o t & e for esm systems and the use of simulation for...
TRANSCRIPT
O T & E for ESM Systems and the use of simulation for
system performance clarification
Tuesday 11 March 2014 © EW Defence Limited Slide 1 of 52
Dr. Sue Robertson
EW Defence Limited
United Kingdom
e-mail: [email protected]
Tuesday 11 March 2014 © EW Defence Limited Slide 2 of 52
Summary of Presentation
Analysis of Trials Data
Why ESM Testing is needed in an Operational Environment
The Choice of a suitable Trials Area
The simulation of Radar Pulses
Example of the use of Trials Data to identify and solve ESM Performance Issues
The use of trials data for ELINT database and ESM Emitter Library Population
Why OT&E is needed
Tuesday 11 March 2014 © EW Defence Limited Slide 3 of 52
Although testing of ESM systems in a laboratory environment gives useful
indications of system performance, testing in the real world is needed to
determine how the system will work and to aid the optimisation of the system.
Factors such as multipath reflections from the surface of the aircraft and from
objects on the ground affect the performance of the system.
Features of the ESM installation on an aircraft, such as the location of the
antennas also affect system performance.
Common problems with ESM systems such as multi-tracking for single
radars, poor parameter measurement and errors in direction-finding can be
identified and rectified by using test flights in an Operational setting.
© EW Defence Limited Slide 4 of 52
Ideal picture seen by an Electronic Surveillance System
Radar 1
Radar 2
Radar 3
Aircraft
Route
Tracks at correct DOA Fixes with Error Ellipses
Tuesday 11 March 2014
© EW Defence Limited Slide 5 of 52
Actual picture seen by an Electronic Surveillance System
Radar 1
Radar 2
Radar 3
Aircraft
Route
Tracks at correct DOA
Extra Tracks at incorrect DOA
Location Fixes with
Error Ellipses
Tuesday 11 March 2014
Tuesday 11 March 2014 © EW Defence Limited Slide 6 of 52
Planning for a Trial
Simulation of Pulse Data for Comparison with Recorded Trials Data
Selection of a Suitable RF Environment
Definition of Aircraft Route and Altitude Profile
Gathering of Ground truth
Definition of Trials Objective
E.g. To Test Direction Finding Accuracy
To check the Performance of the pulse De-interleaver
To make sure that intercepts are Created for radars that should be seen
Ideally an area where the radar pulse density is low
Find locations and types of land-based radars
Build up a picture of shipping in the trials area
Tuesday 11 March 2014 © EW Defence Limited Slide 7 of 52
AIS Data Showing Ships in Malacca Straight and Singapore
Selection of a Suitable Trials Environment
Tuesday 11 March 2014 © EW Defence Limited Slide 8 of 52
AIS Targets in South China Sea
Selection of a Suitable Trials Environment
Tuesday 11 March 2014 © EW Defence Limited Slide 9 of 52
Air Traffic Control radars in East Malaysia
1. A 60 NM Primary Surveillance Radar (PSR) co-mounted with 200 NM monopulse SSR located on Bukit
Kepayang, 1 NM NE of Kota Kinabalu International Airport (055638.7N 1160352.3E)
2. A 60 NM Primary Surveillance Radar (PSR) co-mounted with 200 NM monopulse SSR located in Kuching
International Airport ( 012848.08 N 1102003.90 E)
3. A 60 NM Terminal Primary Approach Radar co-mounted with a 200 NM monopulse SSR located at Miri Airport
(041929.8N 1135915.8E)
4. A 50 NM Terminal Approach Radar with co-mounted 250 NM conventional SSR located at Labuan Air Force
Base (051526.9N 1150937.7E)
Source: Malaysia Aeronautical Information Publication (AIP)
Data for Ground Truth
Other Airfield radars in East Malaysia & Brunei
Sibu (021523.2N 1115904.4E)
Tawau (041415.4N 1180005.3E)
Bandar Seri Bagawan (4.9442 N 114.9283E)
Sibu Radar Tower
Tuesday 11 March 2014 © EW Defence Limited Slide 10 of 52
Weather Radars
Coastal Surveillance Radar in East Malaysia
There are 3 coastal surveillance radars in East Malaysia – provided as
part of project 1206.
