Draft Date: January 31, 2008

Analysis and Design of Launch Vehicle Flight

Control Systems
Bong Wie and Wei Du
Iowa State University, Ames, IA 50011-2271
and
Mark Whorton
NASA Marshall Space Flight Center, Huntsville, AL 35812
This paper describes the fundamental principles of launch vehicle flight control analysis and design. In
particular, the classical concept of drift-minimum and load-minimum control principles is re-examined
and its performance and stability robustness with respect to modeling uncertainties and a gimbal angle constraint is discussed. It is shown that an additional feedback of angle-of-attack or lateral acceleration can
significantly improve the overall performance and robustness, especially in the presence of unexpected large
wind disturbance. Non-minimum-phase structural filtering of unstably interacting bending modes of large
flexible launch vehicles is also shown to be effective and robust.

I. Introduction
Note to Session Organizer/Reviewers: This draft manuscript summarizes very preliminary results obtained during an early phase of a project for the launch vehicle flight control systems analysis and design as applied to Ares-I
Crew Launch Vehicle. During the next several months, a more detailed, rigorous study will be conducted in the areas
of drift-minimum vs load-minimum control, flexible-body stabilization and analysis, gain scheduling vs. adaptive
control, etc. A companion paper on dynamic modeling of large flexible launch vehicles is also being submitted to this
Space Exploration and Transportation GNC session.

II. Rigid-Body Control Analysis

Consider a simplified linear dynamical model of a launch vehicle [15], as illustrated in Fig. 2, as follows:
= M + M
F
N
T
Z =
+
m
m
m
Z
= + + w
V
F = To + T D

(1)
(2)
(3)
(4)

where is the pitch attitude, the angle of attack, Z the inertial Z-axis drift position of the center-of-mass, Z the
inertial drift velocity, m the vehicle mass, To the ungimbaled sustainer thrust, T the gimbaled thrust, N = N the
aerodynamic normal (lift) force acting on the center-of-pressure, D the aerodynamic axial (drag) force, F the total
Vance Coffman Endowed Chair Professor, Space Systems and Controls Laboratory, Department of Aerospace Engineering, 2271 Howe Hall,
Room 2355, (515) 294-3124, [email protected], Associate Fellow AIAA.
Ph.D. Student, Space Systems and Controls Laboratory, Department of Aerospace Engineering, Student Member AIAA.
Chief, Guidance, Navigation, and Mission Analysis Branch, [email protected], Associate Fellow AIAA.

1 of 11

Figure 1. Comparison of Space Shuttle, Ares I, Ares V, and Saturn V Launch Vehicles [1].

x-axis force, the gimbal deflection angle, V the vehicle velocity, w = Vw /V the wind-induced angle of attack, Vw
the wind disturbance velocity, and
M = xcp N /Iy
M = xcg T /Iy
1
N = V 2 SCN
2

(5)
(6)
(7)

where Iy is the pitch moment of inertia. For an effective thrust vector control of a launch vehicle, we need
M max > M max
where max is the gimbal angle constraint and max is the maximum wind-induced angle of attack.
The open-loop transfer functions from the control input (s) can then be obtained as

(s)
s
N
M T
=
M s +
+
(s)
(s)
mV
mV

M (F + N )
Z(s)
1
T 2
=
s M
(s)
(s) m
m

(s)
s
T 2
M F
=
s M s +
(s)
(s) mV
mV
where

N 2
M F
(s) = s s3 +
s M s +
mV
mV

(8)

(9)
(10)
(11)

(12)

Consequently, the 4th-order system described by Eq. (1) - (3) is completely controllable by and is observable by Z;
however, the system is not observable by and .
In 1959, Hoelkner introduced the drift-minimum and load-minimum control concepts as applied to the launch
vehicle flight control system [6]. The concepts have been further investigated in [7-14]. Basically, Hoelkners controller utilizes a full-state feedback control of the form
= K1 K2 K3
2 of 11

(13)

Inertial Reference
Vw Wind Disturbance
X
. V
Z
#w
V
!

cp

Effective!
Wind !
Velocity

N = N# #

z
D

x cp

Z
cg

..

zm

xa

Accelerometer

x cg

To
"

Figure 2. Simplified pitch-axis model of a launch vehicle.

for a 3rd-order dynamical model of the form

0
1
0

0
0
d

=
0
0
M
+
M
+ 0
dt

F/(mV ) 1 N /(mV )

T /(mV )
w

(14)

This 3rd-order system is observable by or . The feedback gains are to be properly selected to minimize the lateral
drift velocity Z = V ( w ) or the bending moment caused by the angle of attack. Note that
Z
= w
V

(15)

where is often called the flight-path angle.

