Decoy Simulation

NAVAL
POSTGRADUATE
SCHOOL
MONTEREY, CALIFORNIA
THESIS
A LOW-FIDELITY SIMULATION TO EXAMINE THE

DESIGN SPACE FOR AN EXPENDABLE ACTIVE
DECOY
by
Devon A. Cartwright
December 2017
Thesis Advisor: Alejandro Hernandez

Co-Advisor: Oleg Yakimenko
Second Reader: Brian Connett
Approved for public release. Distribution is unlimited.

THIS PAGE INTENTIONALLY LEFT BLANK
REPORT DOCUMENTATION PAGE Form Approved OMB
No. 0704–0188
Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing
instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of
information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions
for reducing this burden, to Washington headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson
Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project
(0704-0188) Washington, DC 20503.
1. AGENCY USE ONLY 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
(Leave blank) December 2017 Master’s thesis
4. TITLE AND SUBTITLE 5. FUNDING NUMBERS
A LOW-FIDELITY SIMULATION TO EXAMINE THE DESIGN SPACE FOR
AN EXPENDABLE ACTIVE DECOY
6. AUTHOR(S) Devon A. Cartwright
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) 8. PERFORMING
Naval Postgraduate School ORGANIZATION REPORT
Monterey, CA 93943-5000 NUMBER
9. SPONSORING /MONITORING AGENCY NAME(S) AND 10. SPONSORING /
ADDRESS(ES) MONITORING AGENCY
N/A REPORT NUMBER
11. SUPPLEMENTARY NOTES The views expressed in this thesis are those of the author and do not reflect the
official policy or position of the Department of Defense or the U.S. Government. IRB number ____N/A____.
12a. DISTRIBUTION / AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
13. ABSTRACT (maximum 200 words)
A new type of expendable radar decoy, one that leverages the advances in technology of the last 50
years, is required to improve the survivability of non-stealth aircraft against radar-guided threats. This thesis
applies modeling and simulation based systems engineering to explore the design space of a notional active
expendable decoy to determine the combination of key performance parameters that will maximize
probability of survival. The thesis focuses on development of a Microsoft Excel-based simulator, which
accepts inputs from the user, uses a nearly orthogonal Latin hypercube experimental design, and performs
multiple simulator runs with stochastic model variables. The final output includes statistical analysis of
results to provide the user with the best combination of design variables to improve survivability against
specified threat parameters. Results show that either low power with high gain, or high power with low gain,
was the best combination for maximizing probability of survival. In addition, low altitude deployments are
to be avoided unless the missile is very near the target aircraft. The results provided by the tool are notional,
based on realistic approximations of radar, aircraft, and decoy systems. The results will inform the
acquisition strategy of NAVAIR.
14. SUBJECT TERMS 15. NUMBER OF

expendable active decoy, EAD, radar countermeasures, modeling and simulation based PAGES
systems engineering, MSBSE 99
16. PRICE CODE
17. SECURITY 18. SECURITY 19. SECURITY 20. LIMITATION
CLASSIFICATION OF CLASSIFICATION OF THIS CLASSIFICATION OF OF ABSTRACT
REPORT PAGE ABSTRACT
Unclassified Unclassified Unclassified UU
NSN 7540–01-280-5500 Standard Form 298 (Rev. 2–89)
Prescribed by ANSI Std. 239–18
i
ii
A LOW-FIDELITY SIMULATION TO EXAMINE THE DESIGN SPACE FOR

AN EXPENDABLE ACTIVE DECOY
Devon A. Cartwright
Lieutenant, United States Navy
B.S., University of Kansas, 2009
Submitted in partial fulfillment of the

requirements for the degree of
MASTER OF SCIENCE IN SYSTEMS ENGINEERING
from the
NAVAL POSTGRADUATE SCHOOL

December 2017
Approved by: Alejandro Hernandez

Thesis Advisor
Oleg Yakimenko
Co-Advisor
CDR Brian Connett

Second Reader
Ronald Giachetti
Chair, Department of Systems Engineering
iii
iv
ABSTRACT
A new type of expendable radar decoy, one that leverages the advances in
technology of the last 50 years, is required to improve the survivability of non-stealth
aircraft against radar-guided threats. This thesis applies modeling and simulation based
systems engineering to explore the design space of a notional active expendable decoy to
determine the combination of key performance parameters that will maximize probability
of survival. The thesis focuses on development of a Microsoft Excel-based simulator,
which accepts inputs from the user, uses a nearly orthogonal Latin hypercube experimental
design, and performs multiple simulator runs with stochastic model variables. The final
output includes statistical analysis of results to provide the user with the best combination
of design variables to improve survivability against specified threat parameters. Results
show that either low power with high gain or high power with low gain was the best
combination for maximizing probability of survival. In addition, low altitude deployments
are to be avoided unless the missile is very near the target aircraft. The results provided by
the tool are notional, based on realistic approximations of radar, aircraft, and decoy
systems. The results will inform the acquisition strategy of NAVAIR.
v
vi
TABLE OF CONTENTS
I. INTRODUCTION..................................................................................................1
A. BACKGROUND ........................................................................................1
B. THE PROBLEM ........................................................................................5
C. THESIS ORGANIZATION ......................................................................7
II. LITERATURE REVIEW .....................................................................................9

A. CURRENT EAD RESEARCH .................................................................9
B. COMPUTER SIMULATIONS FOR ANALYSIS ................................11
C. BENEFITS OF MODELING AND SIMULATION IN THE
SYSTEM ENGINEERING PROCESS ..................................................12
D. PRIMARY MODEL REFERENCES ....................................................14
III. STUDY METHODOLOGY AND MODELING EFFORT .............................17

A. PROPOSED PROBLEM SOLUTION ..................................................17
B. MODELING AND SIMULATION BASED SYSTEMS
ENGINEERING DISCUSSION .............................................................17
C. SIMULATION ENGAGEMENT SCENARIO .....................................18
D. RADAR AND WEAPONS SYSTEMS MODEL ..................................22
E. AIRCRAFT MODEL ..............................................................................25
F. EAD MODEL ...........................................................................................30
G. ENVIRONMENT MODEL.....................................................................31
H. PROBABILITY DENSITY FUNCTIONS ............................................34
I. VERIFICATION AND VALIDATION OF THE INTEGRATED
SIMULATION MODEL .........................................................................35
J. MEASURES OF EFFECTIVENESS .....................................................37
K. FACTORS ................................................................................................37
L. EXPERIMENTATION ...........................................................................38
M. DATA ANALYSIS ...................................................................................40
IV. SIMULATOR OPERATION ..............................................................................41
V. SIMULATION EXPERIMENTS AND ANALYSIS ........................................47

A. SIMULATION SETUP ...........................................................................47
B. SENSITIVITY ANALYSIS ....................................................................48
C. REGRESSION ANALYSIS ....................................................................55
VI. CONCLUSION AND RECOMMENDATIONS ...............................................59
vii
A. CONCLUSIONS ......................................................................................59
B. RECOMMENDATIONS .........................................................................60
APPENDIX A. EQUATION VARIABLE DEFINITIONS..........................................63
APPENDIX B. MODELS ................................................................................................65
LIST OF REFERENCES ................................................................................................71
INITIAL DISTRIBUTION LIST ...................................................................................75
viii
LIST OF FIGURES
Figure 1. Towed Decoy behind Eurofighter Typhoon. Source: Onnis (2012). ...........3
Figure 2. Two Types of U.S. Navy Chaff. Source: Wikipedia (2008). .......................5
Figure 3. Example of Radar Returns from an Expendable Active Decoy ..................7
Figure 4. Expendable Active Decoy Systems Fielded by Russia and the UK.
Sources: Kuzmin (2009); Chuter (2016)....................................................10
Figure 5. Prototype RT-1489/ALE GEN-X Decoy and Canister. Source:

Koppenberger (2017). ................................................................................11
Figure 6. ESAMS Model Elements. Source: Defense Systems Information

Analysis Center (2017). .............................................................................12
Figure 7. Systems Engineering Process “V” Model. Adapted from Rausch et

al. (2007). ...................................................................................................13
Figure 8. Type 345 (Castor 2 J/C) Engagement Radar. Source: Air Power
Australia (2009). ........................................................................................20
Figure 9. Engagement Scenario Flow Diagram and OV-1 .......................................21
Figure 10. Integration Improvement Factor. Source: Harney (2013a, 370). ...............24
Figure 11. Basic 3D Model of MH-60S with Hellfire. Source: POFACETS

MATLAB 2016a. .......................................................................................26
Figure 12. MH-60S Bistatic Radar Cross Section at 16 GHz, 0° Angle of

Elevation. Source: POFACETS MATLAB 2016a. ...................................27
Figure 13. Additional Carrier-to-Noise Required for Fluctuating Targets.

Source: Harney (2013a). ............................................................................28
Figure 14. Receiver Operating Characteristic Curves for Swerling II Statistics.

Source: Harney (2013a). ............................................................................29
Figure 15. Geometry Definition of Multipath Returns. Source: Harney (2013a) .......32
Figure 16. Visual Simulation of Engagement .............................................................36
Figure 17. Nearly Orthogonal Latin Hypercube Experimental Design.......................39
Figure 18. Inputs to the EAD Model ...........................................................................42
ix
Figure 19. Multipath and Sea Surface Clutter Models ................................................43
Figure 20. Design of Experiment Variable Levels ......................................................43
Figure 21. Data Output from Multiple Simulation Runs.............................................44
Figure 22. Example of Statistical Output on the “DOE” tab.......................................44
Figure 23. Inputs Requested by the Simulator for Multiple Runs...............................45
Figure 24. Effect of Decoy Power on the Probability of Success ...............................50
Figure 25. Effect of Decoy Gain on Probability of Success .......................................53
Figure 26. Regression Analysis Results ......................................................................57
Figure 27. Radar Model...............................................................................................65
Figure 28. Multipath Model ........................................................................................66
Figure 29. Swerling Case I/II Receiver Operating Characteristics .............................67
Figure 30. Sea Surface Clutter ....................................................................................68
Figure 31. EAD Model ................................................................................................69
x
LIST OF TABLES
Table 1. Castor 2 J/C Specifications. Source: Streetly (2006); Air Power Australia
(2014). ........................................................................................................19
Table 2. Efficiency of Non-matched Filters. Source: Brooker (2007, 290). .....................25
Table 3. Normalized Mean Backscatter Coefficient for 0.3° Grazing Angle. Source:
Harney (2013b, 115) ..................................................................................33
Table 4. Experimentation Factors ......................................................................................38
Table 5. Statistical Data of Design of Experiments 1 ........................................................49
Table 6. Noteworthy Results from Design of Experiments 2 ............................................51
Table 10. Design Points with Greater than 90% Probability of Success ...........................58
xi
xii
LIST OF ACRONYMS AND ABBREVIATIONS
ASPD Advanced Self Protection Decoys

C2 command and control
CEADS Cartwright Expendable Active Decoy Simulator
CNR carrier-to-noise ratio
DOE design of experiments
DRFM digital radio frequency memory
DOD Department of Defense
EAD expendable active decoy
ECM electronic countermeasure
EM electromagnetic
ESAMS Enhanced Surface-to-Air Missile Simulation
EW early warning, electronic warfare
GEN-X Generic Expendable
J/S jam-to-signal ratio
KRET Radioelectronic Technologies Concern
LO low observable
M&S modeling and simulation
MBSE model based systems engineering
MOE measure of effectiveness
MSBSE modeling and simulation based systems engineering
MTS multi-spectral targeting system
NAVAIR Naval Air Systems Command
NOLH nearly orthogonal Latin hypercube
RCS radar cross section
ROC receiver operating characteristics
SE systems engineering
SOI system of interest
xiii
xiv
EXECUTIVE SUMMARY
Currently fielded expendable countermeasure systems cannot adequately protect

friendly aircraft from modern radar-guided threats. The primary method of radar
countermeasures employed on most aircraft is chaff technology, but other methods include
towed decoys and/or electronic jamming. Unfortunately, each of these methods has
shortcomings such as high cost, feasibility issues, and capability limitations. Chaff is
inexpensive and may be used on nearly all aircraft but has a limited capability. Towed
decoys are very capable systems, but they have limited numbers of decoys, relatively high
cost, and their employment is not feasible on all aircraft. Electronic jamming is similarly
capable as towed decoys, but it is not feasible to have a dedicated jamming platform for
every mission or to fit the electronics on every airframe. Developing a new type of
countermeasure that combines the best characteristics of each method with fewer
limitations is possible with today’s technology.
Developing an Expendable Active Decoy (EAD) can leverage evolving

technologies from the last five decades. An EAD is a system that when launched, samples
the electromagnetic (EM) spectrum broadcast by a radar, records the signal, applies phase
and frequency shifts to the recording, and then rebroadcasts the signal in order to mislead
the enemy radar. The radar-guided threat will then steer toward a false target that appears
in a different location.
Other countries have already developed EAD systems, such as the BriteCloud
system created by Selex ES (now Leonardo) of the UK, Advanced Self Protection Decoys
(ASPD) created by Reut Systems & Technologies of Israel, and the President-S system
developed by Radioelectronic Technologies Concern (KRET) of Russia. The EAD systems
fit into the physical space occupied by traditional chaff cartridges and are expended much
the same way. However, with the EAD’s onboard digital radio frequency memory (DRFM)
technology, it can sample the electromagnetic spectrum produced by a threat radar, record
the signal, and replay it with a stronger return than the target aircraft. This process enables
an EAD to capture a missile signal in its terminal phase and “walk” it off the target.
xv
One of the benefits of attempting to design an EAD, versus a new type of chaff, is
that one removes the difficulty of modeling a complex chaff cloud and instead models the
dynamics of a single object. This simplifies the modeling process so that a low-fidelity
discrete event simulation may be able to predict the performance parameters of the system
of interest (SOI) accurately. The systems engineering process utilizes the low-fidelity
simulation developed in this thesis during the early phases of the systems acquisition
process: namely, the concept exploration, concept of operations, and system requirements
phases. This low-fidelity simulation allows designers and engineers to explore the design
space of the SOI and to perform an analysis of alternatives to provide high-level system
requirements. Therefore, using this simulation now to ensure rapid acquisition of a
desperately needed capability is important.
Initial modeling was conducted utilizing Microsoft Excel to act as a validation tool
for the simulation that would eventually be developed. The author reviewed a thesis written
by Jeremy Braud in 2014 at NPS that examined the possibility of using Microsoft Office
products, namely Excel, to create an Electronic Attack Route Optimization tool to validate
the potential of using non-compiled software to develop military specific applications
(Braud 2014). The findings of his thesis showed that Microsoft Office software provided a
means for constructing useful modeling and simulation tools. This convinced the author
that the capabilities of pre-approved software installed on Department of Defense (DOD)
computers would be the preferred method for development of the simulation. Additionally,
the wide proliferation of Microsoft Excel ensures that users of the simulation will require
minimal training in order to maximize the simulator’s utility.
To employ a modeling and simulation based systems engineering approach requires

