Proceedings of the EuroGNC 2013, 2nd CEAS Specialist Conference on ThBT1.

1
Guidance, Navigation & Control, Delft University of Technology, Delft,
The Netherlands, April 10-12, 2013

Flight Control Algorithms for a Vertical Launch

Air Defense Missile

Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

Abstract The necessity of high maneuverability and vertical launching require

thrust vector control additional to aerodynamic control. That hybrid usage of aero-
dynamic and thrust vectoring controls effectively increases the agility of the missile
against air defense threats. This requirement and the rapidly changing dynamics
of this type of missiles renders the guidance and control design critical. However,
the findings suggest that classical guidance and control design approaches are still
valuable to apply and can have successful performance within the effective flight
envelope. It is very rare that a study concerns from detailed dynamics and analysis
of the dynamics covering flight mission and algorithms. In this study, together with
the modeling of the agile dynamics of a vertical launch surface to air missile and
the corresponding thrust forces and moments depending on linear supersonic the-
ory, the application of the flight control algorithms are presented. Two classic linear
autopilot structures are studied. During autopilot design process, an additional term
related to short period dynamics of boost phase is proposed and the drastic effect
of this term is shown. In addition to control algorithms, guidance algorithms are
also defined to fulfill the mission of the missile. Body pursuit algorithm is applied
for rapid turnover maneuver and midcourse guidance. Proportional navigation guid-
ance is chosen for terminal phase. In addition, an alternative maneuvering technique
is proposed to reduce further side slip angle during vertical flight.

Raziye Tekin
Institute of Robotics and Mechatronics, German Aerospace Center - DLR, Muenchner Str. 20,
82234 Wessling, Germany e-mail: [Link]@[Link],
Ozgur Atesoglu
Mechanical Engineering, Middle East Technical University, 06531, Ankara, Turkey
Kemal Leblebicioglu
Electrical and Electronics Engineering, Middle East Technical University, 06531, Ankara, Turkey

777
 ThBT1.1

2 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

1 Introduction

This paper presents the practical application of guidance and control methods for
a vertical launch surface to air missile (VLSAM). Through the paper, the challeng-
ing dynamics of the mentioned missile is briefly presented with the modeling of the
thrust vector control forces and moments, that are formed by the jet vane deflec-
tions, originating from the linear supersonic theory. There are many advanced con-
trol methods for such a rapidly varying dynamics. However, the authors of this pa-
pers advocate from a practical point of view that developing advanced control tech-
niques should be an option only after classical control techniques have been proven
to be inadequate. Hence, classical control techniques and ad-hoc gain scheduling
is applied for the VLSAM. The autopilot design is pursued separately for the mid-
course and terminal guidance phases of the flight. Angle autopilots are designed for
the mid-course, including the rapid turnover maneuver and acceleration autopilot
for terminal guidance phase respectively. The gains of the autopilot are scheduled
with respect to time during boost phase and Mach for the post boost phase. The
performance of the autopilots are analyzed within nonlinear simulation.
The effect of the axial acceleration during the boost phase is emphasized. Recent
studies covering boost phase do not present such a term. The effect is illustrated
within the simulation results regarding the comparison of schedules linearized sys-
tem with the nonlinear system. The second important issue is to define a constant
hybrid control ratio that interconnects the thrust vector with aerodynamic control; a
singular value analysis of the linear control influence is conducted for that reason.
In addition, autopilots are integrated with body pursuit and proportional naviga-
tion guidance (PNG) guidance schemes. The overall guidance and control design
is tested for a defense maneuver to defeat an approaching target. In literature, a
study regarding the initial roll maneuver for interceptors based on fuzzy guidance
has been found [1]. Here an alternative maneuver based on a basic approach: initial
roll command generation to minimize side slip angle during vertical flight is studied
which is very practical and efficient to implement.