Instead of measuring the angle-of-attack, we may employ a body-mounted accelerometer, as illustrated in Fig. 2,
as follows:
= K1 K2 + Ka zm

N
T
xa

= K1 K2 + Ka
+ +

m
m
m
xa
N
T
= K1 K2 + Ka Ka
+ Ka
m
m
m
Because the resulting effect of zm feedback is basically the same as the feedback, we consider here only the control
logic described by Eq. (13).
Substituting Eq. (13) into Eq. (1) - (2) or Eq. (14), we obtain the closed-loop transfer function from the wind

disturbance w (s) to the drift velocity Z(s)

as
Z
A2 s2 + A1 s + Ao
= 3
w V
s + B2 s2 + B1 s + Bo
3 of 11

(16)

where

N
K3 +
T

K2 T
M N
= M (K1 + K3 ) M +
M +
mV
T

T K1
M N
F
=
M +

(M K3 M )
mV
T
mV

T
N
=
K3 +
mV
T

K2 T
M N
=
M +
mV
T
= Bo

T
B2 = M K2 +
mV
B1
Bo
A2
A1
Ao

For a unit-step wind disturbance of w (s) = 1/s, the steady-state value of Z can be found as
Z ss
(A2 s2 + A1 s + Ao )
Ao
= lim 3
=
= 1
s0 s + B2 s2 + B1 s + Bo
V
Bo

(17)

The launch vehicle drifts along the wind direction with Z ss = Vw and also with = = = = 0 as t 1.
It is interesting to notice that the steady-state drift velocity (or the flight path angle) is independent of feedback gains
provided an asymptotically stable closed-loop system with Bo 6= 0.
If we choose the control gains such that Bo = 0 (i.e., one of the closed-loop system roots is placed at s = 0), the
steady-state value of Z becomes
Z ss
(A2 s + A1 )
A1
1
= lim 2
=
=
s0 s + B2 s + B1
V
B1
1+C
where
C=

mV [M (K1 + K3 ) M ]
M K2 T + M N /T

(18)

(19)

For a stable closed-loop system with M (K1 + K3 ) M > 0, we have C > 1 and
|Z ss | < Vw
when Bo = 0. The drift-minimum condition, Bo = 0, can be rewritten as

M K3 M
N
xcp
=
1+
M K1
F
xcg

(20)

(21)

Consider the following closed-loop transfer functions:

s(s2 + M K2 s + M K1 )
= 3
w
s + B2 s2 + B1 s + Bo

(22)

s(K3 s2 + M K2 s + M K1 )
=
w
s3 + B2 s2 + B1 s + Bo

(23)

For a unit-step wind disturbance of w (s) = 1/s, we have = = 0 as t 1. However, for a unit-ramp wind
disturbance of w (s) = 1/s2 , we have
lim (t) = M K1

t1

lim (t) = M K1

t1

Consequently, the bending moment induced by and can be minimized by choosing K1 = 0, which is the loadminimum condition introduced by Hoelkner [6]. The closed-loop system with K1 = 0 is unstable because
Bo =

F
(M K3 M ) < 0
mV
4 of 11

(24)

However, the load-minimum control for short durations has been known to be acceptable provided a deviation from
the nominal flight trajectory is permissible.
A set of full-state feedback control gains, (K1 , K2 , K3 ), can be found by using a pole-placement approach or the
linear-quadratic-regulator (LQR) control method [21-22], as follows:
Z 1
(25)
min
(xT Qx + 2 )dt

subject to x = Ax + B and = Kx where x = [ ]T and K = [K1 K2 K3 ].