development of an acceptable model. The computer simulation scenario uses an open-
ocean engagement of a friendly helicopter by an enemy surface vessel. An open-ocean
engagement removes the possibility of terrain masking from the friendly helicopter’s
repertoire of defensive tactics. This thesis supports an EAD system in development by
NAVAIR, which will be used on naval aircraft in the maritime environment. Additionally,
open-ocean reduces the possibility of multi-path effects playing a role and simplifies their
calculation when they do occur.
xvi
The scenario employs the MH-60S helicopter because it is the most predominant
U.S. Navy rotary-wing aircraft and fulfills the anti-surface warfare mission set.
Additionally, the inherent speed and maneuverability limitations of the airframe preclude
the possibility of out maneuvering or out running the threat missile. Lastly, this aircraft
represents the worst-case scenario in terms of radar cross section (RCS) of naval aircraft
because of complex interactions of the airframe and rotor systems. If an EAD is capable of
decoying a radar-guided threat when employed from a helicopter, then logically the
effectiveness should improve when employed from a fighter or transport aircraft that has
greater maneuverability and speed.
The frequency range of interest for this thesis is from 2 – 18 Ghz. To validate the
simulation later in the study, the threat radar chosen was the Castor 2 J/C radar shown in
Figure 8. It operates within the frequency range of interest and is a commonly used fire
control radar on foreign naval vessels. The missile systems paired with the Castor 2 J/C are
semi-active guided systems. According to a one-on-one interaction with Dr. Robert Harney
(2017), bistatic radars best represent semi-active guided systems with the radar being the
transmitter and the missile being the receiver. According to Streetly’s Jane’s Radar and
Electronic Warfare Systems 2005–2006, the Castor 2 J/C radar is a broadband, J-band,
monopulse-Doppler tracking radar. It has a peak power of 30 kilowatts, a maximum range
of 30 km, and a tracking accuracy of 5 m. It also incorporates an EO/IR suite for passive
tracking and target acquisition (Streetly 2006).
The Cartwright Expendable Active Decoy Simulator (CEADS) is the result of the
integration of all of the models into a single simulator. Each of the aforementioned models
is integrated into the Excel-based simulation on individual tabs. A tab exists to represent
the EAD, radar, missile, multipath, and sea surface clutter models. The models are cross-
linked so that if a shared value changes on one tab, it changes on all corresponding models
that use the same value.
CEADS accepts input from the user on the radar, missile, aircraft, EAD, and
environment variables. The performance parameters of the missile, aircraft, EAD,
environment, and Castor 2 J/C radar are input into the associated models. The radar,
missile, and EAD models calculate the CNR of the aircraft as seen by the radar or missile
xvii
and compare them to the required CNR for a specified detection probability. The EAD
model goes further by also calculating the CNR generated by the EAD given the EAD
performance parameters input by the user. The result is either non-detection of the aircraft
due to the CNR being below the detection threshold, aircraft detection because the CNR is
above the threshold or aircraft non-detection because the EAD CNR causes a jam-to-signal
ratio of greater than one.
The simulation allows a systems engineer to examine the key measure of

effectiveness (MOE) for the EAD system. These key MOEs drive the value of a new system
to the stakeholder and drive the requirements that the systems engineer will set for the
system. The key MOE for this simulator focuses on whether or not the EAD can
successfully drive the missile to a geometry that it is no longer able to engage the aircraft.
This happens when the J/S produced by the EAD is significant enough that the missile
remains focused on the EAD as it falls away from the aircraft until the missile’s range and
altitude no longer permit it to receive the radar returns off the target. At this point, the MOE
declares the simulator run a success and moves on to the next run. If the EAD fails to decoy
the incoming missile, then the missile’s focus will remain on the target aircraft and will
eventually strike the target resulting in a failure.
The factors that the author believes will have the greatest outcome on the success
of the EAD are those factors that are inherent to the system. These factors are the variables
contained in Equations 8 and 9 for jam-to-signal ratio. Those factors are EAD transmitter
power, EAD loss factor, EAD gain, EAD bandwidth, and EAD deployment range. Other
factors that the author believes to be significant, but are not necessarily captured in any
equations, are EAD vertical and horizontal accelerations, and EAD vertical and horizontal
velocities. These factors will have an effect on how long the EAD remains airborne and
how quickly it can drive the missile to an unfavorable geometry. These ten factors shown
combine to create the design of experiments (DOE) for CEADS.
Sequential experimentation enables the researcher to focus on the most relevant

areas of the design space. As such, a screening experiment first identifies the most relevant
factors that affect the MOE, as well as significant interactions among the factors. The next
set of experiments incorporates a nearly orthogonal Latin hypercube (NOLH) design. A
xviii
NOLH allows for efficient exploration of the system design space and provides a good
amount of information about the experimental region. NOLH also allows for simplified
analysis of results though regression analysis and similar methods. This study examines 10
factors that the researcher believes or is uncertain about the influence on the MOE. The
design of experiments used for this thesis is an 11 x 33 NOLH from Cioppa (2002). Each
design point is unique in its combination of values, and constitutes a singular instantiation
of a potential EAD design.
CEADS uses the NOLH DOE to test many possible combinations of the EAD
design space against an individual radar and missile system. This allows for analysis of the
data in order to determine the best combination of performance parameters that maximizes
the EAD effectiveness against a particular threat. Testing the CEADS against the Castor
2J/C radar will prove the simulator is a useful tool. Beyond the scope of this thesis, CEADS
allows for data collection for a multitude of threat systems, analysis of data, and aides in
developing high-level system requirements.
The data that resulted from the sensitivity analysis demonstrates that CEADS
produces statistically relevant results. The author collected the results of the sensitivity
analysis and sorted them to show which combinations of variables produced the highest
probability of success. The data shows that either a low power setting with a high gain, or
a high power with a low gain, is necessary to achieve high probability of success. None of
10–100-watts design points succeeded in achieving a probability of success of 90%. Design
Point 3 had the most instances of 90% or greater probability of success. These results will
be crucial to the development of system level requirements for a future EAD.
References
Braud, Jeremy J. 2014. “Electronic Attack Platform Placement Optimization.” Master’s

thesis, Naval Postgraduate School. https://calhoun.nps.edu/handle/10945/43881
Cioppa, Thomas M. 2002. “Efficient Nearly Orthogonal and Space-Filling Experimental

Designs for High-Dimensional Complex Models.” PhD Diss. Naval Postgraduate
School. https://calhoun.nps.edu/handle/10945/9808
xix
Harney, Robert C. 2017. Discussion on Radar Model Characteristics, edited by Devon
A. Cartwright.
Streetly, Martin. 2005. Jane's Radar and Electronic Warfare Systems 2005-2006. Jane's
Information Group Limited. pp. 160-161
xx
ACKNOWLEDGMENTS
I would like to thank my advisors, Professor Alejandro Hernandez and Professor

Oleg Yakimenko. Your mentorship and guidance helped make sure that my thesis was of
top quality. I also thank my second reader, Commander Brian Connett. Our one-on-one
sessions helped me put things in perspective when I was struggling with certain concepts.
You always had good insight and were a solid sounding board.
I would also like to thank my beautiful wife, Audrey, for supporting me through all
of the hard times. Her love and devotion made it easier to shoulder the burdens of the
difficult days and appreciate the easy ones. I want to thank my amazing daughter,
Everleigh. Watching her learn and grow in the first seven months of her life encourages
me to never stop learning.
xxi
xxii
I. INTRODUCTION
The introduction chapter provides an overview of methods currently employed to

protect aircraft from radar-guided threats. The author compares these methods, and
shortcomings are noted to identify a problem statement: Currently fielded expendable
countermeasure systems cannot adequately protect friendly aircraft from modern radar-
guided threats. Lastly, the author provides a summary of the thesis organization.
A. BACKGROUND
In the age of stealth aircraft eluding detection, destroying their targets, and
returning to base, some may consider the idea of expendable radar countermeasures to be
antiquated; however, the author does not believe so. Only a few aircraft have ever
implemented a full suite of what is known as low observable (LO) technology. Fighter
aircraft such as the F-22 Raptor and F-35 Lightning II, and bombers such as the B-2 Spirit
and F-117 Nighthawk have radar cross sections the size of a bird or smaller (Yue 2001).
While LO technology protects them from a large number of threats, they are not impervious
to detection. In 1999, an F-117 was shot down over Kosovo by an SA-3 infrared guided
missile. In 2001, a British defense firm known as Roke Manor developed a prototype
method of detecting stealth aircraft utilizing currently installed cellphone antenna
infrastructure to act as a type of multi-static radar (Yue 2001). As recently as November
2016, Popular Mechanics magazine reported that China claims that its newest radars
demonstrated the capability to detect America’s stealthiest aircraft, the F-22 Raptor
(Mizokami 2016). With the recent demonstrations of America’s most survivable aircraft
being detectable, even they can benefit from countermeasures. Additionally, because most
other aircraft in the military’s inventory are not LO, a method of protecting them from
missile engagement is still necessary. The primary method of radar countermeasures
employed on most aircraft is chaff technology, but other methods include towed decoys
and/or electronic jamming. Unfortunately, each of these methods has shortcomings such
as high cost, feasibility issues, and capability limitations.
1
Towed decoys, such as the Raytheon AN/ALE-50 employed on non-LO fighter
aircraft, restrict the ability to maneuver and require a very specific missile-aircraft
orientation to be effective. Additionally, the aircraft can only carry a very limited number
of single-use towed decoys. Figure 1 shows a towed decoy deployed behind a fighter
aircraft.
Helicopters do not employ towed decoys because of the possibility of the line
entanglement in the main or tail rotor. Transport and other non-fighter aircraft do not
employ towed decoys because they do not operate in threat environments that require radar-
guided threat defense. Unfortunately, the missile engagement zones of many radar-guided
threats are constantly expanding, which will likely put these aircraft in range. In a recent
interview, Neil Ashdown, deputy editor of IHS Jane’s Intelligence Review stated that the
Chinese HQ-9 system, a derivative of the Russian S-300, has a 230 km range (Ashdown
2016). This could put American command and control (C2), early warning (EW), and
refueling aircraft at risk.
While aircraft can use towed decoys effectively to protect themselves, currently no
system exists to retrieve a decoy if it has been deployed but not used. At present, if a decoy
is deployed, the aircraft must cut the fiber optic cable to release it prior to landing. Defense
Industry Daily states that the price of a single decoy is $22,000 and has a 10-year shelf life.
The loss of a towed expendable each time it is deployed and the limited shelf life are
shortcomings that can be overcome.
2
Figure 1. Towed Decoy behind Eurofighter Typhoon. Source: Onnis (2012).
Electronic jamming is an active measure that can be effective at defending against

radar-guided threats; however, due to the sophistication of integrated electronic systems
involved, jamming often requires a separate dedicated platform. Such jamming is infeasible
for helicopters and other aircraft, because such aircraft cannot fit the electronic systems in
their already space-constrained airframes.
Since jamming is an “active” measure, it is possible for a radar operator to

determine that the radar is experiencing jamming indicating the presence of a threat. The
operator may tune the radar system to the situation to overcome the jamming or to alert
other radar systems to the presence of the threat. Tactically, remaining undetected to the
radar as long as possible and responding only if detection occurs is the best strategy.
Aircraft with limited electronic warfare (EW) or electronic countermeasure (ECM)
capabilities require an on demand expendable system to remain undetected while also
retaining the capability to defeat a threat if encountered.
3
Chaff is a passive expendable countermeasure that provides an on demand defense
against radar-guided threats. Chaff consists of millions of fibers that are about the thickness
of a human hair and only a few centimeters or less in length. These fibers, known as dipoles,
are typically aluminum coated fiberglass filaments cut to a specific length in order to have
the greatest effect against a predetermined radar frequency. This restricts chaff
effectiveness to a small number of threats that emit the wavelength corresponding to the
length of the chaff dipoles.
Chaff works by deploying a canister full of chaff fibers into the moving airstream
around an aircraft of ship. Much like blowing on a dandelion, the air moving past the
vehicle causes the millions of dipoles to spread out into a large cloud within a few seconds.
This cloud absorbs and re-transmits the incident radar energy to cause a radar return. If the
radar dipoles are of the appropriate length for the radar system, the return that an operator
sees on their display from the chaff cloud may be greater in magnitude than the return of
the actual vehicle.
To create a chaff countermeasure that is useful against a broad range of radar

frequencies, multiple dipole lengths per expendable canister, or multiple expendables with
different length dipoles is required. The limited space on the aircraft and standard
expendable canister sizes limits the total number of chaff dipoles carried in a single
expendable canister. The Navy’s primary chaff expendable RR-129A/AL, shown in the
bottom of Figure 2, follows the first method and has three different dipole lengths
corresponding to specific frequencies. Radars that do not transmit in those specific
frequencies will likely be unaffected by the chaff. Additionally, chaff is less effective
against modern threat radars that rely on Doppler and moving target indicator, as the chaff
cloud will remain essentially stationary when compared to the target. The other chaff
shown at the top of Figure 2 is RR-144A/AL, which is a training chaff that does not affect
civilian air traffic control radars. The see through canisters are to show the internal
arrangement of the chaff. The cluster of dipoles on the right side of the image shows the
fibers separated from the canister.
4
U.S. Navy RR-144A/AL training chaff is shown on top and RR-129A/AL Chaff shown on
bottom. Notice the different length in the RR-129A/AL compared to the uniform lengths
in the RR-144A/AL. The fibers on the right show the dipoles outside of the canister.
Figure 2. Two Types of U.S. Navy Chaff. Source: Wikipedia (2008).
B. THE PROBLEM
Currently fielded expendable countermeasure systems cannot adequately protect

friendly aircraft from modern radar-guided threats. Each of the aforementioned methods of
radar countermeasures has drawbacks in capability, cost, or feasibility. Chaff is
inexpensive and may be used on nearly all aircraft but has a limited capability. Towed
decoys are very capable systems but they have limited numbers of decoys, relatively high
cost, and their employment is not feasible on all aircraft. Electronic jamming is similarly
capable as towed decoys but it is not feasible to have a dedicated jamming platform for
every mission or to fit the electronics on every airframe. Developing a new type of
5
countermeasure that combines the best characteristics of each method with fewer
limitations is possible with today’s technology.
Chaff has been the primary expendable countermeasure used since World War II
as a method of fooling enemy radar. However, with the advent of LO technologies, the
United States halted its countermeasure research and redirected funding into this
technology, which promised to make aircraft nearly invisible to radar. According to
NAVAIR’s Advanced Tactical Aircraft Protection Systems (PMA-272), there has been
minimal research in the area of countermeasures since the 1970s (Koppenberger 2017).
The chaff designs developed from this limited research are still the primary expendable
radar decoys employed by all the U.S. military services nearly 50 years later. This chaff
was designed to deal with early Cold War Soviet threats and while countermeasure research
diminished, radar technology continued to improve.
A cursory search identifies few countermeasure research papers scattered

throughout the decades since the 1970s. Of those papers, many cite the problems of
accurately modeling a chaff cloud as the largest hurdle to overcome. (Knott et al. 1981)
Research exists characterizing individual chaff cloud dipoles; however, it is difficult to
model accurately the random nature of millions of individual dipole elements in differing
environmental conditions without better fluid flow dynamic models (Scholfield et al.
2011).
Developing an Expendable Active Decoy (EAD) can leverage evolving

technologies from the last five decades. An EAD is a system that when launched, samples
the electromagnetic (EM) spectrum broadcast by a radar, records the signal, applies phase
and frequency shifts to the recording, and then rebroadcasts the signal in order to mislead
the enemy radar. The radar-guided threat will then steer toward a false target that appears
in a different location. Figure 3 provides an example of what a radar operator might see in
a range-Doppler map.
One of the benefits of attempting to design an EAD, vice a new type of chaff, is
that one removes the difficulty of modeling a complex chaff cloud and instead models the
dynamics of a single object. This simplifies the modeling process so that a low-fidelity
6
discrete event simulation may be able to predict the performance parameters of the system
of interest (SOI) accurately.
Figure 3. Example of Radar Returns from an Expendable Active Decoy
C. THESIS ORGANIZATION
The remainder of this thesis is organized as follows: Chapter II introduces the

