Analysis of first generation MANPAD attacks on fast jets

James Jackman
a
, Mark Richardson*
a
, Brian Butters
b
, Roy Walmsley
b
, Nic Millwood
b
, Peter Yuen
a
,
David James
a

a
Dept. of Informatics & Sensors, Cranfield University, Defence Academy of the UK, Shrivenham,
Swindon, SN6 8LA;
b
Chemring Countermeasures Ltd, High Post, Salisbury, Wiltshire, SP4 6AS
ABSTRACT
The proliferation of early generation Man-Portable Air-Defence (MANPAD) weapon worldwide results in a significant
threat to all aircraft. To develop successful countermeasures to the MANPAD a more detailed understanding of the
factors affecting the missile engagement is needed. This paper discusses the use of CounterSim, a missile engagement
and countermeasure simulation software tool, to model such scenarios. The work starts by analysing simple engagements
of a first generation MANPAD against a fast jet with no countermeasures being employed. The engagement simulations
cover typical MANPAD ranges and aircraft altitudes quoted in open source literature. From this set of base runs,
individual engagements are chosen for further analysis. These may have resulted in hits, misses or near misses. At each
time interval in the simulation the aircraft and missile velocities are used to calculate a projected point of closest
approach. This is then compared with the simulated impact point. The difference is defined as the d error and plots are
produced for hits, misses and near misses. Features of the d error plots are investigated to gain insights into the
potential countermeasure capability. Finally, the analysis of the d error plots is used to investigate the possibility of
replicating the factors in a simulation that produce a miss through a pre-emptive flare deployment.
Keywords: Countermeasure, pre-emptive, simulation, MANPAD
1. INTRODUCTION
The standard response by a fast jet, once alerted by the Missile Approach Warning (MAW) system to a MANPAD, is to
reactively deploy flares and, in some cases, perform a banking manoeuvre. In this instance the missile has been fired and
there is always the possibility of the target aircraft being hit. The ideal solution is to defeat the infrared seeker pre-
emptively and thereby prevent the missile launch. This presents a greater challenge and, in order to understand how to
achieve this objective, the factors affecting an engagement outcome need to be studied in more detail. This paper reports
the early work in the simulation of attacks of a first generation MANPAD on a fast jet and an analysis of the results as a
precursor to developing a pre-emptive countermeasure strategy.
2. MODELLING
The software used to model the missile engagement is called CounterSim. This is designed and developed by Chemring
Countermeasures and allows the user to specify the type of missile, aircraft and environment in which the simulation
takes place. Parameters have been taken from open source literature in order to model a first generation MANPAD. A
generic fast jet model has been used with an engine plume signature taken from the open literature.
2.1 Missile parameters
The type of missile modeled is a first generation MANPAD with an AM spin scan seeker and the following parameters;
The missile body has a diameter of 70mm, mass 9.6kg, lateral acceleration (latax) limit 15g, velocity factor
580m/s
[1]
and a drag coefficient of 0.3
[2]
.
The seeker has a FOV of 1.9
o
, gimbal rate of 6
o
/s, detector waveband of 2-2.7m and a rising sun reticle with
12 spokes
[1]
.
*M.A.Richardson@cranfield.ac.uk
Technologies for Optical Countermeasures VI, edited by David H. Titterton, Mark A. Richardson,
Proc. of SPIE Vol. 7483, 74830I 2009 SPIE CCC code: 0277-786X/09/$18 doi: 10.1117/12.829464
Proc. of SPIE Vol. 7483 74830I-1


At the start of the simulation the horizontal distance of the MANPAD operator from the aircraft ranges from 1km to
5.5km. The aircraft altitude ranges from 100m to 3km. The azimuth angle of the MANPAD operator is varied over 360
o

in 15
o
steps. The timings for the ignition delay, boost thrust and sustain thrust were taken from a video of a soldier firing
an SA-7
[3]
.
2.2 Aircraft parameters
For convenience a 3D model of an AMX-A1 aircraft is used. The aircraft engine plume shown in Figure 1 is modelled as
3 concentric cones all with transparency of 0.5 and with temperatures set at 371
o
C, 149
o
C and 66
o
C
[4]
. The aircraft with
its engine plume is therefore an unclassified, generic fast jet model and results should not be interpreted as representative
of the AMX-A1. In Figure 1, the aircraft is lit from the side to show the 3D model more clearly.