2 Modeling of the Vertical Launch Missile Dynamics

Dynamic modeling of the VLSAM is carried out by implementing the well known
Newton-Euler equations with rigid body assumption. The VLSAM, analyzed in this
paper, is axi-symmetric and has a blunt nose. It is a tail controlled missile and uses
both the aerodynamic tail fins and jet vanes. Two main coordinate systems as the
body coordinate system (B) and the earth fixed inertial coordinate system (E) are
defined and the equations of motion are derived with respect to them. The origin of
the body axis system is assumed to be at the final center of gravity location after
burnout. Also, since the propellant of the missile is burning throughout the flight,
the mass, inertia and the position of center of gravity are formulized as a function
of thrust and total impulse values and included in the model. Hence, since the thrust

778
 ThBT1.1

Flight Control Algorithms for a Vertical Launch Air Defense Missile 3

and impulse are modeled as a function of time, the mass, inertia and center of gravity
position of the missile change as a function of time. Hence, the translational and
rotational motion of the missile can be written as Eq. 1. Detailed information about
the dynamics and aerodynamics can be found in [2].
   
u̇ rv − qw + (FAx + FGx + FT x )/m
 v̇   pw − ru + (FAy + FGy + FTy )/m 
   
ẇ  qu − pv + FAz + (FGz + FT z )/m 
 =  (1)
 ṗ  
   (MAx + MT x )/Ixx 

 q̇   (MAy + MTy − pr(Ixx − Iyy ))/Iyy 
ṙ (MAz + MT z − pq(Ixx − Iyy ))/Iyy

(B) (B) (B)

Here, F̄A(x,y,z) , F̄T (x,y,z) and F̄G(x,y,z) are the Cartesian components of the aerody-
(B) (B)
namic, thrust and gravity forces. M̄A(x,y,z) , M̄T (x,y,z) and are the Cartesian compo-
nents of the aerodynamic and thrust moments. u, v, w are the components of missile
velocity in body coordinates, p, q, r are the missile body rates, m is mass. As, the
VLSAM in this study is axi-symmetric and has cruciform geometry, Iyy = Iyy .
As for the calculation of the rates of the Euler angles, to avoid the singularity
when the pitch angle is equal to ∓90◦ , direction cosine matrix (DCM) formulation
is rather preferred to the Euler angle formulation to avoid the singularity problem.
Although the quaternion formulation is computationally more efficient, it is not cho-
sen because the DCM is more practical to apply and interpret physically.

2.1 Aerodynamic Forces and Moments

The aerodynamic forces/moments are functions of dynamic pressure (Qd ), missile

reference area (S ) and the aerodynamic force coefficients, i.e. Ci ’s. Hence the aero-
dynamic force vector in matrix representation can be written:
(B)
   
= FAx FAy FAz = Qd S Cx Cy Cz
F̄A
 
0 (2)
(B)  
M̄A = Qd S Cl Cm Cn + (xcgre f − xc (t))FAz 
(xcgre f − xc (t))FAy

The aerodynamic force/moment coefficients as nonlinear functions of flight vari-

ables; Ma, α, β , δa , δe , δr where Ma is the Mach number, α is the angle of attack
and β is the angle of sideslip, δa , δr , δr are the aileron, elevator, rudder deflections
of aerodynamic control surfaces. xcre f is the final, i.e. after the propellant burn-out,
and xc (t) is the instantaneous position of the center of gravity, Qd is the dynamic
pressure, S is reference surface. the VLSAM has high angle of attack flight regime,
so that an aerodynamic database is created including ±90 degrees of angle of attack

779
 ThBT1.1

4 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

and cross coupling terms. The details of the aerodynamic modeling and analysis of
the VLSAM can be found in [2].

2.2 Thrust Forces and Moments

The thrust forces and moments, are generated by deflecting the thrust vector by the
jet vanes located at the nozzle exit of the missile. The magnitude of thrust force (T ),
duration of the boost phase, the time to reach the maximum thrust level and the ge-
ometrical properties of the jet vanes are the critical parameters for thrust vectoring.
These parameters are adapted to achieve desired maneuvering capabilities for the
VLSAM [3]. The maximum thrust vectoring forces and moments are dependent on
the maximum jet vane deflection angle and vane characteristics such as thickness
(tk ), chord (c) and thrust motor characteristics as nozzle exit pressure (P) and flow
velocity (M∞ ). To determine the lift and drag forces, linear supersonic theory is di-
rectly applied. The area of the jet vane is not changing or may be neglected. There
are studies that include the jet vane erosion phenomenon which effects the lift and
drag forces created by vane deflections, [4], [5], [6] and [7].