Pulau Balambagan 7.283085N 116.92744E
Pulau Gaya 6.022125N 116.02650E
Pulau antanani 6.707025N 116.34932E
Data for Ground Truth
Kota Kinabalu 5.76472N 116.058E 5 GHz
Kuching 1.36472N 110.337E 5 GHz
Miri 4.21528N 113.988E 3GHz
Sandakan 5.90167N 118.048E 5 GHz
Bintulu 3.12417N 113.019E 5 GHz
There are 5 Weather Radars in East Malaysia
Tuesday 11 March 2014 © EW Defence Limited Slide 11 of 52
Radars in East Malaysia
ATC Weather Radar Coastal Surveillance
Aircraft Flight Profile
Tuesday 11 March 2014 © EW Defence Limited Slide 12 of 52
The best flight profile is one where a repeatable flight profile can be flown, for example
a race track or a box route.
Ideally the aircraft should fly for about 10 minutes on a straight track before turning to
allow the ESM picture to stabilise or to build up a new picture if the system is reset.
Aircraft Speed (kts) Approx Distance (km)
80 25
120 40
200 60
300 90
500 150
1000 300
The distance to be travelled on each straight track of the route will depend on the
aircraft speed.
.
Aircraft Flight Profile
Tuesday 11 March 2014 © EW Defence Limited Slide 13 of 52
Altitude
The altitude at which the trial is flown will determine how far the ESM can “see”.
The radar range (R) is limited by the radar equation
Power Received ∝ 1/R4
The ESM only has to receive pulses that have been transmitted by the radar, so
for the ESM
Power Received ∝1/R2
For example, a radar with a range of 60 nautical miles could be seen at a range
of 3600 nm by an ESM, if is wasn’t for the curvature of the earth.
The radar horizon which increases with altitude is the real governing factor for the
range at which the ESM can see the radar.
Aircraft Altitude (ft) Radar Horizon(km)
200 32
1000 72
2000 102
5000 162
10000 230
20000 320
Tuesday 11 March 2014 © EW Defence Limited Slide 14 of 52
Illustration of Radar Horizon Rings
ATC Weather Radar Coastal Surveillance
Altitude 1000ft,
horizon 75 km
Altitude 2000ft,
horizon 100 km
Altitude 4500ft,
horizon 150 km
Altitude 12000ft,
horizon 250 km
Example of ESM Trial
Tuesday 11 March 2014 © EW Defence Limited Slide 15 of 52
Trial Objectives
To Test the De-interleaver using two of the same type of Coastal Surveillance Radar
To check the creation of an intercept for a radar as it comes into the radar horizon
To check that multi-tracking does not occur
To Test DF performance
Aircraft Parameters
Speed = 200 kts, Maximum Altitude = 20000 ft
Straight line route needed = 60 km
Start of Route = 6.1 N 115.5 E End of Route = 6.5 N 116.0 E
ESM Sensitivity
- 60 dBmi
Range to Radars of Interest
Site Name Range at Route Start (km) Range at Route End (km)
Pulau Gaya 60 53
Pulau Mantanani 115 44
Pulau Balambagan 205 134
Altitude = 5000 ft for radar horizon of 160 km
Tuesday 11 March 2014 © EW Defence Limited Slide 16 of 52
Aircraft Route to Test De-interleaver and Detection Range
Tuesday 11 March 2014 © EW Defence Limited Slide 17 of 52
-65
-60
-55
-50
-45
-40
-35
-30
-25
-20
-4 -3 -2 -1 0 1 2 3 4Azimuth Angle (degrees)
Am
pli
tud
e (
dB
mi)
Range to Radar
= 15nm
3 dB Beamwidth
= 1.5 o
ES Threshold
Sin(x)/x
Beam Shape
1st Sidelobe Level
= -23 dB
The shape of the radar beam determines how many pulses are seen from
the radar each time the radar scans past the ES system
What does the ESM see from each radar?
Tuesday 11 March 2014 © EW Defence Limited Slide 18 of 52
Simulation of Radar Beam Pattern
The inputs to the calculation are the 3dB beamwidth of the radar and the first sidelobe
level.