III. Rigid-Body Control Example

Consider a launch vehicle control design example discussed by Greensite in [15]. Its basic parameters are given as
in [15]

Iy = 2.43E6 slug-ft2 ,
F = 375, 000 lb,
V = 1320 ft/sec,
N = 240, 000 lb/rad,

m = 5830 slug,
xcp = 38 ft,
Vw = 132 ft/sec,
M = 3.75 s2 ,

T = 341, 000 lb
xcg = 32.3 ft
w = 5.73 deg
M = 4.54 s2

(26)

The open-loop poles of this example vehicle are: -1.9767, 0.0488, 1.8967
Note that the wind-induced angle of attack of 5.73 deg considered for this example in [15] is somewhat unrealistic
because it will require a maximum gimbal deflection angle of
max >

M
w = 4.73 deg
M

Most practical thrust vector control systems have a maximum gimbal angle constraint of about 5 deg. In this paper,
we also assume a second-order gimbal actuator dynamics of the form
(s) =

s2

n2
c (s)
+ 2n s + n2

(27)

where = 1 and n = 50 rad/s.

Table 1. Summary of rigid-body control analysis and design

Case No.
1
2
3
4
5
6

Controller Type

(, )-Feedback Control [15]

Drift-Minimum Control [15]
Load-Minimum Control [15]
LQR Control (Q = 0)
Drift-Minimum Control
Load-Minimum Control

Feedback Gains (K1 , K2 , K3 )

(2, 0.8, 0)
(2, 0.8, 3.614)
(0, 0.8, 3.614)
(0.6852, 0.8491, 0.9542)
(0.3220, 0.8352, 1.2765)
(0, 0.8352, 1.2765)

Closed-Loop Poles
-1.74881.3934j, -0.1596
-1.90874.2774j, 0.0
-1.93233.0533j, 0.0471
-1.9767, -1.8967, -0.0488
-1.9767, -1.8967, 0.0
-3.1323, -0.7816, 0.0405

IV. Flexible-Body Control Analysis

More detailed control and stability analysis results for Figs. 10 and 11 will be included in the final manuscript.

5 of 11

! (deg)

5
0
0
0

Zdot (ft/s)
Z (ft)

2
3
Time (seconds)

!5
!10
0
4

# (deg)

" (deg)

10

2
0
0
0

!50
!100
0
0
!100
!200
0

Zdot (ft/s)
Z (ft)

10
5
0
0
20

2
3
Time (seconds)

0
!20
0
5

# (deg)

" (deg)

! (deg)

Figure 3. (, )-feedback
control (Case 1).

0
!5
0
0
!5

!10
0
0
!10
!20
0

Figure 4. Drift-minimum control (Case 2).

6 of 11

! (deg)
" (deg)
# (deg)

0
!10
0
20

2
3
Time (seconds)

0
!20
0
10
0
!10
0
50
0
!50
0
100

Z (ft)

Zdot (ft/s)

10

0
!100
0

Zdot (ft/s)
Z (ft)

6
4
2
0
0

2
3
Time (seconds)

!5
!10
0
1

# (deg)

" (deg)

! (deg)

Figure 5. Load-minimum control (Case 3).

0
!1
0
0

!20
!40
0
0
!50

!100
0

Figure 6. LQR control (Case 4).

7 of 11

! (deg)
" (deg)

4
2
0
0

2
3
Time (seconds)

!5
!10
0
5

# (deg)
Zdot (ft/s)
Z (ft)

0
!5
0
0
!5

!10
0
0
!20
!40
0

Z (ft)

Zdot (ft/s)

# (deg)

" (deg)

! (deg)

Figure 7. Drift-minimum control (Case 5).

10
5
0
0
0

2
3
Time (seconds)

!5
!10
0
10
0
!10
0
10
0
!10
0
0
!20
!40
0

Figure 8. Load-minimum control (Case 6).

8 of 11

! (deg)
" (deg)

4
2
0
0

2
3
Time (seconds)

!5
!10
0
5

# (deg)
Zdot (ft/s)
Z (ft)

0
!5
0
0

!20
!40
0
0
!50

!100
0

Figure 9. Case 5 (drift-minimum control) with max = 5 deg.

Figure 10. Illustrations of dominant bending modes and sensor locations (Ref. 2).

9 of 11

Figure 11. Nichols plot for a baseline pitch-axis flight control system (Ref. 2).