PCL γM∞2 S jv PCD γM∞2 S jv

L= , D= (3)
2 2
Here, lift and drag force coefficients (CL and CD ) are functions of deflection of jet
vanes (δJV ), nozzle exit pressure, thickness and chord of the jet vanes:

4δJV 4 2 tk
CL = p , CD = p (δJV + ( )2 ) (4)
2
M∞ − 1 2
M∞ − 1 c

Within the scope of this paper, the area of the jet vanes are assumed to be fixed.
Linear supersonic theory may sometimes overestimate the lift and drag forces, and
it has to verified with 3D computational fluid dynamics analysis and experiments.
However, it is known that the accuracy of linearized theory is high when jet vanes
are located at enough distance to with respect to each other and outside the nozzle
[8]. Also, the dynamics between the jet vane deflection and total thrust deflection
is taken as unity, because of the high the inflow rate [9]. The moments created by
the deflection of the jet vanes are calculated by using the forces and the moment
arms which are the distance between the nozzle exit diameter and jet vane center of
pressure and the distance between the jet vane center of gravity and missile center
of gravity.

780
 ThBT1.1

Flight Control Algorithms for a Vertical Launch Air Defense Missile 5

3 Autopilot Model

There are studies on different algorithms for the control techniques applied at high
angle of attack flight regimes. They can be mentioned as the robust control design
for IRIS-T [10], the sliding mode controller [11], the adaptive control [12] and some
nonlinear control strategies comparison with classical control [13]. As mentioned
earlier, the aim of this study is to investigate the applicability of the classical control
design techniques on that challenging VLSAM dynamics and identify the possible
advantages and disadvantages. Other control techniques may then come into picture
to defeat the discrepancies that the classical control design techniques cannot han-
dle. In what follows, it will be demonstrated that autopilots designed with classical
control techniques can meet the mission requirements of the VLSAM.

3.1 Linearized Systems

The state space matrix calculated with the Jacobian linearization of the nonlinear
missile dynamics, at different instants of the vertical flight, can be written in the
following form to express the control efficiency originating from the control surfaces
separately:
∆ x̄˙ = Â∆ x̄ + B̂A ūA + B̂ūT (5)
where ūA = [δAa δAe δAr ] and ūT = [δTa δTe δTr ], and B̂A , B̂T are the control matrices
for the aerodynamic and thrust vector control. Here, ∆ x̄ =[∆U α β p q r φ ∆ θ ] is
in general form. The eigenvalues and system matrices of vertical flight at different
velocities and altitudes before and after burnout are given in [14].
To apply linear control techniques, nonlinear missile dynamics is decoupled into
three simplified representations of the overall motion since there is no clear distic-
tion between lateral and longitudinal dynamics because of its axi-symmetry prop-
erty. The pitch plane state space equations can be formulated as:
       
α̇ Z − u̇/U 1 + Zq α Z ZTδe δAe
= α + A δe (6)
q̇ Mα Mq q MAδ e MTδ e δTe

Qd Sd Qd S Qd Sd 2 Qd Sd
where Mα = Iyy Cmα , Zα = mU Czα , Mq = 2UIyy Cmq , Zq = C ,
2U 2 m zq
MAδe =
Qd Sd
Iyy CmAδe , ZAδe = QmUdS
CzAδe , MTδe = TIzyylx , and ZTδe = mU
Tz
including aerodynamic
derivatives e.g. Cmα , Czα etc.

3.1.1 Enhanced Short Period Approximation

The addition of the term u̇/U as seen in Eq. 6 enhances short period approximation
[15]. Conventional short period approximation assumes that the directional velocity
component of the air vehicle is constant (u̇ = 0), however the missile under study,

781
 ThBT1.1

6 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

especially accelerates rapidly in the boost phase. Thus, the u̇/U becomes significant
especially at the beginning of the vertical climb.
In Fig. 1, the normal acceleration time histories for a given longitudinal control
input is presented to show the comparison between linearized and nonlinear models.
Here, the linearized systems are scheduled with respect to time in the boost phase.
As it is seen from the figure, the linear system behavior is drastically separated from
the nonlinear system behaviour especially where u̇/U is high, i.e. at the beginning
of the boost phase. Thus, as for the acceleration autopilot design, it is crucial to
add u̇/U compensation to the short period approximation when the missile velocity
is comparably low. Otherwise, the normal acceleration controller, designed without
u̇/U compensation, may show either inadequate performance or unstable behavior
[15].

20
without udot
with udot
10 nonlinear sim.