Fig. 1. View of the aircraft and plume contours.
The aircraft was set on a constant bearing at a speed of 200m/s and released no countermeasures and performed no
manoeuvre during the base run simulations.
2.3 Simulation
To simulate a missile engagement in CounterSim a number of items or models have to be specified in an appropriate
hierarchy and then parameters specified for each item. Figure 2 shows the scenario model hierarchy used for the
simulations.

Fig. 2. CounterSim simulation item hierarchy.
Previously MANPADS have been modelled in CounterSim with the scenario starting either at launch or with a missile
assumed to be already in flight. In order to model the full pre-launch phase of an engagement it was necessary to identify
and add some additional options - particularly to the existing CounterSim Missile System item. The MANPAD system
parent item is the Tracker which models the operator of the MANPAD tracking the target to obtain lock-on. The Tracker
uses an Alpha, Beta, Gamma track filter. The Designator merely identifies an object to track in this case the Aircraft.
Once the Tracker is tracking the Aircraft and the Generic Seeker is locked on, the advanced options in the Missile
Proc. of SPIE Vol. 7483 74830I-2


System simulate the different stages in the firing sequence. The seeker is uncaged at a user defined time after lock-on.
Next, the lead and super elevation angles are applied. These are either fixed or are related to the calculated crossing rate.
The missile is then launched at a set time after applying the launch angles assuming that lock-on is maintained. Pre
launch horizontal and vertical gimbal rates set in the Generic Seeker affect the time taken to fire the missile. Maximum
pre launch gimbal limits are also set and therefore if these are exceeded the seeker will lose lock. Preset logic choices in
the Generic seeker may cause the seeker to be re-caged until track is assumed or the seeker may remain uncaged. When
track is resumed the lead and super elevation angles are reapplied and the launch is again attempted.
The parameters used for the missile system were to uncage the seeker 0.1s after lock-on, apply a super elevation of 5
o
at
a rate of 6
o
/s and launch 0.1s after super elevation. A test run was undertaken for one altitude to study the effect of
applying a lead angle based on the crossing rate. Taking the gimbal rate and the time for the firing sequence the
maximum possible lead angle applied in a beam on engagement was estimated to be 10
o
. This value was used to
calculate the lead angle for each aircraft azimuth based on the sinusoidal law 1u - sin (18u
o
-Aiiciaft Azimuth). The
results of this test were exactly the same as when no lead angle was applied. With first generation missiles the open
source value for gimbal rates is low and is quoted as 6/s. This limits the lead angle that can be applied in the time it
takes to fire.
The base run simulations did not include any atmospheric attenuation or noise. This was to give the best possible results
that a first generation MANPAD could achieve. The results are shown in the wheel plot of Figure 3. Each point of the
wheel plot represents the distance and azimuth of the missile system to the target. A hit is defined as a miss distance
between 0m and 2m, near miss is between 2m and 10m and a miss is more than 10m.

Fig. 3. Results of the base run simulation for an altitude of 1000m with no noise.
To investigate the effect of noise on the outcome of the simulations a low level of noise was included in the seeker. The
dark noise included is assumed to be random with a normal distribution about a mean of zero with a standard deviation
equal to 1. The units are in photons and the noise is signal independent. The results are shown in Figure 4 and are
consistent with a first generation MANPAD being limited to a tail on engagement at a maximum range of 3.5km.
Proc. of SPIE Vol. 7483 74830I-3