10 of 11

V. Conclusions

References
[1] Cook, S., Ares Project Status, Presented at 2nd AIAA Space Exploration Conference, December 4-6, 2006.
[2] Whorton, M., Hall, C., and Cook, S., Ascent Flight Control and Structural Interaction for the Ares-I Crew
Launch Vehicle, AIAA 2007-1780, April 2007.
[3] Jang, J.-W., Bedrossian, N., Hall, R., Norris, H., Hall, C., and Jackson, M., Initial Ares-I Bending Filter
Design, AAS 07-059, February 2007.
[4] Betts, K. M., Rutherford, R. C., McDuffie, J., Johnson, M. D., Jackson, M., and Hall, C., Time Domain
Simulation of the NASA Crew Launch Vehicle, AIAA 2007-6621, August 2007.
[5] Betts, K. M., Rutherford, R. C., McDuffie, J., Johnson, M. D., Jackson, M., and Hall, C., Stability Analysis of
the NASA ARES I Crew Launch Vehicle Control System, AIAA 2007-6776, August 2007.
[6] Hoelkner, R. F., The Principle of of Artificial Stabilization of Aerodynamically Unstable Missiles, ABMA
DA-TR-64-59, September 25, 1959.
[7] Hoelkner, R. F., Theory of Artificial Stabilization of Missiles and Space Vehicles with Exposition of Four
Control Principles, NASA TN D-555, June 1961.
[8] Harris, R. J., Trajectory Simulation Applicable to Stability and Control Studies of Large Multi-Engine Vehicles, NASA TN D-1838, August 1963.
[9] Harvey, C. A., An Alternate Derivation and Interpretation of the Drift-Minimum Principle, NASA Contract
NASw-563, MH MPG Report 1541-TR 15, Minneapolis-Honeywell, November 22, 1963.
[10] Garner, D., Control Theory Handbook, NASA TM X-53036, April 22, 1964.
[11] Rheinfurth, M. H., The Alleviation of Aerodynamic Loads on Rigid Space Vehicles, NASA TM X-53397,
February 21, 1966.
[12] Martin, D. T., Sievers , R. F., OBrien, R. M., and Rice, A. F., Saturn V Guidance, Navigation, and Targeting,
J. Spacecraft, Vol. 4, No. 7, 1967, pp. 891-898.
[13] Frosch, J. A. and Valley, D. P., Saturn AS-501/S-IC Flight Control System Design, J. Spacecraft, Vol. 4,
No. 8, 1967, pp. 1003-1009.
[14] Haeussermann, W., Description and Performance of the Saturn Launch Vehicles Navigation, Guidance, and
Control System, NASA TN D-5869, July 1970.
[15] Greensite, A. L., Analysis and Design of Space Vehicle Flight Control Systems, Spartan Books, New York,
1970.
[16] Blackburn, T. R. and Vaughan, D. R., Application of Linear Optimal Control and Filtering Theory to the
Saturn V Launch Vehicle, IEEE Transactions on Automatic Control, Vol. AC-16, No. 6, Dec. 1971, pp. 799-806.
[17] Wie, B., Thrust Vector Control Design for a Liquid Upper Stage Spacecraft, Journal of Guidance, Control,
and Dynamics, Vol. 8, No. 5, 1985, pp. 566-572.
[18] Wie, B. and Byun, K. W., New Generalized Structural Filtering Concept for Active Vibration Control Synthesis, Journal of Guidance, Control, and Dynamics, Vol. 12, No. 2, 1989, pp. 147-154.
[19] Byun, K. W, Wie, B. and Sunkel, J., Robust Non-Minimum-Phase Compensation for a Class of Uncertain
Dynamical Systems, Journal of Guidance, Control, and Dynamics, Vol. 14, No. 6, 1991, pp. 1191-1199.
[20] Wie, B., Liu, Q. and Bauer, F., Classical and Robust H1 Control Redesign for the Hubble Space Telescope,
Journal of Guidance, Control, and Dynamics, Vol. 16, No. 6, 1993, pp. 1069-1077.
[21] Bryson, A. E., Control of Spacecraft and Aircraft, Princeton University Press, 1994.
[22] Bryson, A. E., Applied Linear Optimal Control, Cambridge University Press, 2002.
[23] Wie, B., Space Vehicle Dynamics and Control, AIAA Education Series, 1998.

11 of 11