−10
an(m/s )
2

−20

−30

−40

−50
0 0.5 1 1.5 2 2.5 3 3.5
time (s)

Fig. 1 Effect of missile axial acceleration in linearized missile acceleration dynamics

3.2 Thrust Vector-Aerodynamic Control Effectiveness

Control surfaces of such missiles are generally actuated using the same servo ac-
tuator. The challenging design problem is to set the ”hybrid control ratio” which
is directly related with the desired control capability. This ratio has to be consid-
ered together with the mission requirements and the control effectiveness. A static
control effectiveness analysis is conducted for that purpose. As expected, at low ve-
locities aero control is less effective than TVC and becomes powerful as the speed
increases. However, it looses control efficiency at high angle of attack values and
also at relatively high altitudes. The control effectiveness analysis of the aerody-
namic and thrust vectoring controls for the VLSAM at different altitudes is shown
in Fig. 2. Here, the singular values of the B̂ matrices are computed. TVC efficiency
stays nearly constant. This is an expected result that only the total mass of the missile
is decreasing while thrust level is nearly constant and it does not effect the TVC ef-

782
 ThBT1.1

Flight Control Algorithms for a Vertical Launch Air Defense Missile 7

3
10
Altitude (m)
0

5000

TVC
2
10

1
10

Aero Control

0
10 0
10
Mach

Fig. 2 Maximum singular value of B matrix

ficiency much. Further, the aerodynamic control becomes more effective than TVC
after 0.7 or 0.9 Mach depending on the altitude. In order to have maximum maneu-
verability, the control allocation between the two control schemes is distrubed as 1,
but a detailed analysis and optimization study is done for this purpose [14].

3.3 Autopilot Simulations

In order to be used for the mid-course guidance, autopilot is designed to operate

on roll and pitch angles [16]. In that phase, autopilot gains are time scheduled. The
performance of autopilot for pitch/roll angle reference commands and deflection
histories of aileron and elevator deflections are presented in Fig. 3(a) and Fig. 3(b).

Demand 10
80 Response
0
δ (°)
θ (°)

60 −10

40 −20
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4

50 20
Demand
40
Response 10
30
δ (°)
φ (°)

0
a

20
10 −10

0 −20
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
time (s) time (s)

(a) Angle autopilot response histories (b) Elevator and aileron deflections

Fig. 3 Angle autopilot nonlinear simulations

783
 ThBT1.1

8 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

VT (Mach)
2
1
0
0 0.5 1 1.5 2 2.5 3 3.5 4

10
0

α (°)
−10
−20
−30
0 0.5 1 1.5 2 2.5 3 3.5 4
5
x 10
6

Q(Bar)
4
2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
time (s)

Fig. 4 Velocity, angle of attack and dynamic pressure histories

Time histories of critical parameters are also depicted in Fig. 4. Although large angle
difference (45◦ ) are demanded for both the roll and pitch attitudes, autopilot results
are quite successful considering high variation of parameters such as α (0 − 30◦ ),
speed (0-3 Ma) and drastic increase in dynamic pressure.
For the terminal guidance, acceleration autopilot is designed using the 3-loop
acceleration autopilot scheme [13]. Autopilot gains are scheduled with respect to
Mach. The performance of the autopilot to square command input of 15g is illus-
trated in Fig. 5(a) and the elevator deflection in Fig. 5(b). Since the speed is decreas-
ing and demand stays constant in magnitude (15g), control commands increase to
compensate this kinetic energy loss.

20 Demand 15
a
n
15
10

10
5
5
a (g)

δ (°)

0
n

0
e

−5
−5

−10
−10
−15

−20 −15
4 5 6 7 8 9 10 11 12 4 5 6 7 8 9 10 11 12
time (s) time (s)

(a) Acceleration autopilot response (b) Elevator deflection

Fig. 5 Tracking performance of acceleration autopilot

784
 ThBT1.1

Flight Control Algorithms for a Vertical Launch Air Defense Missile 9

4 Guidance Algorithms

One of the main advantages of vertical launching is the engagement capability to

the targets in all possible directions that brings the necessity to direct the missile to
the plane of motion of target as soon as possible. Here, body pursuit guidance is a
possible candidate, and simple to apply, to align the missile body axis to the line
of sight. This procedure tries to minimize the look angle and effectively increases
the target detection possibility. The aim of the body pursuit guidance algorithm is to
produce reference body angle commands to be processed in the previously designed
angle command controller. As for the terminal guidance after the boost phase, the
conventional PNG guidance methodology is chosen for its proven performance and
ease of application.
In this study, the design of the guidance algorithms are divided into two phases
as the mid-course and the terminal guidance. The mid-course guidance starts in the
launch phase and operates until the hand-over to the terminal guidance phase. The
switching condition from mid-course to terminal guidance generally depends on
the current state of the missile and the target, trajectory constraints and the target
detection sensor, i.e. seeker, properties. A simple switching condition is defined and
set to occur when the lock-on range is less than 5 km and the field of view is less than
3 degrees. Intercept condition is defined based on the achievable minimum value of
the closing distance as 1 m.