Fig. 4. Results of the base run simulation for an altitude of 1000m with noise included in the seeker.
Table 1 shows the percentage hit rates of the base run simulations for each altitude. For the whole of the base run
simulation the average hit rate was 75%.
Table 1. Percentage hit rates for each altitude of the base run simulations with no countermeasures.
Altitude m Hit Rate % Altitude m Hit Rate % Altitude m Hit Rate %
100 0 1100 80.4 2100 90.0
200 5.0 1200 82.1 2200 90.8
300 15.4 1300 83.8 2300 90.4
400 40.0 1400 84.2 2400 90.0
500 57.1 1500 86.3 2500 89.2
600 67.9 1600 86.7 2600 87.9
700 75.0 1700 87.1 2700 89.6
800 77.5 1800 87.5 2800 88.8
900 80.4 1900 89.2 2900 87.9
1000 80.8 2000 89.2 3000 88.8
3. ANALYSIS
From the base run simulations individual scenarios have been chosen to analyse in greater detail. For each time step the
velocity vectors of the missile and aircraft are used to calculate the projected point of closest approach. This is then
compared with the actual impact point and the difference is defined as the projected miss distance d.
3.1 Calculation of d
The aircraft travels at a constant speed and on a fixed bearing. The missile is launched at the aircraft which uses
proportional navigation guidance. At time t
1
the aircraft is at A
1
and the missile at M
1
. Their velocity vectors intersect at
point P
1
. The difference between P
1
and the actual impact point of a hit or point of closest approach for a miss is defined
as the projected miss distance, d
1
. Likewise, at a later time t
2
, the projected miss distance is d
2
. d is calculated for
each time interval and a graph produced of d against time.
Proc. of SPIE Vol. 7483 74830I-4



Fig. 5. Shows the definition of d and how it is calculated.
It is unlikely that two lines in three dimensions will intersect exactly, however they can be connected by a unique
shortest line segment
[5]
. Given two lines in three dimensions with endpoints p
1
, p
2
and p
3
, p
4
, as shown in Figure 6, a
point on the two lines will be defined by the following two equations
p
u
= p
1
+p
u
(p
2
-p
1
) (1)
p
b
= p
3
+ p
b
(p
4
- p
3
). (2)
Because the shortest line segment will be perpendicular to the two lines two equations can be written for the dot product
and expanded to solve for
a
and
b

(p
u
- p
b
) (p
2
-p
1
) = u (3)
(p
u
- p
b
) (p
4
-p
3
) = u. (4)

Fig. 6. Shortest line segment between two lines in three dimensions.
The point to use for the projected hit point has to be chosen. Point p
a
will lie on the aircraft velocity vector, p
b
which will
lie on the missile velocity vector or the midpoint of the line segment p
a
p
b
can be used. The results are nearly identical
and because the aircraft is on a constant trajectory in 2 dimensions the point p
a
has been chosen for the projected hit
point.
3.2 Graphs of d
Graphs of d can be classified into different types depending on the angle of attack of the missile. For tail on
engagements that result in a hit, shown by the curve in Figure 7, the graphs are similar in shape. When the missile misses
it is due to the fact that the aircraft is too far away to lock-on to the signal. For head on engagements all the misses occur
because the missile is unable to lock-on to the target. This is because of the geometry of the scenario whereby the seeker
cannot see the plume of the jet and no signal is detected. The graphs of d for head on hits all closely resemble the
example in Figure 8.
Proc. of SPIE Vol. 7483 74830I-5



Fig. 7. Graph of d for a tail on engagement with coordinates distance 3500m, altitude 500m and azimuth 0
o
.


Fig. 8. Graph of d for a head on engagement with coordinates distance 3000m, altitude 1000m and azimuth 180
o
.