A radiation pattern of the general form (sin(u)/u)2 is generated, using the following equation
for u:
u = C p sin ( q )
B
where: B is the 3dB beamwidth in radians
C is a factor dependent on the level of the first sidelobe
For sidelobe levels other than -13.2 dB (which is the default for a sin(u)/u pattern), the
radiation pattern equation must be modified by the factor z to give a pattern with the
required sidelobe levels. The actual equation used is:
Power, Aq = sin √ (u2 - z2) 2
√ (u2 - z2)
The factor C ranges from 0.885 when the first sidelobe level is -13.2 dB to about 1.4
when the first sidelobe level is -40 dB.
The factor z ranges from 0 when the first sidelobe level is -13.2 dB to about 5.4 when
the first sidelobe level is -40 dB.
Tuesday 11 March 2014 © EW Defence Limited Slide 19 of 52
Simulation of Radar Beam Pattern
The radiation pattern is altered in two ways by the general equation which provides the
power, Aq.
In addition to the suppression of the sidelobes, the main peak is enhanced in power. The
power must be normalised by subtracting the power at the peak of the beam as follows to
reduce the effects of the sidelobe generation to a zero gain:
A = A0 - Aq
where A0 is the calculated power at the peak of the beam, i.e. q = 0.
The final stage in the calculation of the emitter power is the correction for the peak ERP
of the emitter at the particular range.
Finally, the radar scan period and the pulse repetition intervals are needed to find the
power of each pulse as the radar scans past the ESM.
© EW Defence Limited Slide 20 of 52
Range Pulses % of Radar
Duration Scan Time
200 nm 50 1.25
150 nm 50 1.25
100 nm 66 1.65
50 nm 150 3.75
25 nm 300 7.50
15 nm 1500 37.50
-70
-60
-50
-40
-30
-20
-10
0 50 100 150 200 250 300 350
Am
pli
tud
e (
dB
)
Time (ms)
50 nm
150 ms
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0 50 100 150 200 250 300 350
Am
pli
tud
e (
dB
)
Time (ms)
25 nm
300 ms
-70
-60
-50
-40
-30
-20
-10
0
0 100 200 300
Am
pli
tud
e (
dB
)
Time (ms)
100 nm
66 ms
Tuesday 11 March 2014
Beam Patterns for Different Ranges
Tuesday 11 March 2014 © EW Defence Limited Slide 21 of 52
Time-line of All radars in trial
0
1
2
3
4
5
6
7
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000
Em
itte
r in
Sc
en
ari
o
Time (ms)
Labuan ATC
KK ATC
KK Weather
Palau Balambagan
Palau Mantanani
Palau Gaya
Tuesday 11 March 2014 © EW Defence Limited Slide 22 of 52
Predicted Amplitude of all radars Showing Pulse Density
in one minute of flight
-60
-55
-50
-45
-40
-35
-30
-25
-20
0 10000 20000 30000 40000 50000 60000
Pu
lse
Am
pli
tud
e (
dB
)
Time (ms)
KK WeatherRadarKK ATC
Labuan ATC
PulauBalambaganPulauMantananiPulau Gaya
(~ 4000 pulses from 6 radars)
Tuesday 11 March 2014 © EW Defence Limited Slide 23 of 52
Detail of Predicted Amplitude Profiles
-60
-55
-50
-45
-40
-35
-30
-25
-20
35000 36000 37000 38000 39000 40000
Pu
lse
Am
pli
tud
e (
dB
)
Time (ms)
KK WeatherRadarKK ATC
Labuan ATC
PulauBalambaganPulauMantananiPulau Gaya
Tuesday 11 March 2014 © EW Defence Limited Slide 24 of 52
Fine Detail of Predicted Amplitude Profiles
-60
-55
-50
-45
-40
-35
-30
-25
-20
37300 37400 37500 37600 37700 37800 37900 38000
Pu
lse
Am
pli
tud
e (
dB
)
Time (ms)
KK WeatherRadarKK ATC
Labuan ATC
PulauBalambaganPulauMantananiPulau Gaya
Tuesday 11 March 2014 © EW Defence Limited Slide 25 of 52
Analysis of Trials Data
Graphical Representation of Pulse Data
Visualisation / Playback of Operator’s Screen
Calculation and Plot of Correct DOA Profiles
Graphical representation of “Track” Data
Comparison of Pulse Data with Simulation
Tuesday 11 March 2014 © EW Defence Limited Slide 27 of 52
DOA and Range Profiles of Radars of Interest
0
30
60
90
120
150
180
210
0 100 200 300 400 500 600
DO
A (
de
gre
es
) Pulau Gaya
PulauMantanani
PulauBalambagan
0
40
80
120
160
200
0 100 200 300 400 500 600
Ran
ge
(km
)
Time (seconds)
© EW Defence Limited Slide 28 of 52
The analysis of ESM Track Data
Tuesday 11 March 2014
0
20
40
60
80
100
120
140
160
180
200
0 100 200 300 400 500 600
DO
A (
de
gre
es
)
Time (seconds)
Pulau Gaya
Pulau Mantanani
Pulau Balambagan
53km
60km
115 km
205 km134 km
44 km
To see the DOA performance of the ESM the correct DOA of the radar is plotted (found by using
navigation data from the trials platform) and the DOA of Track Updates created by the ESM are
shown on the same graph.