4.1 Guidance and Control Simulations

The proposed flight control algorithms are implemented in a defense scenario on a

closing target that has a velocity of 1 Mach and starts a pull-up maneuver with 7g
when the target to missile range becomes less than 3 km. The initial position of the
target with respect to the missile is pT = [10, 2, −2] km. In Fig. 6(a), the acceleration

40
look angle(°)

80

20 40
a (g)
y

0
0 0 2 4 6 8 10
20
−20 0
0 2 4 6 8 10
α (°)

−20
10 −40

0 0 2 4 6 8 10
5
a (g)

−10 0
z

β (°)

−5
−20 −10
−15
−30
0 2 4 6 8 10 0 2 4 6 8 10
time (s)

(a) Acceleration histories (b) Look angle, angle of attack and sideslip

Fig. 6 Engagement simulations

785
 ThBT1.1

10 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

time histories are presented. Fig. 6(b) illustrates time histories of the look angle,
angle of attack, sideslip angle of the missile. Look angle is decreasing from 80◦ to 0
during the mid-course guidance phase. Whenever the guidance algorithm switches
to PNG, look angle starts increasing again (the behavior of the look angle at the
end is not a numerical but dynamical trend). More maneuvers on different type of
targets are simulated, see [14]. In all of the target types, flight control algorithms are
succesfull. Besides, the flight envelope has to be clarified and the overall success of
the flight control algorithms has to be examined.

5 Turnover Strategy

In the previous section, the turnover strategy used in mid-course guidance was skid-
to-turn strategy in which missile does not roll and yaw/pitch channel commands
are applied together to maneuver the missile towards the desired direction. Here,
the proposed turn-over strategy is actually a mixed ascend that turnover maneuver
composed of an initial roll followed by skid-to-turn (Fig. 7). That kind of maneu-
vering is also used in the mid-course guidance of air to air missiles [17] and [18].
[1] discusses a turn over strategy implemented with back to turn and roll maneu-
vers, however a direct comparison of skid-to-turn and skid-to-turn with initial roll
maneuver, and their advantages and disadvantages are not explicitly conducted.

Fig. 7 Turnover with initial roll maneuver

In this turn-over strategy, the missile has an initial roll maneuver and then starts
turning towards the target. The objective of the initial roll maneuver is to align the
pitch plane of motion of the body of the missile to the same vertical plane with the
target. Thus, after the initial roll maneuver that aligns the missile’s pitch plane of
motion, a maneuver in that single plane is required to head on towards the target.
This brings the advantage of flying with less control effort and reduces the side slip
angle values. In order to analyze the differences between standard turn-over and the

786
 ThBT1.1

Flight Control Algorithms for a Vertical Launch Air Defense Missile 11

turnover with initial roll maneuver, they are implemented for different simulations
for the same static target which is at pT = [1, 1, −0.2] km with respect to the
missile. This necessitates a roll angle command of 45 degrees. Upon executing the
roll maneuver, the engagement will become planar; so, only pitching control can
be used to capture the target. Fig. 8 shows angle of attack and side slip angle time

20

α (°)
−20
initial roll + skid to turn
−40 skid to turn

−60
0 0.5 1 1.5 2 2.5 3 3.5

20

0
β (°)

initial roll + skid to turn

−20
skid to turn

−40
0 0.5 1 1.5 2 2.5 3 3.5
time (s)

Fig. 8 Angle of attack and sideslip histories

histories compared for the standard skid-to-turn and skid-to-turn with initial roll
maneuvers. With the proposed maneuver, yaw maneuver is not required to head on
the target. However, as a draw back, it necessitates higher angle of attack than the
standard skid-to-turn maneuver as also seen in Fig. 8.