experiences and steps in the research process that led to the development of the thesis.
Chapter III covers the development of the models that represent the various systems and
environment and the integration of those models into the simulator. Chapter IV describes
the Excel-based simulation developed during the project and provides an overview of its
use case, and Chapter V presents the data obtained from running an example of the
simulation and analysis/interpretation of those results. Lastly, Chapter VI presents the
conclusion and recommendations.
7
8
II. LITERATURE REVIEW
The author’s tactical training to employ the MH-60S helicopter piqued an interest
in radar-guided threats and countermeasures. Methods employed to defend against radar-
guided threats have limited effectiveness because of the pilots’ lack of understanding of
the threats as well as the capabilities of the countermeasures. This judgement was validated
when, while deployed in the Pacific theater, a non-allied ship engaged the author’s aircraft
with a targeting radar. The author’s experience showed that the onboard countermeasures
were likely ineffective against the targeting radar. The non-allied ship turned off the radar
but, had the engagement progressed, the outcome would have been unfavorable for the
aircraft. After arriving at Naval Postgraduate School, these judgements led the author to
reach out to Naval Air Systems Command (NAVAIR) to determine whether it saw any
value in continued countermeasure research. While this thesis does not aim to improve the
understanding of the threats by the operator, the thesis addresses the capabilities of the
countermeasures.
Meeting with representatives from NAVAIR made clear to the author that minimal
research had been completed in the realm of radar countermeasures but the U.S. Navy
determined that this was an issue that required attention. The author attended a survivability
symposium in March of 2017 that brought together groups from various industries,
educational institutions, and each military branch. One of the more promising methods
presented to fill this capability gap was the development of an EAD system.
A. CURRENT EAD RESEARCH
Other countries have already developed EAD systems such as the BriteCloud
system created by Selex ES (now Leonardo) of the UK, Advanced Self Protection Decoys
(ASPD) created by Reut Systems & Technologies of Israel, and the President-S system
developed by Radioelectronic Technologies Concern (KRET) of Russia. Figure 4 shows
examples of these systems. The EAD systems fit into the physical space occupied by
traditional chaff cartridges and are expended much the same way. However, with the
EAD’s onboard DRFM technology, it can sample the electromagnetic spectrum produced
9
by a threat radar, record the signal, and replay it with a stronger return than the target
aircraft. This process enables an EAD to capture a missile signal in its terminal phase and
“walk” it off the target.
The President-S Disposable Jamming Transmitter is shown on the left. The BriteCloud
Expendable Active Decoy is shown on the right. The BriteCloud EAD is enlarged to show
detail.
Figure 4. Expendable Active Decoy Systems Fielded by Russia and the UK.
Sources: Kuzmin (2009); Chuter (2016).
The U.S. Navy, in cooperation with Raytheon and Texas Instruments, developed
the RT-1489/ALE Generic Expendable (GEN-X) in the early 1990s; however, the
production line no longer exists. The GEN-X is a basic EAD with a limited frequency
range. At the symposium, PMA-272 stated that one of the primary concerns with
developing a new EAD system is to avoid the same mistakes encountered in the GEN-X
program, specifically, the battery and storage life problems. Besides improving the
capability of the new EAD system to work across a broader frequency spectrum, it will be
important to address the storage life concern. The battery storage-life concern is beyond
the scope of this thesis. Figure 5 is an image of the GEN-X decoy.
10
Figure 5. Prototype RT-1489/ALE GEN-X Decoy and Canister. Source:
Koppenberger (2017).
B. COMPUTER SIMULATIONS FOR ANALYSIS
Computer programs presented at the symposium simulated engagements of aircraft

by radar-guided threats. One of the limitations discussed was outdated programming
languages utilized to create the simulations, making it difficult to update. The simulations
were also unreliable and required a large amount of computing power along with
specialized training to execute. One of the more useful software programs presented was
the Enhanced Surface-to-Air Missile Simulation (ESAMS). ESAMS takes multiple
models, developed over the last several decades, and combines them into a single
simulation to leverage the capabilities of each, creating an in depth simulation tool. Figure
6 shows the input models for ESAMS.
While this simulation is an extremely powerful tool, it represents a very high-fidelity

model that is better used in the later stages of the systems engineering (SE) process that can
be used to verify earlier SE efforts. This experience convinced the author that a low-fidelity
simulation is necessary to replicate the ESAMS results earlier in the SE process. This
replication effort will require less computing power, minimal training to operate, and can
retain appropriate classification levels for the environment in which it is exercised.
11
Figure 6. ESAMS Model Elements. Source: Defense Systems Information
Analysis Center (2017).
C. BENEFITS OF MODELING AND SIMULATION IN THE SYSTEM

ENGINEERING PROCESS
According to the INCOSE Systems Engineering Handbook, the benefits of

modeling and simulation (M&S) in the SE process include confirming of system
requirements and behavior prior to development and proving insight and clarification to
designers and engineers to minimize errors while maximizing productivity. This helps to
reduce cost and schedule overruns by obtaining information about the system early in the
design process and prior to major commitment of resources. The data that results from
M&S provides data to the analyst, supports decision-making, and informs stakeholders of
the outcome of their preferences as well as system limitations and capabilities.
Additionally, the models provide a common framework for engineers that aid in developing
a shared understanding of the system requirements.
12
The systems engineering process utilizes the low-fidelity simulation developed in
this thesis during the early phases of the systems acquisition process: namely, the concept
exploration, concept of operations, and system requirements phases as shown in Figure 7.
This low-fidelity simulation allows designers and engineers to explore the design space of
the SOI and to perform an analysis of alternatives to provide high-level system
requirements. According to Benjamin Blanchard and Wolter Fabrycky, systems
engineering efforts undertaken early in the system acquisition process can result in cost
savings, reduced acquisition times, and a reduction in risks. All of these combine to ensure
stakeholder satisfaction with the final product (Blanchard and Fabrycky 2006, 48).
Therefore, using this simulation now to ensure rapid acquisition of a desperately needed
capability is important.
Figure 7. Systems Engineering Process “V” Model.

Adapted from Rausch et al. (2007).
As recently as October of 2017 at the National Defense Industrial Association’s

Systems Engineering Conference, the commander of NAVAIR Vice Admiral Paul
Grosklags communicated that NAVAIR requires a new model-based systems engineering
13
environment that solves integration and training problems, and reduces the time to design,
test, and evaluate new systems. The current process involves giving 500-page specification
documents to industry, asking them to interpret the requirements, and build a system from
scratch. This practice hinders the ability to make little changes along the way, while testing
changes in a virtual environment allows for rapid solution generation. With accurate
information from stakeholders, it is possible to build a computer model, insert notional
aircraft or weapons systems, and provide the model to industry to facilitate system design.
The intent is also to use the models as verification tools later in the SE process to assess
program maturation and to validate earlier assumptions. He added that the benefits of
model-based systems engineering extend throughout test and evaluation, production, and
sustainment phases and can demonstrate system robustness over the system life cycle.
NAVAIR plans to implement model-based systems engineering into any future program to
provide the opportunity to learn as quickly as possible and to achieve the benefits
previously listed. Working in direct contact with NAVAIR, this thesis and the resulting
models and simulation directly support the vice admiral’s call to action (Eckstein 2017).
D. PRIMARY MODEL REFERENCES
Modeling the individual elements of the simulation, requires a deeper review of

physics and equation-based approximations inherent in the engineering of the radar,
missile, aircraft, and EAD systems. For this, the author turned to the writings of Dr. Robert
Harney of the Naval Postgraduate School and Dr. Graham Brooker of the University of
Sydney Australian Centre for Field Robotics. The aforementioned authors’ publications
provide the necessary mathematical models to represent the tracking radar, multipath,
clutter, aircraft, and EAD system. Physics-based models were adapted from the author’s
operational knowledge to model the flight paths of the missile, aircraft, and EAD.
Initial modeling was conducted utilizing Microsoft Excel to act as a validation tool
for the simulation that would eventually be developed. The author reviewed a thesis written
by Jeremy Braud in 2014 at NPS that examined the possibility of using Microsoft Office
products, namely Excel, to create an Electronic Attack Route Optimization tool to validate
the potential of using non-compiled software to develop military specific applications
14
(Braud 2014). The findings of his thesis showed that Microsoft Office software provided a
means for constructing useful modeling and simulation tools. This convinced the author
that the capabilities of pre-approved software installed on Department of Defense (DOD)
computers would be the preferred method for development of the simulation. Additionally,
the wide proliferation of Microsoft Excel ensures that users of the simulation will require
minimal training in order to maximize the simulator’s utility.
15
16
III. STUDY METHODOLOGY AND MODELING EFFORT
This chapter introduces a proposed problem solution to answer the problems

statement in Chapter I; namely, fielded expendable countermeasure systems cannot
adequately protect friendly aircraft from modern radar-guided threats. The solution
includes developing a low-fidelity simulation to allow designers and engineers the ability
to examine the design space of a new system that can fill this capability gap. The chapter
introduces the engagement scenario on which the simulation is based as well as identifies
the engagement variables that require modeling. The chapter discusses the individual
models of the simulation in detail including the mathematical equations used to represent
the scenario elements. Individual models represent the radar system, missile system,
aircraft, EAD, and environment. All equation variables, their definitions, and the units are
contained in Appendix A.
A. PROPOSED PROBLEM SOLUTION
In order to facilitate the design of a new EAD system by the U.S. military, the first
step is to develop a low-fidelity simulation that allows for low-cost exploration of the
design space, fulfilling the request of Vice Admiral Grosklags. The basis for this thesis is
the construction of a simulation that will receive input from a user on the characteristics of
a radar-guided threat, aircraft, expendable, and environment. The models of the systems
will be equation-based and physics-based approximations. The simulation executes
hundreds of times to collect samples of the decoy’s performance based on the variation of
stochastic variables. Statistical analysis of the data then determines the decoy’s
effectiveness against the selected radar-guided threat.
B. MODELING AND SIMULATION BASED SYSTEMS ENGINEERING

DISCUSSION
This study uses a modeling and simulation-based systems engineering (MSBSE)

approach, employs scenario methodologies, and experimentation using a low-fidelity
computer simulation to identify the required key performance parameters for a new radar
decoy system. According to the INCOSE SE Vision 2020 point paper, “model-based
17
systems engineering (MBSE) is the formalized application of modeling to support system
requirements, design, analysis, verification and validation activities beginning in the
conceptual design phase and continuing throughout development and later life cycle
phases” (INCOSE-TP-2004-004-02 2007).
The difference between a model and a simulation is “a model is a physical,

mathematical, or logical abstract representation of a system entity, while a simulation is
the implementation of a model over time that brings the corresponding model to life”
(Gianni et al. 2015). MSBSE incorporates models in the context of a simulation to evaluate
the properties of a complex system, in a specific environment, and typically allows for
execution on a time scale, to provide a more comprehensive analysis of the design space
than can be achieved with MBSE alone. This thesis employs models to represent the
various individual elements of helicopter and missile engagement, which include the
operational scenario, radar and weapon systems, aircraft, decoy systems, environment, and
decision processes. Using these models within a simulation construct allows for the
analysis and design of the SOI.
C. SIMULATION ENGAGEMENT SCENARIO
To employ an MSBSE approach requires development of an acceptable model. The

computer simulation scenario uses an open-ocean engagement of a friendly helicopter by
an enemy surface vessel. An open-ocean engagement removes the possibility of terrain
masking from the friendly helicopter’s repertoire of defensive tactics. This thesis supports
an EAD system in development by NAVAIR, which will be used on naval aircraft in the
maritime environment. Additionally, open-ocean reduces the possibility of multi-path
effects playing a role and simplifies their calculation when they do occur.
The scenario employs the MH-60S helicopter because it is the most predominant
U.S. Navy rotary-wing aircraft and fulfills the anti-surface warfare mission set.
Additionally, the inherent speed and maneuverability limitations of the airframe preclude
the possibility of out maneuvering or out running the threat missile. Lastly, this aircraft
represents the worst-case scenario in terms of radar cross section (RCS) of naval aircraft
because of complex interactions of the airframe and rotor systems. If an EAD is capable of
18
decoying a radar-guided threat when employed from a helicopter, then logically the
effectiveness should improve when employed from a fighter or transport aircraft that has
greater maneuverability and speed.
The frequency range of interest for this thesis is from 2 – 18 Ghz. To validate the
simulation later in the study, the threat radar chosen was the Castor 2 J/C radar shown in
Figure 8. It operates within the frequency range of interest and is a commonly used fire
control radar on foreign naval vessels. The missile systems paired with the Castor 2 J/C are
semi-active guided systems. According to a one-on-one interaction with Dr. Robert Harney
(2017), bistatic radars best represent semi-active guided systems with the radar being the
transmitter and the missile being the receiver. According to Streetly’s Jane’s Radar and
Electronic Warfare Systems 2005–2006, the Castor 2 J/C radar is a broadband, J-band,
monopulse-Doppler tracking radar. It has a peak power of 30 kilowatts, a maximum range
of 30 km, and a tracking accuracy of 5 m. It also incorporates an EO/IR suite for passive
tracking and target acquisition (Streetly 2006). Table 3.1 covers the Castor 2 J/C
specifications that were derived from Jane’s Radar and Electronic Warfare Systems and
Air Power Australia, an independent defense think tank.
Table 1. Castor 2 J/C Specifications. Source: Streetly (2006); Air Power

Australia (2014).
Peak Power 30,000+ watts

Frequency Operating Band J-band; 15.7 – 17.7 GHz
Range 0.7 – 30 km
Tracking Accuracy 5m
Pulse Repetition Frequency (PRF) 3600 – 7200 pps
Pulse Duration 7.4 – 7.6 μs
Transmitter Gain 43 dB
Beam width 0.67° elevation
19
Figure 8. Type 345 (Castor 2 J/C) Engagement Radar.
Source: Air Power Australia (2009).
Figure 9 represents the engagement scenario flow diagram. A friendly aircraft will
approach a naval surface vessel. The naval surface vessel will be radiating its detection and
tracking radars in an effort to detect the aircraft. Once the aircraft is in range of the targeting
radar, the radar-guided threat engages the aircraft. The aircraft, having detected the
engagement, will deploy an EAD in an attempt to defend against the threat. Either the
decoy will succeed and the missile will fail to destroy the aircraft, or the missile will
successfully impact the aircraft. Red boxes represent the inputs and outputs related to the
radar and missile models. Blue boxes are the inputs and outputs of the aircraft model. Green
boxes represent the inputs and outputs related to the EAD model. The inset image shows
the Department of Defense Acquisition Framework Operational View (OV-1). This is a
high-level representation of the engagement that the simulation replicates.
20
Figure 9. Engagement Scenario Flow Diagram and OV-1
21
D. RADAR AND WEAPONS SYSTEMS MODEL
The radar model uses the radar range equation to represent the radar and missile
systems. This equation provides the carrier-to-noise ratio (CNR) of the radar. The CNR is
the ratio of the radar-generated signal that returns from a target to the noise that the radar
receives from background phenomena. The greater the CNR, the greater the probability of
detection by the radar. This factor is important because it defines the ability of the radar to
detect a target. The radar range equation, derived from Harney (2013a), is given as
PT LT LRGT AR e 2 R

CNR  (1)
16 2 (kTBF ) R 4
Because the semi-active guided missile represents a bistatic radar system, the
author modified Equation 1 to differentiate between the transmitter and receiver ranges;
that is to say, as the missile moves closer to the aircraft, the receiver range will decrease
while the transmitter range remains approximately the same. Equation 2 reflects these
changes and applies to the missile system only.
PT LT LRGT AR e  ( RT  RR )
CNR  (2)
16 2 (kTBF ) RT 2 RR 2
Research provides some of the values for the variables such as power and gain.
Other variables must use reasonable assumptions. Assumptions made for the model
include:
 transmitter and receiver loss factor of -2 dB (Brooker 2007, 299)
 clear weather resulting in no atmospheric attenuation due to the frequency

being less than 20 GHz (Harney 2013c, 96)
 electronic temperature is 290K (Brooker 2007, 283)
 Radar receiver on missile has a radius of 0.07 m. The HQ-7 missile is one
of the missiles commonly paired with the Castor 2J/C and has a diameter of
22
0.15 m according to Air Power Australia; therefore, a receiver radius of 0.07
m is reasonable. (Air Power Australia 2014)
 Noise Factor is 4 dB (Brooker 2007, 298)
 The receiver area for both the missile and the radar is circular.
In order to improve detection probability, modern radar systems employ pulse

integration to improve detection probability. They do this by collecting many return pulses
and integrating them over time instead of using individual pulses, which can vary in
intensity between subsequent pulses. Equation 3 from Brooker (2007) represents a rotating
radar’s hits per scan of the radar pulses on the target. This is useful for determining total
pulses received so that the radar can integrate the pulses to improve detection probability.
This equation is in the model for future use; however, the Castor 2 J/C does not rotate. For
this reason, the author assumes 100 hits per scan for the purposes of the simulation.
Derivation of the integration improvement factor comes from Figure 10 after determining
the hits per scan. The user will use the hits per scan calculated in Equation 3 to enter the
chart from the bottom at the appropriate value, find where the line intersects the desired
Swerling case (discussed in Aircraft Model Section III.E), and read the associated
integration improvement factor off the left side of Figure 10.
b f p
nb  (3)
s
23
Figure 10. Integration Improvement Factor. Source: Harney (2013a, 370).
A radar operates with a central frequency of interest known as the modulating

frequency; however, due to the method of generating the radar pulse by the transmitter, the
receiver takes in some of the surrounding frequency bands as well. This is known as the
radar bandwidth and it is the difference between the upper and lower cut-off frequencies
that surround the modulating frequency. The wider the bandwidth, the more noise allowed
into the system and the less sensitive the radar system becomes; however, a narrow
bandwidth might also distort the radar pulse shape resulting in degradation of receiver
performance (Payne 2010, 36–37). Equation 4, adapted from Brooker (2007), provides a
method for calculating an unknown radar bandwidth when given the pulse width. Table 2
comes from Brooker (2007) and lists the efficiency of non-matched filters. In keeping with
planning for the worst-case scenario, the author assumes that the radar will use a Gaussian
matched filter with a Gaussian pulse signal that results in no loss of signal to noise ratio on
the part of the receiver. This gives a Bτ value of 0.44.
  B (4)
24
Table 2. Efficiency of Non-matched Filters. Source: Brooker (2007, 290).
The probability of false alarm is the chance that radar noise will exceed the
detection threshold, resulting in a false detection. False detections are undesirable because
they utilize crew resources and cause distraction to the operator, which may result in
missing an actual target. With Equation 5, adapted from Harney (2013a), one can calculate
probability of false alarm (Pfa) when one knows the bandwidth and time between false
alarms. Without knowing the time between false alarms for the Castor 2J/C, the author
assumes that Tfa is two hours (7200 seconds). This assumption derives from the author’s
experience that target radars only operate for a short duration during an engagement and a
low Tfa is acceptable.
1
Pfa  (5)
T fa B
E. AIRCRAFT MODEL
The choice of aircraft brings with it the difficulties of calculating radar reflection
characteristics brought on by the complex geometry and spinning rotor systems. The
primary characteristic of these radar reflections is known as the radar cross section (RCS).
The RCS in important because the greater the RCS, the greater the CNR of a radar return
resulting in an increased probability of detection.
25
In order to determine the approximate RCS of an MH-60S, a MATLAB program
called POFACETS, developed by Hellenic Air Force Major Filippos Chatzigeorgiadis at
NPS in 2004, calculates the bistatic radar cross section. A basic 3D representation of the
MH-60S was created using published and derived dimensions. A rectangle box represents
the fuselage, a wedge the tail boom, flat plates for the wings, rotors, vertical, and horizontal
stabilizers, ellipses for the cockpit and nose, a sphere for the multi-spectral targeting system
(MTS), cylinders for landing gear, and cylinders with ogive nose cones for the Hellfire
missiles. Figure 11 is a display of the basic 3D model. All axes are in meters.
Figure 11. Basic 3D Model of MH-60S with Hellfire. Source: POFACETS