© EW Defence Limited Slide 29 of 52
The analysis of ESM Track Data
Tuesday 11 March 2014
50
51
52
53
54
55
56
57
58
59
60
300 310 320 330 340 350 360 370 380 390 400
DO
A (
de
gre
es
)
Time (seconds)
This graph shows the tracks created for two of the radars, with the track for Palau Balambagan
appearing as the radar comes into the radar horizon of the aircraft.
Palau Balambagan
Palau Mantanani
© EW Defence Limited Slide 30 of 52
The analysis of ESM Pulse Data
Tuesday 11 March 2014
Each of the updates to the ESM tracks is created from sets of pulses. In this plot the correct
DOA of the radar is plotted and the DOAs of individual pulses are plotted on the same graph.
0
20
40
60
80
100
120
140
160
180
200
0 100 200 300 400 500 600
DO
A (
de
gre
es
)
Time (seconds)
Pulau Mantanani
Pulau Balambagan
Pulau Gaya
Pulau Gaya
Pulau Mantanani
Pulau Balambagan
53km
60km
115 km
205 km
134 km
44 km
© EW Defence Limited Slide 31 of 52
The Detail of the Pulse Data
Tuesday 11 March 2014
In this graph the pulses from individual scans of each of the 3 radars can be seen.
40
60
80
100
120
140
160
370 375 380 385 390 395 400
DO
A (
de
gre
es
)
Time (seconds)
Pulau Mantanani
Pulau Balambagan
Pulau Gaya
Pulau Gaya
Pulau Mantanani
Pulau Balambagan
53km60km
115 km
205 km 134 km
44 km
Tuesday 11 March 2014 © EW Defence Limited Slide 32 of 52
Comparison of Theoretical Beam Shape with Received Pulses
Theoretical Amplitude
Profile
Example of Amplitude
Profile from Pulse Data
-65
-60
-55
-50
-45
-40
-35
-30
0 5 10 15 20 25
Am
pli
tud
e (
dB
mi)
Time (ms)
Number of Pulses = 24
Range to Radar = 60 km
-65
-60
-55
-50
-45
-40
-35
-30
0 5 10 15 20 25
Am
pli
tud
e (
dB
mi)
Time (ms)
Number of Pulses = 23
Range to Radar = 60 km
Tuesday 11 March 2014 © EW Defence Limited Slide 33 of 52
Comparison of Theoretical Beam Shape with Received Pulses
Theoretical Amplitude
Profile
Example of Amplitude
Profile from Pulse Data-56
-54
-52
-50
-48
-46
-44
-42
-40
0 2 4 6 8 10
Am
pli
tud
e (
dB
mi)
Time (ms)
Number of Pulses = 11
Range to Radar = 160 km
-56
-54
-52
-50
-48
-46
-44
-42
-40
0 2 4 6 8 10
Am
pli
tud
e (
dB
mi)
Time (ms)
Number of Pulses = 11
Range to Radar = 160 km
radar2
3
1
4
5
Tuesday 11 March 2014 © EW Defence Limited Slide 34 of 52
Route taken by aircraft in trial
Example of Use of Pulse Data to diagnose DF / Multi-tracking
4
radar
1
3
5
2
Tuesday 11 March 2014 © EW Defence Limited Slide 35 of 52
Track formation during the Trial
Tuesday 11 March 2014 © EW Defence Limited Slide 36 of 52
Pulse data during a single run of the flight trial
Tuesday 11 March 2014 © EW Defence Limited Slide 37 of 52
Single scans of an ATC Radar showing the slope in DOA
across the main beam
Tuesday 11 March 2014 © EW Defence Limited Slide 38 of 52
ESM Antenna Installation on a Large Aircraft similar to the antenna
configuration in the trial where the DOA spread was first seen
Amplitude Comparison DOA Measurement
Tuesday 11 March 2014 © EW Defence Limited Slide 39 of 52
The beam shape for the ESM antennas can be represented by a Gaussian model with a nominal 3dB beamwidth of
about 70o.