6 Discussion and Conclusion

The paper illustrates practical flight control algorithms for an advanced missile, the
VLSAM. First, six degrees of freedom flight dynamics is briefly introduced. An en-
hancing linearization term for the short period approximation which has not directly
mentioned in literature is proposed. Its drastic effect to similarity of scheduled lin-
earized systems and nonlinear dynamics is shown. Two different autopilots, angle
and acceleration, are designed using classical control techniques for the flight enve-
lope of the VLSAM in co-operation with guidance algorithms requirements. Aero-
dynamic and thrust vector control capabilities are blended in 1 : 1 ratio to have more
agility. The scheduled autopilots demonstrate satisfying performance in a highly
nonlinear, rapidly parameter and time varying environment which is a promising
start up for design process of industrial applications.
There are also advanced guidance techniques for agile missiles, but the analysis
and numerical results show that a body pursuit guidance for midcourse guidance
phase and proportional guidance for terminal guidance can be directly applied to
the VLSAM. Moreover, a turnover maneuver, which is the initial roll maneuver, is

787
 ThBT1.1

12 Raziye Tekin, Ozgur Atesoglu and Kemal Leblebicioglu

also accomplished for midcourse guidance phase that decreases maneuver require-
ments in lateral direction. To sum up, classical approaches are still applicable for
such an agile system. In order to start up a design, from industrial point of view,
the flight control algorithms which are applied in this paper can be implemented
easily and effectively. For further studies, maneuvers and autopilots are going to be
optimized in order to maximize total energy and increase flight time. Regarding the
optimization results, advanced flight control algorithms may be considered.

References

1. C. L. Lin, Y. P. Lin, and T.L. Wang. A fuzzy guidance law for vertical launh interceptors.
Journal of Guidance, Control, and Dynamics, 17(8):914–923, 2009.
2. R. Tekin, O. Atesoglu, and K. Leblebicioglu. Modeling and vertical launch analysis of an aero-
and thrust vectored surface to air missile. In AIAA Atmospheric Flight Mechanics Conference,
2010.
3. R. Tekin, O. Atesoglu, and K. Leblebicioglu. Vertical launch phase and effective turnover
maneuver analysis of an air defence missile. In 5. Defence Technologies, Ankara, Turkey,
2010.
4. A. I-Sharkawy. Experimental studies on thrust vector control in solid propellant rockets. In
13rd AIAA/SAE Propulsion Conference, 1977.
5. P.A Giragosian. Theoretical and experimental aerodynamic correlation of jet vane control
effectiveness. In AIAA Atmospheric Flight Mechanics Conference, 1981.
6. V. Harrisson, A. deChamplin, D. Kretschmer, R. Farinaccio, and R. A. Stowe. Force mea-
surement evaluating erosion effects on jet vanes for a thrust vector control system. In 39th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, 2003.
7. E. Orbekk. Thrust vector model and solid rocket motor firing validations. In 44th
AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, 2008.
8. E. Orbekk. Novel tvc system utilizing guide vanes with jet flap’s into a high efficiency compact
nozzle. In AIAA Atmospheric Flight Mechanics Conference, 2005.
9. G. Sung and Y. S. Hwang. Thrust-vector characteristics of jet vanes arranged in x-formation
within a shroud. Journal of Propulsion and Power, 20(3), 2004.
10. H. Buschek and M. McDaniel. Design and flight test of a robust autopilot for iris-t air to air
missile. Control Engineering Practice 11, pages 551–558, 2003.
11. A. Thukral and M. Innocenti. A sliding mode missile pitch autopilot synthesis for high angle
of attack maneuvering. IEEE Transactions on Control Systems Technology, 6(3):359–371,
1998.
12. C. Song. Adaptive Control of Linear Time Varying Systems with Applications to Missile Au-
topilots. PhD thesis, University of California, 1989.
13. E. Devaud, H. Siguerdidjane, and S. Font. Some control strategies for a high angle of attack
missile autopilot. Control Engineering Practice 8, pages 885–892, 2000.
14. Raziye Tekin. Design, modeling, guidance and control of a vertical launch surface to air
missile. Master’s thesis, Middle East Technical University, 2010.
15. K. S. Erer and R. Tekin. An investigation on the effects of forward speed change on lateral
missile dynamics. In National Automatic Control Conference, 2010.
16. R. Solis. An analysis of the vertical launch phase of a missile concept. In AIAA Aerospace
Sciences Meeting, 1983.
17. P. K. Menon and E. Ohlmeyer. Integrated design of an agile missile guidance and control
systems. In Proceedings of the 7th Mediterranean Conference on Control and Automation,
1999.
18. K. A. Wise and D. J. Broy. Agile missile dynamics and control. Journal of Guidance, Control,
and Dynamics, 21(3):441–449, 1998.

788