MATLAB 2016a.
26
The POFACETS program allows the determination of radar cross section from any
model angle in decibels per square meter. The author used POFACETS to calculate the
bistatic radar cross section for the model in 1 GHz increments from 2 to 18 GHz. Since the
Castor 2 J/C operates between 15.7 and 17.7 GHz, the author chose 16 GHz for the model.
Figure 12 shows the polar plot of the RCS at 16 GHz and 0° elevation from 0 to 360
degrees. The vertical axis is the RCS in decibels per square meter. Based on operational
experience and the results of the POFACETS calculation, the author assumes an average
RCS of 24 dBsm. This value falls within the bounds of realistically expected values when
compared to values provided in Harney’s text (2013b, 108).
Figure 12. MH-60S Bistatic Radar Cross Section at 16 GHz, 0° Angle of

Elevation. Source: POFACETS MATLAB 2016a.
27
Most real-world objects do not exhibit a steady state RCS. Interactions between
reflections off complex vehicle geometries and spinning parts will cause the RCS to
fluctuate in intensity. Because of the reflective geometry and rotor system of the MH-60S,
it exhibits fluctuations in its RCS. Swerling models characterize these fluctuations. Figure
13 shows the additional CNR required in order to detect a fluctuating target over a non-
fluctuating one. The worst-case scenario is a target exhibiting Swerling Cases I and II
characteristics. Any realistic probability of detection (≥0.90) requires a greater CNR for
Swerling I/II than it does for Swerling III/IV. Therefore, Swerling Case II best represents
the worst case scenario for fluctuating characteristics of the helicopter model; though
Swerling Case IV likely better represents a helicopter due to the fast fluctuations of a main
and tail rotor.
Figure 13. Additional Carrier-to-Noise Required for Fluctuating Targets.

Source: Harney (2013a).
The simulation allows the user to choose a desired probability of detection. For the
purposes of this thesis, the author chooses a probability of detection of 0.90. Combining
the desired probability of detection with the probability of false alarm calculated by
Equation 5, helps determine the CNR required by the radar to detect a Swerling Case II
28
target. This Swerling CNR is compared to the CNR produced by the target return from
Equation 2 to determine whether detection occurs. Equation 6 from Harney (2013a) is the
closed form solution for Swerling Case II only. Alternatively, one may use receiver
operation characteristic (ROC) curves, shown in Figure 14, to analytically determine the
CNR required.
PD  PF 
1/ 1 CNR  
(6)
Figure 14. Receiver Operating Characteristic Curves for Swerling II Statistics.

Source: Harney (2013a).
a. Flight Path
The kinematic equations of motion are mathematical representations that explain

the movement of objects over time. With a known starting range, velocity, and acceleration,
one can calculate the position of an object after a known duration of time. The equations
are separated into the vertical and horizontal components. For the purposes of the model,
29
the author assumes the flight path of the aircraft to be a straight, level, and un-accelerated.
Therefore, the aircraft model only uses the horizontal component. This movement model
is used to calculate the change in range of the aircraft as the simulation progresses as well
as provides a portion of the EAD’s forward velocity on deployment. The horizontal
component kinematic equation in Equation 7 represents the flight path where acceleration
is zero.
1 2
R  R0  Vxt  axt  R0  Vxt (7)
2
F. EAD MODEL
The jam-to-signal ratio (J/S) of a decoy is the ratio of the signal power produced by
the decoy to the signal power produced by the target radar return. A J/S of greater than one
is desirable for a decoy. If the J/S is greater than one, the decoy is able to produce a signal
that is greater than the return signal coming from the target resulting in reduced detection
probability. This is a critical measure for the simulation and defines one of the primary
measures of performance of the system. The simulation uses this equation to determine
whether or not the EAD is successful.
Decoys that record and repeat a radar signal are known as deception jammers. A
deception jammer best represents an EAD. Equation 8, adapted from Harney (2013c), best
represents these types of jammers. Since, per the scenario, the radar is engaging the friendly
helicopter, the author assumes that the radar is pointing directly at the aircraft and SLL is
1 (0 dB). Equation 3.8 applies to the jam to signal ratio for the radar only.
J PJ LT GJ  SLL  4 RT
4
 (8)
S PT LRGT  RJ 2
Because the radar is being treated as a bistatic system, as the missile moves closer
to the target the range from the receiver decreases while the jammer range and radar range
remain approximately the same (RT ≈ RJ). Therefore, Equation 8 becomes Equation 9 when
representing the missile.
30
J PJ LT GJ  SLL  4 RT RR PJ LT GJ  SLL  4 RR 2
2 2
  (9)
S PT LRGT  RJ 2 PT LRGT 
The aircraft only had horizontal flight path components; however, the EAD has
both horizontal and vertical components that require modeling. The kinematic equations of
motion in Equations 10 and 11 represent the flight path and velocities of the EAD in the
horizontal and vertical directions. In the model, the user can change the acceleration and
initial velocity of the EAD to simulate the type of deployment such as falling under
parachute or free fall. These equations will determine the change in range and altitude of
the EAD during the engagement scenario as well as the change in velocity. The author
assumes that acceleration is constant.
1
R  R0  Vxt  axt 2
2
(10)
1
H  H 0  Vy t  a y t 2
2
Vx  Vx  ax t
(11)
Vy  Vy  a y t
G. ENVIRONMENT MODEL
Multipath is a phenomenon that occurs when the radar signal bounces off the
ground plane before it bounces off the target. This creates multiple paths for the radar signal
to return to the radar receiver. Typically, four paths exist and Figure 15 shows the
geometry. These paths include a direct path from the radar, to the target, and back to the
receiver (A→C→A). Another path involves the signal bouncing off the ground plane,
hitting the target, and returning to the radar the way it came (A→B→C→B→A). The last
two paths involve either a direct path to the target and return by bouncing off the ground
plane (A→C→B→A) or vice versa (A→B→C→A). Multipath allows aircraft to hide in
null zones where the destructive interference caused by multiple radar returns results in
loss of radar detection capability. The multipath model is only valid when the angle
between the ground bounce point and the target (alpha + beta) is less than the beam width
31
elevation. Equation 12, from Harney (2013a), determines the validity of this criterion. If
multipath criteria are met then the effects can be calculated utilizing Equation 13, adapted
from Harney (2013a). If the ratio is less than one, then the target is in a null zone and is not
detectable. For the purposes of this thesis, multipath is not considered in the determination
of EAD success or failure; however, it is included for future development of the simulation
to add greater fidelity.
Figure 15. Geometry Definition of Multipath Returns. Source: Harney

(2013a)
 HT  H R  1  H T  H R 
    tan 1    tan     BW (12)
 R   R 
  2  
R 
4    HT H R 
 16  DET  sin 4   
PR
 (13)
P0  R   R 
 
 
Sea surface backscatter is a phenomenon that occurs when the radar beam is
sufficiently wide that some radar energy is reflected off the surface of the ocean. This
return, if of sufficient magnitude, can mask a target signal. The magnitude of this return
known as sea surface clutter RCS is dependent on sea state, beam width, and the range
resolution of the radar. In the simulation’s current state, the sea surface clutter RCS is
overwhelmed by the target RCS. Therefore, the computations are not used for the overall
measure of performance of the EAD, but they are in place for further development.
Equation 14, adapted from Harney (2013b), calculates the sea surface clutter RCS.
32
 SC  R BW R 0 (14)
The radar range resolution is the length of the imaginary box created in space by
the pulse duration. Equation 15 from Harney (2013b) calculates the range resolution.
c
R  (15)
2
Table 3 is an example of a look-up table containing empirically derived data

showing the mean backscatter coefficient of various wavelengths depending on grazing
angle, sea state, and radar wave polarization. The grazing angle of the incident radar wave
can be calculated using Equation 16. This thesis defines grazing angle (θGA) as the angle
of elevation of the center of the radar beam to the target minus half the elevation beam
width.
 H  H R   BW
GA  tan 1  T  (16)
 R  2
Table 3. Normalized Mean Backscatter Coefficient for 0.3° Grazing

Angle. Source: Harney (2013b, 115)
33
After choosing the appropriate mean backscatter coefficient, the value is substituted
into Equation 17, adapted from Harney (2013b), in order to calculate the sea clutter CNR.
PT LT LRGT AR ( R BW R 0 )e-2 R

CNRSC  (17)
162 (kTBF ) R 4
H. PROBABILITY DENSITY FUNCTIONS
One can plot the probability density functions of the noise and signal plus noise to
provide a visual comparison of the target radar return intensity to background noise. The
noise inherent in the radar is often narrow-band filtered and results in a Gaussian
probability distribution. Equation 18, from Harney (2013a), represents the Gaussian noise.
V is the voltage of the noise return and V0 is the root mean squared noise level.
2 V2
V  
pN V     e 2V02
(18)
 V0 
The probability density function for Gaussian noise is altered when a target is
present. The radar return adds to the Gaussian noise and the probability density function
becomes Rician. Equation 3.19, from Harney (2013a), is the probability density function
for a non-fluctuating return signal plus noise.
2 V 2
 A2 
V    VA 
pS  N V     e 2V02
I0  2  (19)
 V0   V0 
I0 is a Bessel function of imaginary argument and is a function of (VA/V02). A is

the voltage of the target return. If the CNR of the target return is known, Equation 20 from
Harney (2013a) can substitute CNR for A. After determining the voltage of the target
return, it can be substituted into Equation 20 and the probability density function can be
graphed.
34
A2
CNR  (20)
V0 2
I. VERIFICATION AND VALIDATION OF THE INTEGRATED

SIMULATION MODEL
The Cartwright Expendable Active Decoy Simulator (CEADS) is the result of the
integration of all of the previous models into a single simulator. Each of the aforementioned
models is integrated into the Excel-based simulation on individual tabs. A specific tab
represents each of the following elements of the CEADS: EAD, radar, missile, multipath,
and sea surface clutter. The models are cross-linked so that if a shared value changes on
one tab, it changes on all corresponding models that use the same value. Appendix B
contains examples of each model and Appendix C is the simulator.
CEADS accepts input from the user on the radar, missile, aircraft, EAD, and
environment variables. The performance parameters of the missile, aircraft, EAD,
environment, and Castor 2 J/C radar from Table 3.1 are input into the associated models.
The radar, missile, and EAD models calculate the CNR of the aircraft as seen by the radar
or missile and compare them to the required CNR for a specified detection probability. The
EAD model goes further by also calculating the CNR generated by the EAD given the EAD
performance parameters input by the user. The result is either non-detection of the aircraft
due to the CNR being below the detection threshold, aircraft detection because the CNR is
above the threshold, or aircraft non-detection because the EAD CNR causes a jam-to-signal
ratio of greater than one. It also provides a visual simulation of the engagement showing
the decoy effectiveness. Figure 16 is an example of the visual simulation.
35
Green lines represent the radar. Orange lines represent the missile. Dashed lines are the
upper and lower beam width edges and dotted lines are the beam center.
Figure 16. Visual Simulation of Engagement
In order to verify that the simulator is operating as intended, the author analyzed
the results of a simulator run and compared them to hand calculations. Figure 18 in chapter
IV shows an instantaneous calculation of missile CNR. The author used Equation 2 to
calculate the missile CNR and compared it to the CNR calculated by the simulator. The
result of the hand calculation shows the same missile CNR achieved by the simulator at
the bottom of Figure 18. Similar computations and comparisons for all equations in the
model verify that the equations constructed from a variety of experts in the corresponding
fields are correctly implemented in the simulator and produce appropriate values.
PT LT LR G T A R σe  T R 
α R R
CNR 
16π 2 (kTBF)R T 2 R R 2
(30424)(.52)(.6)(19808.69)(.0154)(3.01)
CNR   93.0
16π (1.38 1023 )(290)(2.2 105 )(6.85)(24367.5)2 (24367.5-11500)2
2
Based on the author’s experience, the outputs provided by the simulator are
reasonable and acceptable. The simulator uses the values to calculate the jam-to-signal
ratio, which it then uses to determine the instantaneous success or failure of the EAD. The
36
simulator demonstrates that increasing EAD power or gain results in a subsequent
instantaneous increase in jam-to-signal ratio, which is to be expected. It then follows that
a greater jam-to-signal ratio results in a greater likelihood of EAD success. As the author
will demonstrate in chapter IV, the data obtained through experimentation supports this
statement. The validation of CEADS proves it to be a useful modeling and simulation based
systems engineering tool that can reduce design and testing time of future EAD systems.
J. MEASURES OF EFFECTIVENESS
The simulation allows a systems engineer to examine the key measure of

effectiveness (MOE) for the EAD system. These key MOEs drive the value of a new system
to the stakeholder and drive the requirements that the systems engineer will set for the
system. The key MOE for this simulator focuses on whether or not the EAD can
successfully drive the missile to a geometry that it is no longer able to engage the aircraft.
This happens when the J/S produced by the EAD is significant enough that the missile
remains focused on the EAD as it falls away from the aircraft until the missile’s range and
altitude no longer permit it to receive the radar returns off the target. At this point, the MOE
declares the simulator run a success and moves on to the next run. If the EAD fails to lure
the incoming missile, then the missile’s focus will remain on the target aircraft and will
eventually strike the target resulting in a failure.
K. FACTORS
The factors that the author believes will have the greatest outcome on the success
of the EAD are those factors that are inherent to the system. These factors are the variables
contained in Equations 8 and 9 for jam-to-signal ratio. Those factors are EAD transmitter
power, EAD loss factor, EAD gain, EAD bandwidth, and EAD deployment range. Other
factors that the author believes to be significant, but are not necessarily captured in any
equations, are EAD vertical and horizontal accelerations, and EAD vertical and horizontal
velocities. These factors will have an effect on how long the EAD remains airborne and
how quickly it can drive the missile to an unfavorable geometry. These 10 factors shown
in Table 4 combine to create the design of experiments (DOE) for CEADS.
37
Table 4. Experimentation Factors
Factor Description Reason for Inclusion

