 67µs a 21µs-lasting re-ply is received. Setting the minimum threshold up to–80dBm on the contrary shows that most of the tele-grams in space are fairly weak.
The number of Mode S replies (below) is ten timessmaller than the conventional radio load. It mainly en-closes the Mode S transponder squitter broadcastsand replies to ACAS interrogations of aircraft.
On the uplink channel 1030MHz a more detailed sub-division of conventional modes can be achieved ac-cording to fig. 3.
The results of pure military Modes 1 and 2 are shownin fig. 7. In contrast to the total number of correlatedinterrogations those with a level above –77dBm areseparated (left) which corresponds to the minimumtrigger level (MTL) of transponders according to [3,sec. 3.8.1.7.5.1]. Generally these evaluations con-cern real interrogations (P1>P2) whereas a distinc-tion between main and side lobe interogations is pos-sible by making full use of the SLS-indicator P2. Theresult (right) describes a high portion of interrogationsreceived from the main lobe.
An expected higher activity can be determined in thesquawk interrogation Mode 3/A and in the civil alti-tude interrogation Mode C rate as depicted in fig. 8.Mode A interrogation rates of civil radars are higherthan Mode C and in addition with military Mode 3 theamount of 3/A is clearly higher than Mode C in gen-eral.
The preceding considerations did not include Inter-mode C interrogations elicited by ACAS Whisper-Shout acitivity. Nowadays more and more aircraft areequipped with ACAS which leads to the measuredhigh interrogation rate as shown in the upper two di-agrams in fig. 9: A large number of more or less faraway ACAS-equipped aircraft transmit whisper-shoutsequences whose initial transmissions are of weakoutput power. This explains the big difference be-tween the total number of correlated telegrams andthose above MTL. Comparing the two numbers aboveMTL one can derive that the activity of ACAS Inter-mode C and Mode C ground interrogation is mostlybalanced.
In the lower diagram the Mode S interrogation rate ispresented. Due to the fact that the Mode S extensionof Gotzenhain radar was disabled during the test flightthe activity is traced back only to ACAS-equipped air-craft. Like before the majority of detected telegrams isbelow MTL what indicates a huge number of aircraftflying at a large distance.
0 100 200 300 400 5000
5
10
15
20
25
30 Uplink Mode 1 P1>P2
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
0 100 200 300 400 5000
5
10
15
20
25
30
35Uplink Mode 1 and side lobes
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof P1>P2
0 100 200 300 400 5000
5
10
15
20
25
30
35Uplink Mode 2 P1>P2
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
0 100 200 300 400 5000
5
10
15
20
25
30
35
40Uplink Mode 2 and side lobes
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof P1>P2
Figure 7: Military interrogations
The composition of both Intermode C and Mode Sradio load is higher than the combined interrogationrate of civil SSR radars. As a result of this, the uplinkchannel is affected more and more by the increasingequipage grade of ACAS. On the uplink channel onecannot distinguish between military, civil ground sta-tion or ACAS-elicited replies so a definite statementabout the influence of ACAS is not derivable.
CONCLUSIONS
An experimental system capable of receiving andrecording SSR signal-in-space as video raw data wasintroduced in this paper. In context with the describedalgorithms to detect different uplink and downlink tele-grams formats it is now possible to evaluate the SSRradio field load .
Currently all calculations are processed after therecordings are finished. Within the first steps to es-tablish its usage as a radio field monitor it is neces-sary to investigate various methods and algorithmsto improve their performance with respect to a highdetection security. The main aspect is to get reprodu-cable results that are based on proper scientific meth-ods.
In order to generate instant results it is necessary toimplement tested and approved software algorithmsas hardware. Due to the fact that all operations dealwith integer values the implementation could be re-alized with highly integrated Complex ProgrammableLogic Devices (CPLDs). The technical conditions canbe taken for granted nowadays.
The integration of a mobile radio field monitor into
0 100 200 300 400 5000
10
20
30
40
50
60
70
80
90Uplink Mode 3/A P1>P2
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
0 100 200 300 400 5000
20
40
60
80
100Uplink Mode 3/A and side lobes
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof P1>P2
0 100 200 300 400 5000
5
10
15
20
25
30
35
40
45Uplink Mode C P1>P2
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
0 100 200 300 400 5000
10
20
30
40
50
60Uplink Mode C and side lobes
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof P1>P2
Figure 8: Mode 3/A and C interrogations
flight inspection procedures comprises the potentialto get an overview of the current radio field load. Es-pecially in critical air spaces a cyclic SSR flight in-spection would provide a differentiated analysis of thecurrent status and may early detect the reason for areduced reachability of aircraft transponders by SSRground stations due to garbling effects. Furthermore,the analysis of recorded video data could also revealjamming signals which possibly block transpondersand may help to locate the disturbance source. Thesystem’s main benefit is therefore the effective assis-tance of maintaining the backbone of ATC.
ABBREVIATIONS
ACAS Airborne Collision Avoiding SystemACF Autocorrelation Function
A/D Analog to Digital (Conversion)AGC Automatic Gain ControlASR Airport Surveillace RadarsATC Air Traffic ControlCPLD Complex Programmable Logic DeviceDFS Deutsche Flugsicherung GmbHDPSK Differential Phase Shift KeyingFAA Federal Aviation AdministrationFIS Flight Inspection SystemGPS Global Positioning SystemIFF Identification Friend or FoeMTL Minimum Trigger LevelSLS Side Lobe SuppressionSSR Secondary Surveillance Radar
0 100 200 300 400 5000
20
40
60
80
100
120Uplink Intermode C
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
0 100 200 300 400 5000
20
40
60
80
100
120
140Uplink Intermode C and side lobes
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof P1>P2
0 100 200 300 400 5000
50
100
150
200
250
300Uplink Mode S
time [sec]
tele
gram
rat
e [1
0 s−
1 ]
total number thereof > −77dBm
Figure 9: Intermode C and Mode S interrogations
REFERENCES
[1] BREDEMEYER, J.: Funkfeldmessung alsSpezialanwendung der Flugvermessung. In: Or-tung und Navigation (2001), Nr. 1, S. 83–90
[2] ICAO: Annex 10 to the Convention of Inter-national Civil Aviation. Volume IV: SurveillanceRadar and Collision Avoidance Systems. Mon-treal: International Civil Aviation Organization,1995
[3] ICAO: Annex 10 to the Convention of Interna-tional Civil Aviation. Volume I: Radio NavigationAids. Montreal: International Civil Aviation Orga-nization, 1996
[4] ICAO: DOC 8071, Manual on Testing of RadioNavigation Aids. Volume III: Testing of Surveil-lance Radar Systems. Montreal: InternationalCivil Aviation Organization, 1998
[5] LUKE, H. D.: Signalubertragung. 7. Auflage.Berlin : Springer, 1999
[6] MALLWITZ, R. ; WAPELHORST, L. ; PAGANO, T.1030/1090 Megahertz Signal Analysis Frankfurt,Germany. DFS/FAA Report. 1996
[7] STEVENS, M.C.: Secondary Surveillance Radar.Norwood : Artech House, Inc., 1988