The amplitude difference between the mth and either the (m+1)th port or the (m-1)th port can be used to estimate the
direction of arrival of the signal.
The assumption in the operation of an Amplitude Comparison ESM system is that the emitted pulse arrives with
equal amplitude at all receiving antennas, but that the detected amplitude at each antenna depends on the
receiver antenna beam shape.
The DOA of the pulse can therefore be determined from the difference in amplitude of the receiving antennas.
Typical ESM Receiver Antenna
Beam Shapes
with Activation Levels for a
Correct DOA of 70o
Tuesday 11 March 2014 © EW Defence Limited Slide 40 of 52
Typical Beam Shape of an Air Traffic Control (ATC) Radar
Tuesday 11 March 2014 © EW Defence Limited Slide 41 of 52
Typical Beam Shape of an ATC Radar with
overlay of Aircraft Azimuth Extent
Tuesday 11 March 2014 © EW Defence Limited Slide 42 of 52
DOA Errors across the main beam of a scanning radar as an aircraft with
30m ESM antenna separation moves past the radar at a range of 10 nm
Tuesday 11 March 2014 © EW Defence Limited Slide 43 of 52
Theoretical DOA Error Slope due to Antenna Separation
-65
-60
-55
-50
-45
-40
-35
-30
0 5 10 15 20 25 30
Am
pli
tud
e (
dB
mi)
-20
-10
0
10
20
0 5 10 15 20 25 30
DO
A E
rro
r (
deg
ree
s)
Time (ms)
Tuesday 11 March 2014 © EW Defence Limited Slide 44 of 52
Using Trials Data for ELINT Databases
Traditionally an Emitter Type would have been specified in the Emitter Libraries
for ESM Systems, rather having specific parameters for individual Emitter sites.
For example,
A typical Air Traffic Control (ATC)Radar may have the following parameter ranges
specified for it in the Emitter Library:
RF = 2700 to 3100 MHz, PRI = 1000 to 2000 ms, PW = 1 to 2 ms
A typical Marine Radar may have the following parameter ranges:
RF = 3040 to 3065 MHz, PRI = 1200 to 1300 ms, PW = 0.7 to 1.2 ms
Already we can see ambiguity.....
© EW Defence Limited Slide 45 of 52
Example of the Overlap in Parameter ranges for some
ATC/Marine radars
0
500
1000
1500
2000
2500
0 500 1000 1500 2000 2500 3000
pu
lse
wid
th (
ns
) .
PRI (ms)
Tuesday 11 March 2014
Tuesday 11 March 2014 © EW Defence Limited Slide 46 of 52
Using Trials Data for ELINT Databases
There is a trend towards using specific emitter parameters in Emitter
Libraries.
For example, two ATC radars of the same type would be specified
individually in the library
Radar at site A: RF = 2765 MHz, PRI = 1333ms, PW = 1.2ms
Radar at Site B: RF = 2945 MHz, PRI = 1555ms, PW = 1.5ms
Pulse data is needed to aid the specification and testing of the Emitter
Library.
.
Tuesday 11 March 2014 © EW Defence Limited Slide 47 of 52
Interleaved Radar Pulses
Here is an example of what pulses received from typical set of ship radars in
250 ms may look like – even in a busy RF environment it is rare to get more
than 3 or 4 radar pulse streams interleaved at any one time.
-70
-60
-50
-40
-30
-20
0 50 100 150 200 250
Pu
lse
Am
pli
tud
e (
dB
mi)
Time (ms)
Tuesday 11 March 2014 © EW Defence Limited Slide 48 of 52
De-interleaved Radar Pulses
It is therefore straightforward to de-interleave the radar pulses off-line from
the trial and extract sets of pulses for individual radars for use in the ELINT
database.