EAD Transmitter Power in watts produced at the
Affects J/S
Power EAD transmitter
Percentage of remaining power
EAD Loss Factor Affects J/S
prior to transmission
EAD energy focused in a specific
EAD Gain Affects J/S
direction
Difference between upper and
EAD Bandwidth Affects J/S
lower cutoff frequencies
Vertical Vertical acceleration on Affects altitude which
Acceleration deployment affects time of engagement
Horizontal Horizontal deceleration on Affects Range which
Acceleration deployment affects J/S
Affects altitude which
Vertical Velocity Vertical velocity on deployment
affects time of engagement
Horizontal Affects Range which
Horizontal velocity on deployment
Velocity affects J/S
Deployment Altitude of the aircraft at Affects time of
Altitude deployment engagement
Range of the EAD from the missile
Deployment Range Affects J/S
at deployment
L. EXPERIMENTATION
This thesis applies computer simulation and experimentation to examine the design
space of the EAD system. Experimentation is a key component of knowledge acquisition.
It is often infeasible or impractical to conduct physical experimentation with a system due
to safety, money, time, or other constraints. In such situations, one may substitute computer
model and simulation to represent the physical system and to understand the complex
interactions between systems and the environment. Often these computer simulations may
be preferred over physical experimentation and may reduce the risks to safety, cost, and
schedule overruns. Computer experimentation allows analysts to assign causality, visualize
interactions and trends, and develop mathematical expressions that relate factors with
outcomes (Koehler and Owen 1996).
38
Sequential experimentation enables the researcher to focus on the most relevant
areas of the design space. As such, a screening experiment first identifies the most relevant
factors that affect the MOE, as well as significant interactions among the factors. The next
set of experiments incorporates a nearly orthogonal Latin hypercube (NOLH) design. A
NOLH allows for efficient exploration of the system design space and provides a good
amount of information about the experimental region. NOLH also allows for simplified
analysis of results though regression analysis and similar methods. This study examines 10
factors that the researcher believes have, or is uncertain about, the influence on the MOE.
The DOE used for this thesis is an 11 x 33 NOLH from Cioppa (2002). Each design point
is unique in its combination of values, and constitutes a singular instantiation of a potential
EAD design. Figure 17 shows an example of the NOLH used in CEADS.
Units Watts dB dB 10 kHZ m/s2 m/s2 m/s m/s m m

low level 1 0 1 1 0 0 0 0 15 1000 1
high level 10 10 10 1000 -9.8 10 -10 -10 300 25000 33
decimals 0 1 0 0 1 1 0 0 0 0 0
Jammer Transmitter Jammer Transmitter Jammer Gain Jammer Bandwidth EAD Initial ↑ EAD Initial → EAD Initial ↑ EAD Initial → EAD Initial EAD Deployment
factor name Power (PJ) Loss Factor (LT) (GJ) (BJ) Accel Accel Velocity Velocity Deployment Altitude Range
DP1 10 0.9 5 188 -8.6 6.3 -7 -5 300 17500 20
DP2 9 10 2 376 -4.6 1.9 -8 -3 273 11500 28
DP3 9 4.4 9 157 -0.3 5.9 -7 0 104 24250 13
DP4 6 8.8 10 407 -9.2 1.6 -8 -1 140 3250 8
DP5 9 0.3 5 220 -6.7 7.2 -4 -6 33 4000 23
DP6 10 9.4 4 282 -4.3 2.2 -2 -9 24 13750 25
DP7 7 4.7 10 251 0 6.6 -4 -9 264 1000 10
DP8 6 6.9 9 344 -8.9 2.5 -2 -10 166 23500 7
DP9 7 2.5 3 532 -7 3.1 0 -2 184 15250 3
DP10 8 6.6 4 688 -2.1 5.3 -1 -4 255 5500 2
DP11 7 2.2 8 969 -3.4 0.6 -1 -2 122 16000 22
DP12 8 7.2 7 938 -7.4 9.7 -5 -4 77 4750 19
DP13 6 1.6 3 563 -5.8 1.3 -10 -8 113 7750 1
DP14 9 5.9 4 875 -1.5 5.6 -9 -7 86 19750 5
DP15 7 1.9 9 906 -3.7 0 -7 -8 220 9250 30
DP16 8 6.3 6 1000 -8 9.1 -6 -7 247 19000 18
DP17 6 5 6 501 -4.9 5 -5 -5 158 13000 17
DP18 1 9.1 6 813 -1.2 3.8 -3 -5 15 8500 14
DP19 2 0 9 625 -5.2 8.1 -3 -7 42 14500 6
DP20 2 5.6 2 844 -9.5 4.1 -3 -10 211 1750 21
DP21 5 1.3 1 594 -0.6 8.4 -2 -9 175 22750 26
DP22 2 9.7 6 781 -3.1 2.8 -6 -4 282 22000 11
DP23 1 0.6 7 719 -5.5 7.8 -8 -1 291 12250 9
DP24 4 5.3 1 750 -9.8 3.4 -6 -1 51 25000 24
DP25 5 3.1 2 657 -0.9 7.5 -8 0 149 2500 27
DP26 4 7.5 8 469 -2.8 6.9 -10 -8 131 10750 31
DP27 3 3.4 7 313 -7.7 4.7 -9 -6 60 20500 32
DP28 4 7.8 3 32 -6.4 9.4 -9 -8 193 10000 12
DP29 3 2.8 4 63 -2.5 0.3 -5 -6 238 21250 15
DP30 5 8.4 8 438 -4 8.8 0 -2 202 18250 33
DP31 2 4.1 7 126 -8.3 4.4 -1 -3 229 6250 29
DP32 4 8.1 2 95 -6.1 10 -3 -3 95 16750 4
DP33 3 3.8 5 1 -1.8 0.9 -4 -3 68 7000 16
Figure 17. Nearly Orthogonal Latin Hypercube Experimental Design
39
M. DATA ANALYSIS
The author will conduct a sensitivity analysis by running multiple designs of

experiments to determine the effects of changing variable ranges has on the probability of
success. The author will then perform regression analysis to isolate significant main factors
and two-way interactions. This regression analysis will determine whether any factors can
be removed from the design of experiments. Finally, the author analyzes the results of the
sensitivity analysis to determine which combination of factor values produce the highest
probability of success, thereby highlighting the desired characteristics of the EAD and
providing the design teams a focus in their efforts.
40
IV. SIMULATOR OPERATION
This chapter discusses the operation of the simulator. For CEADS to act as a true
simulator that produces unique results each time it runs, a certain amount of randomness,
or stochasticity, must be included. Equations 1 and 2 are deterministic, using the time-
averaged values of the variables. However, the CEADS inserts stochastic values into the
equations. The result is a discrete event simulation that incorporates randomness in its
calculations. Specifically, the author chose to implement stochastic characteristics into the
radar transmitter power, radar transmitter losses, missile receiver losses, missile noise
factor, and the aircraft RCS. These variables, while not all inclusive of the possible
stochastic values, represent the variables that the author believes have the greatest effect
on detectability. Follow-on studies to this thesis may examine new variables.
In order to initialize CEADS for a simulation, the user inserts known values into
the respective models. The user can enter nearly all required values into the EAD model.
The EAD model is linked to the radar and missile models so that any shared values
automatically update. For the purposes of this thesis, the author inserted the values from
Table 1 into the radar section of the EAD model. Values not included in Table 1 are the
assumptions stated in the radar and weapon systems model section III.D. Any cells
highlighted in yellow require the user to input a value rather than choose from a slider bar.
Figure 18 shows an example of the EAD model interface and inputs for each variable.
41
Calculate EAD CNR
Radar Transmitter Power (PT) 30,424 Watts 44.832 dB
Radar Transmitter Loss Factor (L T) 52% -2.82 dB 1
Radar Receiver Loss Factor (L R) 63% -2 dB 2
2
Radar Transmitter Area (AT) 0.5542 m Radius (m) 0.42 42
Frequency (f ) 16 GHz 16000
Wavelength (λ) 0.01875 meters
Radar Gain (GT) 19808.691 42.969 dB
2
Radar Receiver Area (AR) 0.5542 m Radius (m) 0.42 7
Radar Side Lobe Level (SLL) 100% 0 dB 100
Radar Receiver Bandwidth (B R) 2.20E+05 Hz
Radar Noise Factor (F) 6 3.981071706 dB
Missile Receiver Loss Factor (L R) 60% -2.24 dB
2
Missile Receiver Area (AR) 0.0154 m Radius (m) 0.07 7
Missile Side Lobe Level (SLL) 100% 0 dB 100
Missile Receiver Bandwidth (B R) 2.20E+05 Hz
Missile Noise Factor (F) 6.85 4.841723676 dB
Jammer Transmitter Power (PJ) 10 Watts 10.000 dB 0.1
Jammer Transmitter Loss Factor (L T) 81% -0.9 dB 9
2
Jammer Transmitter Area (AT) 0.0000885 m Radius (m) 0.00530666
Jammer Gain (GJ) 3.16227766 5 dB
Jammer Bandwidth (BJ) 1.88E+06 Hz 1880 KHz 188
2
Aircraft RCS (σ) 3.01 m 4.786472717 dBsm
Attenuation factor (α) 0
Single Simulator Run Multiple Simulator
Boltzmann's Constant (k) 1.38E-23 J/K
w/ Visualization Runs w/ Data Output
Temperature (T) 290 K
Radar to Target Range (R T) 25000 m 296.15
Radar to Jammer Range (R J) 24587.5 m Deploy EAD Reset EAD 245.875
Missile to Target Range (RT) 13500 m 245.875
Missile to Jammer Range (RJ) 13500 m Match Missile-Target 135
CNRTa rget to Ra da r 283039.860 54.518 dB
CNRJa mmer to Ra da r 26616.29 44.25 dB
J/SJa mmer to Ra da r 0.094037248 EAD Not Successful
CNRTa rget to Mi s s i l e 93.034 19.686 dB
CNRJa mmer to Mi s s i l e 2192.67 33.41 dB
J/SJa mmer to Mi s s i l e 23.56856811 EAD Successful
Figure 18. Inputs to the EAD Model
On the tab for the multipath model, the only inputs required are the radar altitude
and the maximum detection range of the radar. Currently multipath effects do not influence
the success or failure of the EAD; however, it is being included for future development of
the simulator. On the tab for the surface clutter model, the only input necessary is choosing
the sea state and polarization of the radar wave. Similarly, this model is not used to
determine success or failure of the EAD, but is included for future development. Figure 19
shows the required inputs circled in red for the multipath and surface clutter models.
42
Figure 19. Multipath and Sea Surface Clutter Models
The final inputs required are on the DOE tab. The user decides the minimum and
maximum values of each variable. The NOLH determines the appropriate combination of
variables in order to develop a design of experiments. Figure 20 shows an example of the
variables and their low and high levels.

low level 1 0 1 1 0 0 0 0 15 1000
high level 10 10 10 1000 -9.8 10 -10 -10 300 25000
decimals 0 1 0 0 1 1 0 0 0 0
Jammer Transmitter Jammer Transmitter Jammer Bandwidth EAD Initial ↑ EAD Initial → EAD Initial ↑ EAD Initial → EAD Initial Deployment EAD Deployment
factor name Power (PJ) Loss Factor (LT) Jammer Gain (GJ) (BJ) Accel Accel Velocity Velocity Altitude Range
Figure 20. Design of Experiment Variable Levels
Once all input values are entered, the user returns to the EAD tab. The user makes
the choice of a single or multiple runs by clicking one of the buttons shown in the white
section of Figure 18. If the user chooses a single run by clicking “Single Simulator Run w/
Visualization,” CEADS displays a visualization of the engagement in real-time as shown
in Figure 16. If the user chooses to conduct multiple runs by clicking “Multiple Simulator
Runs w/ Data Output,” the simulator prompts the user for the design points to run, the
number of runs for each design point, the radar transmitter power, and aircraft average
RCS. When all runs are complete, CEADS saves the results from each run on individual
tabs allowing for statistical analysis. Figure 21 shows an example of the raw data output.
CEADS records simple statistical analysis to include proportion of runs resulting in success
(mean), standard deviation, and 95% confidence intervals on the “DOE” tab for each design
point shown in Figure 22.
43
Run #1
Deployment Altitude (m) 300 Deployment Range (m) 17500
Range to Target (m) Radar Power (watts) J/S Aircraft RCS (dB) Upper BW (m) Out of Beamwidth Hit?
25000 29406.71676 0 67.01287972 446.1729905 0
24472.5 30235.57639 0 25.81166814 443.0887404 0
23945 30356.52542 0 43.97835731 440.0044903 0
23417.5 30742.7237 0 69.61543361 436.9202402 0
22890 29546.40812 0 13.72556823 433.8359901 0
22362.5 29717.05693 0 23.52555166 430.75174 0
21835 30642.99107 0 18.3487969 427.6674899 0
21307.5 29878.9131 0 21.95889619 424.5832398 0
20780 29521.88438 0 9.04036212 421.4989897 0
20252.5 29298.23846 0 4.241337233 418.4147396 0
19725 29923.93297 0 13.4286951 415.3304895 0
19197.5 29800.83865 0 31.53339836 412.2462394 0
18670 30611.46593 0 11.09096982 409.1619893 0
18142.5 29874.62872 0 43.45113697 406.0777392 0
17615 30024.54662 0 1.907279679 402.9934891 0
17087.5 30075.88315 5.878798813 10.39540681 399.909239 0
16560 29437.2229 0.013607419 37.58492867 396.8249889 0
16032.5 29935.48554 3.42128E-08 93.43828038 393.7407388 0
15505 29613.7771 27.17029696 4.555632561 370.4814887 0
14977.5 30365.1793 3.44194E-09 104.6582286 387.5722386 0
14450 29964.97577 0.214760034 25.5217732 384.4879885 0
13922.5 29563.81673 29.61863308 5.020411727 321.7037384 0
13395 30204.31995 48.07573342 1.318592583 301.1444883 0
12867.5 30423.60997 23.56856811 4.786472717 278.4352382 1
Figure 21. Data Output from Multiple Simulation Runs
Figure 22. Example of Statistical Output on the “DOE” tab
When the user clicks one of the simulator buttons shown in Figure 18, the simulator
prompts the user for the radar transmitter power. A triangular distribution represents the
radar transmitter power centered on the user’s input. For the purposes of this thesis, the
author used the Castor 2 J/C peak power of 30,000 watts with the minimum and maximum
bounds being 29,000 and 31,000 watts respectively. Currently, the simulator is hard-coded
to bound the minimum and maximum values at plus or minus 1,000 watts, but this could
be changed in a future release. A normal distribution represents both the radar transmitter
and missile receiver losses with a mean of -3 dB and a standard deviation of 1 dB. The
missile noise factor is also normally distributed and has a mean of 6 and a standard
deviation of 2. These values are also hard-coded, but could be changed in a future release
of CEADS
An aircraft that exhibits Swerling Case I/II characteristics best represents the RCS
fluctuations with an exponential probability density function. The author integrated the
44
exponential probability density function provided by Harney (2013a) to achieve a
cumulative density function given by Equation 21. The simulator inserts a random number
from 0 to 1 in place of CDF to determine the corresponding RCS (σ). When the user clicks
one of the simulator buttons shown in Figure 18, CEADS prompts the user for the aircraft
average RCS. For the purpose of this thesis, the author assumes the average radar cross
section (σav) to be 24 dBsm as previously calculated in III.E. Figure 23 shows the inputs
requested by CEADS from the user.


 av
CDF  1  e (21)
Figure 23. Inputs Requested by the Simulator for Multiple Runs
45
46
V. SIMULATION EXPERIMENTS AND ANALYSIS
This chapter discusses the results of multiple simulation runs, performs a sensitivity
analysis, and conducts a regression analysis of the data to determine the performance
parameters of interest that are statistically relevant to the success rate of the EAD. CEADS
uses the NOLH DOE to test many possible combinations of the EAD design space against
an individual radar and missile system. This allows for analysis of the data in order to
determine the best combination of performance parameters that maximizes the EAD
effectiveness against a particular threat. Testing the CEADS against the Castor 2J/C radar
provides a realistic scenario for the simulator and demonstrates its potential utility. Beyond
the scope of this thesis, CEADS allows for data collection for a multitude of threat systems,
analysis of data, and aides in developing high-level system requirements.
A. SIMULATION SETUP
The variables chosen as parameters of interest for the NOLH include jammer
transmitter power, jammer transmitter loss factor, jammer gain, jammer bandwidth, EAD
initial vertical and horizontal acceleration, EAD initial vertical and horizontal velocity,
EAD deployment altitude, and EAD deployment range. The author’s experience leads him
to believe that the factors with the greatest effect on outcome are jammer transmitter power
and gain. Therefore, the author began with a sensitivity analysis to explore how the jammer
transmitter power and gain ranges affect the success rate of the EAD. The author kept the
upper and lower bounds of all other variables the same in subsequent DOEs of the
sensitivity analysis.
For the purpose of this thesis, the author completed 20 runs for each of the 33 total
design points. The result of each run is either a “1” if the EAD succeeds in drawing the
missile to an unfavorable engagement geometry or a “0” if the missile hits the target. The
mean and sample standard deviation of all 20 runs is recorded on the DOE tab for each
design point. The simulator uses this data to calculate a 95% confidence interval around
the mean. If analysts desire a different interval width around an estimated mean, Equation
47
22 from Hayter (2014, 341) provides a way to calculate the necessary number of runs or
replications.
2
t s
number _ of _ runs  4   /2,n 1  (22)
 E 
 