-5000
-4000
-3000
-2000
-1000
0
1000
2000
3000
4000
0 2 4 6 8 10 12 14 16
D
i
s
t
a
n
c
e

m
Time s
-1800
-1600
-1400
-1200
-1000
-800
-600
-400
-200
0
0 1 2 3 4 5 6 7 8
D
i
s
t
a
n
c
e

m
Time s
Proc. of SPIE Vol. 7483 74830I-6



Fig. 9. Graph of d for a beam on engagement with coordinates distance 1500m, Altitude 1500m and azimuth 120
o
.
To study the differences in plots of d for hits and misses it is necessary to look at beam on engagements. In these cases
it is possible that the missile misses due to the limitations of its design and not due to a lack of signal from the target.


Fig. 10. Graph of d for a beam on engagement with coordinates distance 2000m, Altitude 1500m and azimuth 120
o
.
-1200
-1000
-800
-600
-400
-200
0
0 1 2 3 4 5 6 7
D
i
s
t
a
n
c
e

m
Time s
-1200
-1000
-800
-600
-400
-200
0
0 1 2 3 4 5 6 7 8
D
i
s
t
a
n
c
e

m
Time s
Proc. of SPIE Vol. 7483 74830I-7


Figures 9 and 10 show plots of d for two engagements where the only difference is the missile firing points 1500m and
2000m from the aircraft. The simulations ended with different results. The two graphs are similar up until 4s into the
simulation. Figure 9, the plot for a miss shows a continuous smooth curve and d never reaches zero whereas Figure 10,
the plot for a hit shows small fluctuations before attaining a zero miss distance.

3.3 Missile acceleration
The missile body and seeker behaviour during each simulation was examined to determine which factors most affected
the outcome of an engagement. The most significant factor was the missile acceleration.
Figure 11 shows a beam on engagement for a distance of 1500m, altitude 1500m and aircraft azimuth 120
o
that results in
a miss. The lateral acceleration shows a steady increase as the missile tries to use proportional navigation to aim ahead
of the aircraft current position. Once the missile reaches its latax limit it can no longer keep the target in the seeker FOV
and so loses signal. This can also be seen in the graph of d, Figure 9, which exhibits a smooth curve.
Figure 12 shows a beam on engagement for a distance of 2000m, altitude 1500m and aircraft azimuth 120
o
that results in
a hit. The lateral acceleration is the same until 4s into the simulation. At this time the missile starts alternating between
zero and the latax limit as it goes through a process of acquiring the target, then the target moving to the edge of the FOV
and finally reacquires the target in the centre of the seeker view. The large fluctuations in the lateral acceleration appear
as small fluctuations in the graph of d, Figure 10.


Fig. 11. Vertical and horizontal acceleration for a beam on engagement that resulted in a miss.

Proc. of SPIE Vol. 7483 74830I-8



Fig. 12. Vertical and horizontal acceleration for a beam on engagement that resulted in a hit.
3.3 Pre-emptive flare deployment
For the pre-emptive flare deployment scenario, two flares were fired at the start of the simulation when the tracker is
trying to lock-on to the target. The examples in Figures 13 and 14 show the differences in the plot of d for a hit and a
miss caused by pre-emptive flare deployment. Percentage hit rates were also calculated for each altitude, Table 2,
showing the effectiveness of pre-emptive flare deployment. The hit rate for the whole of the pre-emptive flare
deployment scenario was an average of 8.4%. With no noise and an altitude of 1000m the hit rate was 9.2%. When noise
was included in the seeker the hit rate dropped to 0.8%. It was hoped that pre-emptive flares would deny lock-on of the
target and push the seeker gimbal beyond its limits and/or rates keeping the missile in the tube. However, the missile
locked onto the flare and still launched which might be due to limitations in the design of the signal processor.
Table 2. Percentage hit rates for each altitude when two pre-emptive flares are deployed.
Altitude m Hit Rate % Altitude m Hit Rate % Altitude m Hit Rate %
100 0 1100 11.3 2100 10.8
200 0 1200 10.8 2200 10.4
300 2.9 1300 11.3 2300 8.3
400 4.6 1400 11.7 2400 7.1
500 5.4 1500 12.1 2500 7.1
600 7.1 1600 12.5 2600 7.5
700 6.3 1700 12.9 2700 7.5
800 8.3 1800 11.7 2800 6.7
900 8.3 1900 12.9 2900 6.7
1000 9. 2 2000 12.1 3000 7.5
Proc. of SPIE Vol. 7483 74830I-9





Fig. 13. Graph of d for an engagement with coordinates distance 2500m, altitude 1000m and azimuth 195
o
.