-70
-60
-50
-40
-30
-20
0 50 100 150 200 250
Pu
lse
Am
pli
tud
e (
dB
mi)
Time (ms)
Tuesday 11 March 2014 © EW Defence Limited Slide 49 of 52
Example of Pulses received from a Fixed PRI radar
and corresponding ELINT data set
-70
-60
-50
-40
-30
-20
0 50 100 150 200 250
Time (ms)
Radar Pulses
Pulse Amp1 TOA2 PRI3 RF4 PW5
(dBmi) (ms) (ms) (MHz) (ms)
1 -54.7 101.65 3045.1 0.70
2 -52.5 102.98 1331.831 3045.1 0.70
3 -48.5 104.32 1331.831 3045.2 0.71
4 -48.5 105.65 1331.830 3045.0 0.71
5 -47.0 106.98 1331.832 3045.2 0.70
6 -46.2 108.31 1331.831 3045.2 0.70
7 -47.0 109.64 1331.831 3045.1 0.70
8 -48.5 110.98 1331.831 3045.3 0.70
9 -49.5 112.31 1331.832 3045.2 0.70
10 -52.5 113.64 1331.831 3045.3 0.71
11 -54.0 114.97 1331.834 3045.2 0.70
12 -57.0 116.30 1331.833 3045.2 0.70
13 -59.5 117.63 1331.833 3045.2 0.72
Specific EmitterParameters
RF = 3045 MHz PRI Type = Fixed
PW = 0.7 (ms) PRI Element = 1331.83 (ms)
Scan Period = 1.8 seconds 3dB beamwidth = 2 degrees
1 Amplitude
2 Time of Arrival
3 Pulse Repetition Interval
4 Radar Frequency
5 Pulse Width
Tuesday 11 March 2014 © EW Defence Limited Slide 50 of 52
Example of Pulses received from a Staggered PRI
radar and corresponding ELINT data set
-70
-60
-50
-40
-30
-20
0 50 100 150 200 250
Time (ms)
Radar Pulses
Pulse Amp1 TOA2 PRI3 RF4 PW5
(dBmi) (ms) (ms) (MHz) (ms)
1 -59.0 215.20 9410.1 0.30
2 -55.7 216.87 1673.232 9410.0 0.31
3 -53.7 218.56 1685.516 9410.2 0.30
4 -51.7 220.26 1697.217 9410.1 0.30
5 -49.5 221.93 1673.232 9410.1 0.30
6 -47.0 223.61 1685.516 9410.2 0.30
7 -46.7 225.31 1697.217 9410.2 0.31
8 -45.5 226.99 1673.233 9410.1 0.30
9 -44.7 228.67 1685.516 9410.1 0.30
10 -44.7 230.37 1697.217 9410.0 0.30
11 -45.5 232.04 1673.232 9410.1 0.30
12 -46.2 233.73 1685.516 9410.0 0.30
13 -47.0 235.42 1697.217 9410.1 0.30
14 -51.7 237.10 1673.232 9410.0 0.31
15 -54.0 238.78 1685.517 9410.2 0.30
Specific Emitter Parameters
RF = 9410 MHz PRI Type = Staggered
PW = 0.3 (ms) PRI Elements =1673.232, 1685.516, 1697.217(ms)
Scan Period = 1.5 seconds 3dB beamwidth = 1.8 degrees
1 Amplitude
2 Time of Arrival
3 Pulse Repetition Interval
4 Radar Frequency
5 Pulse Width
Tuesday 11 March 2014 © EW Defence Limited Slide 51 of 52
Conclusions
O T & E trials are needed to check aspects of ESM System performance, such as
de-interleaving, detection range and DF accuracy in the real world.
Use of simulation of radar beam shapes based on actual ground truth allows a
comparison to be made with data recorded during a trial.
Issues such as multi-tracking and DOA errors can be diagnosed by looking at
recorded pulse data.
Recorded radar pulse data can be used to populate ELINT databases and for the
creation of ESM radar libraries containing specific Emitter Identities.
Tuesday 11 March 2014 © EW Defence Limited Slide 52 of 52
Dr Sue Robertson, e-mail: [email protected]
Nimrod MRA4