B. SENSITIVITY ANALYSIS
The first DOE of the sensitivity analysis began with a minimum transmitter power
of 1 watt and a maximum of 10 watts. The minimum gain was 1 dB and maximum of 10
dB. This DOE will be the baseline as the author assumes that range of values contains the
minimum values for the EAD to achieve success. All following DOEs will stack the
designs, meaning the power and gain ranges will change, but all other baseline variable
ranges between DOEs. As an example, the baseline EAD loss factor range is from 1 to 10.
All following DOEs will use this same range. According to Hernandez (2008), stacking
NOLHs maintains the orthogonality of the overall design. Table 5 displays the results of
the first DOE of the sensitivity analysis and an example of the statistical data output. The
outcome shows that the design point exhibiting the greatest success rate is DP 33 with a
power of 3 watts and gain of 5 dB resulting in 80% success.
48
Table 5. Statistical Data of Design of Experiments 1
low level 1 0 1 1 0 0 0 0 15 1000 1
high level 10 10 10 1000 -9.8 10 -10 -10 300 25000 33
decimals 0 1 0 0 1 1 0 0 0 0 0
Jammer Transmitter Jammer Transmitter Jammer Bandwidth EAD Initial ↑ EAD Initial → EAD Initial ↑ EAD Initial → EAD Initial Deployment EAD Deployment
factor name Power (PJ) Loss Factor (LT) Jammer Gain (GJ) (BJ) Accel Accel Velocity Velocity Altitude Range
DP1 10 0.9 5 188 -8.6 6.3 -7 -5 300 17500 20
Statistics Mean 0.55 Standard Dev 0.510417786 95% Upper 0.773700458 95% Lower 0.326299542 # of Runs 20
DP2 9 10 2 376 -4.6 1.9 -8 -3 273 11500 28
Statistics Mean 0.15 Standard Dev 0.366347549 95% Upper 0.310558892 95% Lower -0.010558892 # of Runs 20
DP3 9 4.4 9 157 -0.3 5.9 -7 0 104 24250 13
DP4 6 8.8 10 407 -9.2 1.6 -8 -1 140 3250 8
DP5 9 0.3 5 220 -6.7 7.2 -4 -6 33 4000 23
DP6 10 9.4 4 282 -4.3 2.2 -2 -9 24 13750 25
Statistics Mean 0 Standard Dev 0 95% Upper 0 95% Lower 0 # of Runs 20
DP7 7 4.7 10 251 0 6.6 -4 -9 264 1000 10
DP8 6 6.9 9 344 -8.9 2.5 -2 -10 166 23500 7
DP9 7 2.5 3 532 -7 3.1 0 -2 184 15250 3
DP10 8 6.6 4 688 -2.1 5.3 -1 -4 255 5500 2
DP11 7 2.2 8 969 -3.4 0.6 -1 -2 122 16000 22
DP12 8 7.2 7 938 -7.4 9.7 -5 -4 77 4750 19
DP13 6 1.6 3 563 -5.8 1.3 -10 -8 113 7750 1
DP14 9 5.9 4 875 -1.5 5.6 -9 -7 86 19750 5
DP15 7 1.9 9 906 -3.7 0 -7 -8 220 9250 30
DP16 8 6.3 6 1000 -8 9.1 -6 -7 247 19000 18
DP17 6 5 6 501 -4.9 5 -5 -5 158 13000 17
DP18 1 9.1 6 813 -1.2 3.8 -3 -5 15 8500 14
DP19 2 0 9 625 -5.2 8.1 -3 -7 42 14500 6
DP20 2 5.6 2 844 -9.5 4.1 -3 -10 211 1750 21
DP21 5 1.3 1 594 -0.6 8.4 -2 -9 175 22750 26
DP22 2 9.7 6 781 -3.1 2.8 -6 -4 282 22000 11
DP23 1 0.6 7 719 -5.5 7.8 -8 -1 291 12250 9
DP24 4 5.3 1 750 -9.8 3.4 -6 -1 51 25000 24
DP25 5 3.1 2 657 -0.9 7.5 -8 0 149 2500 27
DP26 4 7.5 8 469 -2.8 6.9 -10 -8 131 10750 31
DP27 3 3.4 7 313 -7.7 4.7 -9 -6 60 20500 32
DP28 4 7.8 3 32 -6.4 9.4 -9 -8 193 10000 12
DP29 3 2.8 4 63 -2.5 0.3 -5 -6 238 21250 15
DP30 5 8.4 8 438 -4 8.8 0 -2 202 18250 33
DP31 2 4.1 7 126 -8.3 4.4 -1 -3 229 6250 29
DP32 4 8.1 2 95 -6.1 10 -3 -3 95 16750 4
DP33 3 3.8 5 1 -1.8 0.9 -4 -3 68 7000 16
The goal of the second and third DOEs of the sensitivity analysis was to determine
the effect of increasing the EAD transmitter power on the success rate of the EAD. DOE 2
maintained the same minimum and maximum EAD gain, but increased the minimum
power to 10 watts and maximum to 100 watts. DOE 3 also maintained a fixed gain, but the
power range was from 100 to 500 watts. Again, the baseline DOE variables stack in DOE
2. Figure 24 shows a graphical composite of all DOE success rates based on power.
Visually, one can determine that there is a clear correlation between increased power and
49
increased success rate. While DOE 1 has nearly 60% of all EAD configurations having less
than 50% chance of success, DOE 3 where the range of power is from 100 to 500 watts,
nearly 70% of the EAD configurations are above the 50% success rate.
Figure 24. Effect of Decoy Power on the Probability of Success
Table 6 displays a few of the interesting results of DOE 2 of the sensitivity analysis.
When compared to the results of the baseline DOE in Table 5, the results show that
increasing power by approximately ten times can result in as much as a 50% increase in
success, as demonstrated by DP 11 and DP 16. In another instance, DP 20 and 22, which
previously had a 0% success rate, now have a 25% and 15% chance of success respectively.
The average overall success rate of the baseline DOE is only 28.5%. DOE 2 has an average
success rate of 46.1%.
50
Table 6. Noteworthy Results from Design of Experiments 2
Table 7 displays a few interesting results from DOE 3 of the sensitivity analysis.
increasing power by approximately 50 times can result in as much as a 60% increase in
success, as demonstrated by DP 4. DP 1, 3, and 4 has the highest success rates calculated
thus far. DP 20, which had a 0% success rate in the baseline DOE, now has a 45% chance
of success and DP 22 has a 40% probability of success.
When compared to DOE 2, nearly every design point saw an increase in probability
of success from between 5% up to 40%. The average probability of success for DOE 3 is
58.9% compared to DOE 2 of 46.1% and the baseline DOE of 28.5%. DP 16, which
experienced the greatest improvement in success between the baseline and DOE 2, only
experienced an increase of 5% even though the power was increased by nearly 500%. This
might indicate that further increasing power for these design points will have no further
effect on the probability of success; there is a limit for increasing the probability of success
by only increasing power. Another factor must be investigated to determine if probability
of success can be further increased.
51
The goal of DOE 4 and 5 of the sensitivity analysis was to determine the effect that
only increasing gain would have on the probability of success. EAD transmitter power
resets to the baseline of 1 to 10 watts. DOE 5 used a gain range of 10 to 30 dB and DOE 6
used a range of 30 dB to 50 dB. All other variable ranges remained fixed. Figure 25 shows
a graphical composite of all DOE success rates based on gain. Visually, one can determine
that there is a clear correlation between increased gain and increased success rate. While
DOE 1 has nearly 60% of all EAD configurations having less than 50% chance of success,
DOE 5 where the range of gain is from 30 to 50 dB, nearly 85% of the EAD configurations
are above the 50% success rate.
52
Figure 25. Effect of Decoy Gain on Probability of Success
Table 8 displays a few of the interesting results of DOE 4 of the sensitivity analysis.
increasing gain from 2 dB to 13 dB can result in an impressive 85% increase in success, as
demonstrated by DP 2, which achieved a 100% probability of success. DP 16 saw an
increase from 5% to 70% when increasing gain by 16 dB. DP 20 and 22, which previously
had a 0% success rate, now have a 35% and 45% chance of success respectively.
Table 9 displays a few of the interesting results from DOE 5 of the sensitivity
analysis. When compared to the results of the baseline DOE in Table 5, the results show
53
that increasing the gain range to between 30 dB and 50 dB increases probability of success
across all design points except for those with 0% probability. DP 25 is the lowest success
rate at 55% while all other design points are greater than or equal to 70% probability of
success. DP 20, which had a 0% success rate in the baseline DOE, now has a 75% chance
of success and DP 22 has a 70% probability of success.
When compared to DOE 4, nearly every design point saw an increase in probability
of success with 12 of the design points having 90% or greater probability of success. DP 3
and DP 8 achieved a 100% probability of success, which is an improvement over both
DOEs. The average probability of success for DOE 5 is 71.3% compared to DOE 4 of
57.7% and the baseline DOE of 28.5%. It appears clear that increasing gain provides for
greater probability of success across the entire range of design points when compared to
only increasing power.
One thing that becomes clear from the sensitivity analysis is that, regardless of the
EAD system characteristics, altitude and range, with regard to gain, have an effect on the
success of the EAD. It is clear from the data that the design points with a low altitude
deployment (< 60 m) and a long range from the missile experience failure 100% of the
time. This is because the EAD does not have the altitude to drive the missile to an
unfavorable geometry. Instead, the EAD descends on the surface with the target aircraft
remaining within the missile’s field of view. This allows time for the missile to reacquire
the target. DP 5 demonstrates the only time when a low altitude deployment might succeed.
This design point has a deployment altitude of 33 meters; however, the deployment range
54
from the missile is only 4000 meters. This allows the EAD to succeed in pulling the missile
off the target aircraft prior to surface impact. This indicates that deployment tactics,
techniques, and procedures can have an effect on the success of the EAD. A potential
deployment technique might be for the pilot flying at low altitude to wait until the missile
is closer before deploying the EAD.
C. REGRESSION ANALYSIS
Following the sensitivity analysis, the author performed regression analysis on the
output from the simulation runs to determine which factors were statistically significant to
EAD success or failure. Figure 26 displays the results of the regression analysis at the 0.10
significance level. Not all primary factors are statistically significant on their own;
significant two-way interactions capture primary factors that must still be considered in the
analysis. Notably, the two factors which were explored as the most likely candidates for
having an effect on success are absent from the primary factors. However, their two-way
interaction is statistically significant, which indicates that their primary effects must be
considered in the overall design of the EAD. The two-way interaction between loss and
horizontal acceleration has the greatest effect on EAD success. Examining each factor
shows that they are both independently significant.
Loss factor is significant because it represents the percent of EAD power remaining
after experiencing transmitter losses due to inefficiency. For example, a 3 dB loss factor
represents approximately 50% of power remaining. Because the loss factor is logarithmic
in nature, going from -2 dB to -3 dB represents going from 63% power remaining to 50%
power remaining. This small change can have a big impact on jam-to-signal which depends
on the loss factor. As such, the loss factor is a conduit for the effects of power, which may
explain why power does not show as a significant primary factor. A systems engineer must
examine this factor closely to determine how to make the energy transmission as efficient
as possible to minimize the loss factor.
Horizontal acceleration is significant because it represents the change in velocity

of the EAD after deployment. A high acceleration value represents a rapid deceleration of
the EAD following deployment. According to Equation 9, the jam-to-signal ratio (J/S)
55
increases with the square of the receiver to target range (RR). Therefore, the greater the
range to the EAD, the higher the J/S will be. A higher acceleration value represents a rapid
deceleration of the system. The system decelerates until it reaches a horizontal velocity of
zero which maximizes its range from the missile with respect to time. This maintains a
high J/S for a longer period, which results in a higher probability of success. This might
suggest that a systems engineer find a way to slow the EAD down as fast as possible
following deployment.
Similarly, vertical acceleration is the next most significant factor captured by the
two-way interactions between it and horizontal acceleration and loss factor. Vertical
acceleration is significant because it has a direct effect on both duration of engagement and
how rapidly the missile can be pulled to an unfavorable geometry. Vertical acceleration
affects the duration of the engagement by how quickly the EAD falls away from the
aircraft. The faster it accelerates, the faster it falls, and the sooner it impacts the water. If
this happens too quickly, and the aircraft is still in the missile’s beam width, then the EAD
fails. This becomes the focus of a systems engineering trade-off. If the EAD accelerates
too quickly away from the aircraft, it could fail to draw the missile away before it impacts
the water. If it accelerates too slowly, then the target remains in the missile’s beam width
and there is a chance that it could reacquire the target.
These significant factors are important attributes that must be incorporated in any
design alternatives for a new EAD because they drive the capability requirements of the
system. These are some, but not all, of the factors that engineers must consider to ensure a
future EAD system is successful. Further analysis is required to determine the correct
combination of factors that allow for maximum probability of success. Other
considerations such as ability to produce power may limit factors such as maximum power
output. Another trade-off may consider the requirement of a high-gain EAD needing to
point its antenna directly at the threat and the problems associated with localization of said
threat.
56
Figure 26. Regression Analysis Results
The data that resulted from the sensitivity analysis demonstrates that CEADS
produces statistically relevant results. The author collected the results of the sensitivity
analysis and sorted them to show which combinations of variables produced the highest
probability of success. Table 10 shows those design points that achieved a probability of
success of 90% or greater. The data shows that either a low power setting with a high gain,
or a high power with a low gain, is necessary to achieve high probability of success.
Notably none of the designs with power in the 10 – 100-watt range succeeded in achieving
a probability of success of 90%.
57
DP 3 had the most instances of high probability of success achieving 90% or greater
in three out of the five DOEs. It seems to be a prime candidate for consideration of the
EAD capability requirements. It demonstrates that an EAD with a power of less than 10
watts can still achieve success when coupled with a gain of 28 dB or greater. If power is
increased to 450 watts then a gain of 9 dB will suffice. Additional reasons the EAD was
able to succeed may be deduced from those factors that were identified as the major drivers
during regression analysis. DP 3 has a loss factor of 4.4, which equates to 36% power
remaining. This indicates that only 3.25 watts of power are transmitted, which seems low
and would negatively affect the probability of success. However, the horizontal
acceleration value is rather high which indicates rapid deceleration and works in favor of
increasing probability of success. This rapid deceleration, coupled with zero horizontal
velocity added at deployment, and a deployment range of 24,500 meters, combine to
improve the jam-to-signal ratio. Lastly, the low vertical acceleration coupled with the long
range and 104-meter high deployment work in favor of the EAD. Their combination
provides enough time to draw the missile off the target. These results will be crucial to the
development of system level requirements for a future EAD.
Table 10. Design Points with Greater than 90% Probability of Success
Decoy Power Loss Factor Decoy Gain Decoy Bandwidth Vert Accel Horiz Accel Vert Vel Horiz Vel Altitude Range Success
DP 2 9 10 13 376 -4.6 1.9 -8 -3 273 11500 1
DP 3 9 4.4 48 157 -0.3 5.9 -7 0 104 24250 1
DP 8 6 6.9 49 344 -8.9 2.5 -2 -10 166 23500 1
DP 1 500 0.9 5 188 -8.6 6.3 -7 -5 300 17500 1
DP 4 325 8.8 10 407 -9.2 1.6 -8 -1 140 3250 1
DP 7 7 4.7 49 251 0 6.6 -4 -9 264 1000 0.95
DP 23 1 0.6 44 719 -5.5 7.8 -8 -1 291 12250 0.95
DP 26 4 7.5 46 469 -2.8 6.9 -10 -8 131 10750 0.95
DP 33 3 3.8 38 1 -1.8 0.9 -4 -3 68 7000 0.95
DP 3 450 4.4 9 157 -0.3 5.9 -7 0 104 24250 0.95
DP 8 313 6.9 9 344 -8.9 2.5 -2 -10 166 23500 0.95
DP 3 9 4.4 28 157 -0.3 5.9 -7 0 104 24250 0.9
DP 29 3 2.8 17 63 -2.5 0.3 -5 -6 238 21250 0.9
DP 2 9 10 33 376 -4.6 1.9 -8 -3 273 11500 0.9
DP 4 6 8.8 50 407 -9.2 1.6 -8 -1 140 3250 0.9
DP 5 9 0.3 39 220 -6.7 7.2 -4 -6 33 4000 0.9
DP 11 7 2.2 45 969 -3.4 0.6 -1 -2 122 16000 0.9
DP 15 7 1.9 47 906 -3.7 0 -7 -8 220 9250 0.9
DP 32 4 8.1 33 95 -6.1 10 -3 -3 95 16750 0.9
DP 33 175 3.8 5 1 -1.8 0.9 -4 -3 68 7000 0.9
58
VI. CONCLUSION AND RECOMMENDATIONS
A. CONCLUSIONS
From the results of the analysis, the author recommends that all significant factors
be explored in a tradeoff analysis to determine the combination that results in the highest
probability of success against all threat systems. The focus of design should be on a
combination of high power, low loss factor, high gain, low bandwidth, medium vertical
acceleration and velocity, high horizontal acceleration (deceleration), zero horizontal
velocity, early deployment, and medium to high altitude.
J/S is linearly dependent on most of the aforementioned factors. As noted by the