Fig. 14. Graph of d for same engagement but with two pre-emptive flares deployed.
-1200
-1000
-800
-600
-400
-200
0
0 1 2 3 4 5 6 7
D
i
s
t
a
n
c
e

m
Time s
-1200
-1000
-800
-600
-400
-200
0
0 1 2 3 4 5 6
D
i
s
t
a
n
c
e

m
Time s
Proc. of SPIE Vol. 7483 74830I-10


In reality pre-emptive flare deployment will not occur at exactly the same time as the aircraft is being tracked. There will
be a firing sequence of flares at time intervals governed by the number of flares onboard, the size of area in which there
is a threat and the burn time of the flare. A scenario was set up where an aircraft flew along a constant trajectory at a
speed of 200m/s at an altitude of 1000m firing two flares at 4 second intervals after the start of the run. The aircraft
travelled a distance of 12km and fired 32 flares. MANPADs were positioned with 1km spaces in a grid of 10km by
10km, represented by crosses in Figure 15. The aircraft flew across the centre of the grid as the MANPADs tried to track,
lock-on, launch the missile and hit the target. None of the 121 simulations resulted in a hit and there were no near misses.

Fig. 15. Grid pattern of missile systems.
4. CONCLUSIONS AND FUTURE WORK
This work has explored a means of simulating and assessing pre-emptive countermeasures against a first generation
MANPAD. In deriving the simulation method, improvements have been made to the CounterSim Tracker and Missile
System models to simulate the track, uncage, lead angle and super elevation phases and the recovery process that occurs
when preemptive flares cause the seeker to fail to acquire the aircraft or lose lock.
A set of base run simulations was carried out to test the effectiveness of a first generation seeker against a fast jet with no
countermeasures and in ideal circumstances with no noise and no transmission loss. The model was also tested with
noise included in the missile seeker and showed results typical of a first generation MANPAD. The projected miss
distance d was defined and introduced as a new variable to analyse individual scenarios. Differences in plots of d were
only noted after 4 seconds into beam on engagements indicating that d can only be used to classify the angle of attack,
but not the result of an engagement.
Pre-emptive flares were investigated, starting with two fired at the beginning of the simulation when the tracker is trying
to lock-on to the aircraft. This reduced the average hit rate from 75% to 8.4%. To test a large number of possible
MANPAD positions, pre-emptive flares were fired in a timed sequence as the aircraft flew through an area 10km square.
Proc. of SPIE Vol. 7483 74830I-11


The result of no hits indicates that a sequence of pre-emptive flares is more effective than reactive single flares fired after
missile launch.
So far, simulations have assumed that the MANPAD operator always applies a single value of lead angle and super
elevation. The facility has been included to model the imperfect operator with lead angle and super elevation selected at
random within a prescribed statistical distribution.
Future work will be to further test the pre-emptive flare concept against first generation and more advanced MANPADs
with counter countermeasure (CCM) features. A range of pre-emptive flares with different IR outputs will be tested and
the differences due to the imperfect operator will be tested. Results will also be compared with reactive flare deployment.
REFERENCES
[1]
http://www.fas.org/man/dod-101/sys/missile/row/sa-7.htm
[2]
http://exploration.grc.nasa.gov/education/rocket/shaped.html
[3]
http://www.youtube.com/watch?gl=GB&hl=en-GB&v=nkld7L_pvz4&NR=1
[4]
R. D. Hudson, Infrared Systems Engineering, Wiley, 1969.
[5]
http://local.wasp.uwa.edu.au/~pbourke/geometry/lineline3d/


Proc. of SPIE Vol. 7483 74830I-12