sensitivity analysis, producing either high power or high gain will increase the J/S produced
by the EAD. The author did not examine generating both simultaneously, but the results of
the sensitivity analysis prove that having both high power and high gain is not required for
high probability of success. A tradeoff exists here as previously mentioned. High power
production requires enough battery power to generate the necessary wattage. A large
battery will take up space in an already space-constrained system. This could be overcome
by using a smaller battery with greater energy density. Similarly, a large gain will require
that the EAD be able to geo-locate the threat and point its main antenna in the threat’s
direction. This adds complexity to the design of the system. Therefore, systems engineers
must consider the tradeoff between generating higher power outputs using batteries that
take up more space, or the engineering difficulty of designing the system to leverage high
gain by geo-locating the threat. Doing both simultaneously will cost more time and money
with no added benefit.
High horizontal acceleration and zero horizontal velocity on deployment will

ensure that the EAD creates a large miss distance by ensuring the EAD remains at the
maximum engagement distance at which it was deployed. This allows the aircraft to fly out
of the radar resolution cell while the threat remains focused on the EAD creating a large
miss distance. This capability requirement could manifest in the design of the dispenser,
which could deploy the EAD directly perpendicular to the flight path of the aircraft;
59
ensuring zero added horizontal velocity on deployment. Some form of drag device such as
a parachute, streamer, or drag fins could then be used to ensure a rapid deceleration.
Lastly, the author concludes that one of two deployment schema should be
considered. The first is a medium-to-high altitude engagement with an early deployment
of the EAD at long range. This serves to allow maximum time for the EAD to draw the
missile away from the aircraft and leverages the positive effects of the square of the range
term on J/S. Keeping in mind that a helicopter’s flight profile is typically relegated to 300
meters or less, the sensitivity analysis determined that the design points with low altitude
deployment usually resulted in a 0% probability of success. Those design points that
succeeded with a low altitude deployment did so by waiting until the missile was closer
before deploying. This allows the missile to be drawn away before the EAD impacts the
water. By analyzing DP 5 and 33, which had a high level of success, the author concludes
that when deploying the EAD at a low altitude (<70 meters), the operator should use a
range to altitude ratio of 100 to 1. For example, DP 33 had a deployment altitude of 68
meters and a range of 7000 meters resulting in a high probability of success. Again, a
tradeoff exists here as a system engineer could design the system with a low vertical
velocity and acceleration, which would allow low altitude deployments at longer ranges.
This has the beneficial effect of requiring only one deployment schema but further tradeoff
analysis should be conducted to determine the effect on engagement outcomes.
B. RECOMMENDATIONS
The first recommendation is that subject matter experts at military organizations

such as NAVAIR validate CEADS to ensure that the data provided is reasonable and that
CEADS provides a sound MSBSE tool to aid in the development of the capability
requirements of a future EAD system. Next, NAVAIR should use CEADS in the
acquisition strategy for a new EAD system. The modeling and simulation based approach
may be leveraged to save time and money and is the first step on the path to systems
acquisition. Though the results obtained in Table 10 are for a single threat system, these
design points may prove to be a launching point for future system design space exploration.
The author recommends that the stakeholders at NAVAIR explore the combinations of
60
factors that produced the highest probability of success when researching new threat
systems.
The author recommends that another student continue the research conducted by
this thesis. There is further benefit in exploring the results against a multitude of threat
systems so that a more complete picture may be obtained of the performance required of a
new EAD system. This research could go on to include a cost-benefit analysis to determine
the optimum combination of performance parameters values that achieve the greatest
probability of success at the lowest cost.
The author recommends a few possible improvements to the simulator that would
improve its ability to deliver relevant results. The first recommendation is to implement
Doppler. CEADS only incorporates mono-pulse radar characteristics in the radar model.
Incorporating Doppler will provide greater realism and allows for modeling more
sophisticated threat systems. Incorporating other types of radar such as continuous wave
may also be possible and would improve realism as well as allow for analysis against a
wider range of threat systems. Another improvement could include modeling an active
missile with its own onboard radar since CEADS currently only models semi-active
missiles.
The next recommendation involves implementing a method for the aircraft to

execute a maneuver following EAD deployment. In a real-world engagement, a maneuver
often accompanies any kind of countermeasure deployment. These maneuvers are designed
to reduce aircraft RCS and to exit the radar cell while the threat is focused on the
countermeasure. The individual aspects of the maneuver such as rate of turn could be
entered as a design point variable. This would not only have an effect on the outcome of
the engagement but could drive system deployment doctrine as well. The author has
already determined that a low altitude deployment at long range should be avoided.
The author also recommends implementing slant range into future releases of
CEADS. In an effort to simplify the model, the author assumed that horizontal and slant
range distances were roughly equal due to the low altitudes of the engagement. This may
not prove true if exploring the effects at high altitude engagements.
61
Lastly, improving efficiency of the coding will help in improving the ability to
collect large amounts of data. The author is by no means a coding expert and taught himself
Microsoft Visual Basic coding while working on the thesis. The author assumes that his
code is not as efficient as it could be. The author found that the ability to collect data was
the choke point of the thesis. Running CEADS on a supercomputer or conducting parallel
computing could also help with collecting data faster. Another recommendation to improve
efficiency is to remove the radar and missile models completely and instead roll them all
into the EAD tab. This would prevent much of the cross-linked computation that occurs
and could speed up computation times.
This study will inform the aviation community’s acquisition strategy for an
improved EAD. Constantly improving radar technologies have limited the effectiveness of
currently employed airborne countermeasures and could one day render LO technology
obsolete. Additionally, the increasing ranges seen from recent threat systems now put non-
fighter aircraft at risk of being detected and engaged far away from hostile environments.
These facts, coupled with the massive improvements in battery energy density and
miniaturization of electronics such as DRFM have made it possible to develop small,
lightweight, and smart EADs that greatly improve the capabilities of today’s aircraft to
defend themselves. There is an opportunity to address this issue now. A countermeasure
that is agile, adaptive, and intelligent, which is employable on all aircraft across all
services, is necessary to ensure that the United States may continue to dominate hostile
airspace around the world.
62
APPENDIX A. EQUATION VARIABLE DEFINITIONS
𝛼 Atmospheric Attenuation Coefficient (m-1)

α+β Multipath Beam Width (radians)
ΔR Radar Range Resolution (m)
ε Efficiency of Matched Filter (%)
ηb Hits per Scan (unitless)
θGA Grazing Angle (radians)
θb Azimuth Beam Width (deg)
θBW Elevation Beam Width (radians)
s Scan Rate (deg/s)
λ Wavelength (m)
σ Radar Cross Section (m2)
σav Average Radar Cross Section (m2)
σ0 Mean Backscatter Coefficient (dBsm/m2)
σSC Sea Surface Clutter Radar Cross Section (m2)
τ Pulse Duration (s)
A Target Voltage Return (volts)
AR Receiver Area (m2)
ax Acceleration along x-axis (m/s2)
ay Acceleration along x-axis (m/s2)
B Radar Bandwidth (Hz)
c Speed of Light (m/s)
CNR Carrier-to-noise Ratio (unitless or dB)
E Statistical Margin of Error
F Receiver Noise Factor (unitless)
fp frequency (Hz)
GT Transmitter Gain (unitless)
GJ Jammer Gain (unitless)
HR Height of Receiver (m)
HT Height of Target (m)
J/S Jam-to-signal Ratio (unitless)
k Boltzmann’s constant (J/K)
LR Receiver Loss Factors (%)
LT Transmitter Loss Factor (%)
PD Probability of Detection (unitless)
Pfa Probability of False of Alarm (unitless)
PR/P0 Multipath Detection Ratio (unitless)
PT Transmitter power (watts)
R Range (m)
R0 Initial Range (m)
RDET Radar maximum Detection Range (m)
RR Receiver to Target Range (m)
RT Transmitter to Target Range (m)
63
s Sample Standard Deviation
SLL Side Lobe Level (%)
T Electronics Temperature (K)
t time (s)
tα/s,n-1 student-t distribution (unitless)
Tfa Time Between False Alarm (s)
V Voltage of Noise Return (volts)
V0 Root Mean Squared Noise Level ((V/2)1/2)
Vx Velocity along x-axis (m/s)
Vy Velocity along y-axis (m/s)
64
APPENDIX B. MODELS
Calculate Radar CNR
Radar Probability of Detection
Radar Transmitter Power (PT) 30,000 Watts 44.771 dB 30
Radar Transmitter Loss Factor (L T) 63% -2 dB 20 1
2 0.9 Noise Rayleigh
Radar Transmitter Area (AT) 0.5542 m Radius (m) 0.42 42
0.8 Target CNR 70.879
Wavelength (λ) 0.01875 meters 0.7 Sea Surface Clutter CNR 25.101
Radar Gain (GT) 19808.691 42.969 dB 0.6 Detection Threshold

2
Radar Receiver Area (AR) 0.5542 m Radius (m) 0.42 42
P(V)
0.5
Aircraft RCS (σ) 251.19 m2 24 dBsm 24
0.4
Boltzmann's Constant (k) 1.38E-23 J/K
Temperature (T) 290 K 0.3
Radar Receiver Bandwidth (BR) 2.20E+05 Hz 0.2

Radar Noise Factor (F) 6 3.981071706 dB 0.1
Radar to Target Range (RT) 30000 m 300
0
Attenuation factor (α) 0 -1 1 3 5 7 9 11
CNR 48738.226 46.879 dB V/V0
Pulses to Integrate (Detection of Signals in Noise eq. 10.18)
Multipath Effects On Radar Detection
Pulse Width (τ) 0.000002 s 20
Pulse Repitition Frequency (f p ) 7200 pps 7200 1000000
Scan Rate (ωm) 18 rpm 18 100
Beam Azimuth (θb ) 1.50 deg 150 0.01 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000
Power Relative to Minimum Detectable Power

Beam Elevation (θ e ) 0.67 deg 67 1E-06
Hits per scan (ηb ) 100 1E-10
Calculate Bandwidth of Matched Filter and P fa (Detection of Signals in Noise eq. 10.17) 1E-14
Matched Filter Assumption (Bτ) 0.44 (See Table Below) 1E-18
Radar Receiver Bandwidth (BR) 2.20E+05 Hz 0.2200 MHz 1E-22
Time Between False Alarms (Tfa ) 120 min 1E-26
False Alarm Probability (Pfa ) 6.31E-10 1E-30
Benefit from Integrating 100 pulses for Swerling Case I/II (Chart) 1E-34
24 dB 1E-38
CNR Needed to Detect Swerling Case I/II Target 1E-42
22.982 dB Calculate CNR 1E-46
Calculated CNR 46.879 dB 1E-50

1E-54
Sea Clutter CNR 25.101 dB
1E-58
Total CNR 70.879 dB
1E-62
Multipath a factor? Yes Range (m)
Detected?: Not Detected Multipath
Figure 27. Radar Model
65
Radar Multipath Effect Calculation
α 0.229182 1E-59 30000
Multipath Effects on Radar Detection
β 0.343771 1E+09 30000
α+β 0.572952
1000000
Wavelength (λ) 0.01875 m 100
Power Relative to Minimum Detectable Power (P/P DET)

k (2π/λ) 335.1032 0.01 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000
Max Detection Range (Rdet) 30000 m R (m) P/P
1E-060
Target Height (HT) 150 m 01E-10 0 0
Radar Height (HR) 30 m 1001E-14 1 111.1398
Radar to Target Range (R) 30000 m 2001E-18 1 72.95097
Detection Ratio (P/PDET) 2.36E-58 Is Multipath a Factor? 3001E-22 1 52.76517
Multipath a Factor 4001E-26 1 40.92699
5001E-30 1 33.29461
6001E-34 1 28.00918
7001E-38 1 24.14835
8001E-42 1 21.21115
9001E-46 1 18.90458
10001E-50 1 17.04675
11001E-54 1 15.51914
12001E-58 1 14.24136
13001E-62 1 13.15704 Range (m)
1400 1 12.2255
Figure 28. Multipath Model
66
System of Interest
PF PD PF PD PF PD PF PD PF PD CNR Converted
6.31E-10 0.00 6.31E-10 0.00 1.00E-08 0.00 1.00E-10 0.00 1.00E-12 0.00 0 1 Swerling Case II
0.00 0.00 0.00 0.00 0.00 1.002 1.584893192
0.00 0.00 0.01 0.00 0.00 4 2.511886432
0.90
0.01 0.01 0.02 0.01 0.00 6 3.981071706
0.06 0.06 0.08 0.04 0.02 0.808 6.309573445
0.15 0.15 0.19 0.12 0.08 10 10
0.70
0.28 0.28 0.34 0.25 0.19 12 15.84893192
0.44 0.44 0.49 0.41 0.35 0.60
14 25.11886432
0.60 0.60 0.64 0.57 0.51 16 39.81071706
PD
0.50
0.72 0.72 0.75 0.70 0.65 18 63.09573445
0.81 0.81 0.83 0.80 0.76 0.40
20 100
0.88 0.88 0.89 0.87 0.84 22 158.4893192
0.30
0.92 0.92 0.93 0.91 0.90 24 251.1886432
0.95 0.95 0.95 0.94 0.93 0.20
26 398.1071706
0.97 0.97 0.97 0.96 0.96 28 630.9573445
0.10
0.98 0.98 0.98 0.98 0.97 30 1000
0.99 0.99 0.99 0.99 0.98 0.00
32 1584.893192
0.99 0.99 0.99 0.99 0.99 34 02511.886432 10 20 30 40
0.99 0.99 1.00 0.99 0.99 36 3981.071706 CNR (dB)

1.00 1.00 1.00 1.00 1.00 38 6309.573445
Radar Pfa Missile Pfa 1.00E-08 1.00E-10 1.00E-12
1.00 1.00 1.00 1.00 1.00 40 10000
Calculate CNR for P D Calculate CNR for P D Calculate CNR for P D Calculate CNR for P D Calculate CNR for P D
0.90 0.90 0.90 0.90 0.90
90 0.90 0.90 0.90 0.90 0.90
Calculated CNR Calculated CNR Calculated CNR Calculated CNR Calculated CNR
22.98151676 22.98151675 22.36784679 23.34752815 24.14690074
Calculate PD for CNR Calculate PD for CNR Calculate P D for CNR Calculate PD for CNR Calculate PD for CNR
CNR PD CNR PD CNR PD CNR PD CNR PD
23 0.90 23 0.90 23 0.91 23 0.89 23 0.87
Figure 29. Swerling Case I/II Receiver Operating Characteristics
67
Radar Sea Surface Clutter Normalized Mean Sea Backscatter Coefficient σ0 for Grazing Angle of 0.1°
2
Sea State/Radar Polarization 2H Reflection Coefficient in dBsm/m for Given Frequency (Ghz)
Target Height (HT) 150 m Sea State/Radar Polarization 3 5.6 9.3 17
Radar Height (HR) 30 m 0V - - - -
Closest Radar Frequency 17 GHz 0H -90 -87 - -
Radar to Target Range (R) 30000 m 1V -80 -72 -65 -
Beam Elevation (θ e ) 0.011693706 rad 1H -80 -75 -71 -
Radar Range Resolution (ΔR) 300 m 2V -75 -67 -56 -
Grazing Angle -0.105818104 deg 2H -75 -67 -59 -48
2
Normalized Mean Sea Backscatter Coefficient (σ 0) -48 dBsm/m 3V -75 -60 -51 -
2
σ 1.667994751 m 2.221947 dBsm 3H -68 -69 -53 -
CNRSC 323.6416442 25.10064 dB 4V -67 -58 -48 -
4H -63 -60 -48 -
5V -63 -55 -44 -
5H -63 -58 -42 -
6V -46 - - -
6H - - - -
Normalized Mean Sea Backscatter Coefficient σ0 for Grazing Angle of 0.3°

2
Reflection Coefficient in dBsm/m for Given Frequency (Ghz)
Sea State 3 5.6 9.3 17
0V - - - -63
0H -83 -79 -74 -
1V -64 -60 -58 -54
1H -74 -68 -66 -58
2V -62 -55 -52 -52
2H -66 -60 -56 -53
3V -58 -50 -45 -47
3H -60 -50 -46 -42
4V -57 - -43 -44
4H -55 - -42 -39
5V -52 - -39 -39
5H -52 -44 -39 -38
6V - - -34 -37
6H -46 - -34 -37
Figure 30. Sea Surface Clutter
68
Calculate EAD CNR
Probability of Detection
Radar Transmitter Power (PT) 30,000 Watts 44.771 dB
Radar Transmitter Loss Factor (L T) 63% -2 dB 1 1
Noise Rayleigh
2 0.9 Missile Target CNR 31.316
Radar Transmitter Area (AT) 0.5542 m Radius (m) 0.42 42 Missile Jammer CNR -
0.8 Missile Sea Surface Clutter CNR 4.538
0.7 Radar Target CNR 70.879
Wavelength (λ) 0.01875 meters
Radar Jammer CNR -
Radar Gain (GT) 19808.691 42.969 dB 0.6 Radar Sea Surface Clutter CNR 25.101
2
Radar Receiver Area (AR) 0.5542 m Radius (m) 0.42 7 Detection Threshold
P(V)
0.5
Radar Side Lobe Level (SLL) 100% 0 dB 100
0.4
Radar Receiver Bandwidth (BR) 2.20E+05 Hz
Radar Noise Factor (F) 6 3.981071706 dB 0.3
Missile Receiver Loss Factor (L R) 63% -2 dB 0.2

2
Missile Receiver Area (AR) 0.0154 m Radius (m) 0.07 7 0.1
Missile Side Lobe Level (SLL) 100% 0 dB 100
0
Missile Receiver Bandwidth (BR) 2.20E+05 Hz -1 1 3 5 7 9 11
Missile Noise Factor (F) 6 3.981071706 dB V/V0
2 2
Jammer Transmitter Power (PJ) 1,000 Watts 30.000 dB 10 Altitude (m) Range (m) → Velocity (m/s) ↑ Velocity (m/s) → Accel (m/s ) ↑ Accel (m/s )
Jammer Transmitter Loss Factor (L T) 63% -2 dB 20 Aircraft 150 30000 -55
2
Jammer Transmitter Area (AT) 0.0000352 m Radius (m) 0.00334828 Radar 30 0
Jammer Gain (GJ) 1.258925412 1 dB Missile 0 0 1000 1000
Jammer Bandwidth (BJ) 1.00E+06 Hz 1 MHz 100 EAD Initial 150 30000 0 0 0 -1
Aircraft RCS (σ) 251.19 m2 24 dBsm EAD 0 0 0 0 0 0
Attenuation factor (α) 0 200
Boltzmann's Constant (k) 1.38E-23 J/K Radar Beam
180
Temperature (T) 290 K 0 30 30 30
Radar to Target Range (RT) 30000 m 296.15 30000 150 -25.40759 325.4075886
160
Radar to Jammer Range (RJ) 30000 m Deploy EAD Reset EAD 300 Missile Beam
Missile to Target Range (RT) 30000 m 300 140 0 0 0 0
Missile to Jammer Range (RJ) 30000 m Match Missile-Target 300 30000 150 -25.40759 325.4075886
CNRTa rget to Ra da r 12242488.907 70.879 dB 120
Altitude (m)
CNRJa mmer to Ra da r 0.00 - dB
100
J/SJa mmer to Ra da r 0 EAD Not Successful
CNRTa rget to Mi s s i l e 1353.840 31.316 dB 80
CNRJa mmer to Mi s s i l e 0.00 - dB
J/SJa mmer to Mi s s i l e 0 EAD Not Successful 60
Radar Pulses to Integrate (Detection of Signals in Noise eq. 10.18)
40
Pulse Repitition Frequency (f p ) 7200 pps 7200
20
Scan Rate (ωm) 18 rpm 18
Beam Azimuth (θb ) 1.50 deg 150 0
Beam Elevation (θ e ) 0.67 deg 67 0 5000 10000 15000 20000 25000 30000 35000
Hits per scan (ηb ) 100 Range (m)
Missile Pulses to Integrate (Detection of Signals in Noise eq. 10.18)
Pulse Repitition Frequency (f p ) 7200 pps 7200
Scan Rate (ωm) 18 rpm 18
Beam Azimuth (θb ) 1.50 deg 150
Beam Elevation (θ e ) 0.67 deg 67
Hits per scan (ηb ) 100
Calculate Bandwidth of Radar Matched Filter and P fa (Detection of Signals in Noise eq. 10.17)
Matched Filter Assumption (Bτ) 0.44 (See Table to Right)
Radar Receiver Bandwidth (BR) 2.20E+05 Hz 0.2200 MHz
Time Between False Alarms (Tfa ) 120 min
False Alarm Probability (Pfa ) 6.31E-10
Calculate Bandwidth of Missile Matched Filter and P fa (Detection of Signals in Noise eq. 10.17)
Matched Filter Assumption (Bτ) 0.44 (See Table to Right)
Missile Receiver Bandwidth (BR) 2.20E+05 Hz 0.2200 MHz
Time Between False Alarms (Tfa ) 120 min
False Alarm Probability (Pfa ) 6.31E-10
Radar Benefit from Integrating 100 pulses for Swerling Case I/II (Chart)
24 dB
Missile Benefit from Integrating 100 pulses for Swerling Case I/II (Chart)
0 dB No integration by missile
Radar CNR Needed to Detect Swerling Case I/II Target
22.982 dB Calculate CNR
Calculated CNR 46.879 dB
Total CNR 70.879 dB
Multipath a factor? Yes
Missile CNR Needed to Detect Swerling Case I/II Target
22.982 dB Calculate CNR
Calculated CNR 31.316 dB
Total CNR 31.316 dB
Multipath a factor? Yes
Figure 31. EAD Model
69
70
LIST OF REFERENCES
Blanchard, Benjamin S., and Wolter J. Fabrycky. 2006. Systems Engineering and
Analysis. 4th ed. Upper Saddle River, NJ: Prentice Hall.
Braud, Jeremy J. 2014. “Electronic Attack Platform Placement Optimization.” Master’s

thesis, Naval Postgraduate School. https://calhoun.nps.edu/handle/10945/43881
Brooker, Graham. 2007. Sensors and Signals. Sydney: University of Sydney.
Chuter, Andrew. 2016. “RAF Considers BriteCloud to Counter Radar-Guided Missile

Threat.” Defense News. 30 March. Accessed 15 June 2017.
http://www.defensenews.com/story/defense/international/europe/2016/03/30/raf-
radar-guided-missile-threat/82417308/
Cioppa, Thomas M. 2002. “Efficient Nearly Orthogonal and Space-Filling Experimental

Designs for High-Dimensional Complex Models.” PhD Diss. Naval Postgraduate
School. https://calhoun.nps.edu/handle/10945/9808
Defense Industry Daily. 2013. “Raytheon’s ALE-50 “Little Buddy” Decoys.” 25 April.
Accessed 15 June 2017. http://www.defenseindustrydaily.com/raytheons-ale50-
little-buddy-decoys-02573/
Eckstein, Megan. 2017. “NAVAIR Pushing Paperless Engineering to Speed Aircraft,

Weapon Design.” U.S. Naval Institute News. 24 October. Accessed 25 October
2017. https://news.usni.org/2017/10/24/navair-pushing-paperless-engineering-
speed-aircraft-weapon-design
Gabriel, Dominguez. 2016. “What Is China’s HQ-9 Air Defense System Capable of?”
Interview by author 17 February. Accessed 28 September 2017.
http://www.dw.com/en/what-is-chinas-hq-9-air-defense-system-capable-of/a-
19053690
Gianni, Daniele, Andrea D’Ambrogio, and Andreas Tolk. 2015. Modeling and
simulation-based systems engineering handbook. Boca Raton, LA: CRC Press.
Harney, Robert C. 2013a. Sensor Signals and Functions. Combat Systems Engineering.
Vol. 1. Monterey: Naval Postgraduate School.
———. 2013b. Sensor Technologies. Combat Systems Engineering. Vol. 2. Monterey:

Naval Postgraduate School.
———. 2013c. Command and Control Elements. Combat Systems Engineering. Vol. 5.
Monterey: Naval Postgraduate School.
71
———. 2017. Discussion on Radar Model Characteristics, edited by Devon A.
Cartwright.
Hayter, Anthony J. 2012. Probability and Statistics for Engineers and Scientists. 4th ed.
Boston: Brooks/Cole.
Hernandez, Alejandro S. 2008. “Breaking Barriers to Design Dimensions in Nearly

Orthogonal Latin Hypercubes.” PhD Diss. Naval Postgraduate School.
https://calhoun.nps.edu/handle/10945/10306
Kopp, Carlo, Andrew Martin. 2012. “Type 345 HHQ-7 Naval Sino-Crotale engagement
radar.” Digital Image. Air Power Australia. http://www.ausairpower.net/APA-
HQ-7-Crotale.html#mozTocId698527
Koppenberger, Jim. 2017. “Navy EAD Plans and Programs (PMA-272).” Presentation at
RF Summit #2, New Mexico State University, Las Cruces. 8 March.
Knott, E.F., D. J. Lewinski and S.D. Hunt. 1981. “Chaff Theoretical/Analytical

Characterization and Validation Program.” Georgia Institute of Technology,
Atlanta. 30 September. http://www.dtic.mil/dtic/tr/fulltext/u2/a105893.pdf
Kuzmin, Vitaly V. 2009. “Передатчик помех одноразового использования

(Disposable jamming transmitter)” Accessed 25 March 2017.
http://www.vitalykuzmin.net/Military/MAKS-2009/i-r9sbL7z/A
Mizokami, Kyle. 2016. “China Claims its New Anti-stealth Radars can Detect the F-22.”
Popular Mechanics. 11 November. Accessed 10 September 2017.
http://www.popularmechanics.com/military/weapons/a23846/china-anti-stealth-
radars-detect-f-22/
Onnis, Gian Luca. 2012. “Towed Decoy Behind Typhoon.” Digital Image. The
Aviationist: David Cenciotti’s Weblog. 16 April. Accessed 9 August 2017.
https://theaviationist.com/2012/04/16/towed-decoy/
Payne, Craig M. 2010. Principles of Naval Weapon Systems. Second ed. Annapolis,
Maryland: Naval Institute Press.
Rausch, Robert, David Benevelli, and Mort Serell. 2007. Testing Programs for
Transportation Management Systems: A Technical Handbook.
https://ops.fhwa.dot.gov/publications/tptms/handbook/index.htm
Streetly, Martin. 2005. Jane's Radar and Electronic Warfare Systems 2005-2006. Jane's
Information Group Limited. pp. 160-161
Wikipedia, The Free Encyclopedia. 2008. "USNChaff." Digital Image. Accessed 15 May
2017, https://commons.wikimedia.org/wiki/File:Usnchaff.jpg
72
Yue, Tao. 2001. “Detection of the B-2 Stealth Bomber and a Brief History on ‘Stealth.’”
The Tech. 30 November. Accessed 12 August 2017.
http://tech.mit.edu/V121/N63/Stealth.63f.html
73
74
INITIAL DISTRIBUTION LIST
1. Defense Technical Information Center

Ft. Belvoir, Virginia
2. Dudley Knox Library

Naval Postgraduate School
Monterey, California
75

Decoy Simulation

Uploaded by

Copyright:

Available Formats

Decoy Simulation

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Decoy Simulation

Uploaded by

Copyright:

Available Formats

NAVAL

A LOW-FIDELITY SIMULATION TO EXAMINE THE

Thesis Advisor: Alejandro Hernandez

Approved for public release. Distribution is unlimited.

13. ABSTRACT (maximum 200 words)

14. SUBJECT TERMS 15. NUMBER OF

A LOW-FIDELITY SIMULATION TO EXAMINE THE DESIGN SPACE FOR

Submitted in partial fulfillment of the

MASTER OF SCIENCE IN SYSTEMS ENGINEERING

NAVAL POSTGRADUATE SCHOOL

Approved by: Alejandro Hernandez

CDR Brian Connett

II. LITERATURE REVIEW .....................................................................................9

III. STUDY METHODOLOGY AND MODELING EFFORT .............................17

IV. SIMULATOR OPERATION ..............................................................................41

V. SIMULATION EXPERIMENTS AND ANALYSIS ........................................47

VI. CONCLUSION AND RECOMMENDATIONS ...............................................59

APPENDIX A. EQUATION VARIABLE DEFINITIONS..........................................63

APPENDIX B. MODELS ................................................................................................65

LIST OF REFERENCES ................................................................................................71

INITIAL DISTRIBUTION LIST ...................................................................................75

Figure 3. Example of Radar Returns from an Expendable Active Decoy ..................7

Figure 5. Prototype RT-1489/ALE GEN-X Decoy and Canister. Source:

Figure 6. ESAMS Model Elements. Source: Defense Systems Information

Figure 7. Systems Engineering Process “V” Model. Adapted from Rausch et

Figure 9. Engagement Scenario Flow Diagram and OV-1 .......................................21

Figure 11. Basic 3D Model of MH-60S with Hellfire. Source: POFACETS

Figure 12. MH-60S Bistatic Radar Cross Section at 16 GHz, 0° Angle of

Figure 13. Additional Carrier-to-Noise Required for Fluctuating Targets.

Figure 14. Receiver Operating Characteristic Curves for Swerling II Statistics.

Figure 16. Visual Simulation of Engagement .............................................................36

Figure 17. Nearly Orthogonal Latin Hypercube Experimental Design.......................39

Figure 18. Inputs to the EAD Model ...........................................................................42

Figure 20. Design of Experiment Variable Levels ......................................................43

Figure 21. Data Output from Multiple Simulation Runs.............................................44

Figure 22. Example of Statistical Output on the “DOE” tab.......................................44

Figure 23. Inputs Requested by the Simulator for Multiple Runs...............................45

Figure 24. Effect of Decoy Power on the Probability of Success ...............................50

Figure 25. Effect of Decoy Gain on Probability of Success .......................................53

Figure 26. Regression Analysis Results ......................................................................57

Figure 27. Radar Model...............................................................................................65

Figure 28. Multipath Model ........................................................................................66

Figure 29. Swerling Case I/II Receiver Operating Characteristics .............................67

Figure 30. Sea Surface Clutter ....................................................................................68

Figure 31. EAD Model ................................................................................................69

Table 2. Efficiency of Non-matched Filters. Source: Brooker (2007, 290). .....................25

Table 4. Experimentation Factors ......................................................................................38

Table 5. Statistical Data of Design of Experiments 1 ........................................................49

Table 6. Noteworthy Results from Design of Experiments 2 ............................................51

Table 7. Noteworthy Results from Design of Experiments 3 ............................................52

Table 8. Noteworthy Results from Design of Experiments 4 ............................................53

Table 9. Noteworthy Results from Design of Experiments 5 ............................................54

ASPD Advanced Self Protection Decoys

Currently fielded expendable countermeasure systems cannot adequately protect

Developing an Expendable Active Decoy (EAD) can leverage evolving

To employ a modeling and simulation based systems engineering approach requires

The simulation allows a systems engineer to examine the key measure of

Sequential experimentation enables the researcher to focus on the most relevant

Braud, Jeremy J. 2014. “Electronic Attack Platform Placement Optimization.” Master’s