Академический Документы
Профессиональный Документы
Культура Документы
The Applied Technology Institute specializes in training programs for technical professionals. Our courses keep you
current in the state-of-the-art technology that is essential to keep your company on the cutting edge in today’s highly
competitive marketplace. Since 1984, ATI has earned the trust of training departments nationwide, and has presented
on-site training at the major Navy, Air Force and NASA centers, and for a large number of contractors. Our training
increases effectiveness and productivity. Learn from the proven best.
or Ft = I sp [ w 0 − w ( t )].
• Thus, burning an amount of fuel weighing ∆w over a period of time ∆t produces a total
impulse of F∆t = I sp ∆w
• Hence, Isp measures a fuel’s efficiency in converting weight into impulse, which can be written
I sp = − F = F∆t = U .
w& ∆w g
* i.e, the weight of the fuel decreases at the constant rate w& . © Walter R. Dyer, 2004
1-7
Example--Isp
• Isp is sometimes used incorrectly to compare the thrust-producing capability of rocket
fuels. This can lead to wrong conclusions because Isp is really measure of a fuel’s
efficiency in producing impulse.
• For example, the conventional fuels that propel missiles to supersonic speeds and
send satellites into space have Isp between 150 and 450 seconds, and some produce
over a million lb of thrust.
• On the other hand, NASA’s ion engine, which produces thrust by accelerating
ionized xenon atoms to 30 km/sec, has an Isp of 3(104)/g=3,100 sec. However, the ion
engine produces only 92 mN of thrust (about 1/3 oz)*.
Its mass flow rate is
kg
− m& = F = .092 = 3 (10 − 6 ) , or about 11 gm/hr
I sp g 3100 ⋅ 9 .8 sec
• The ion engine cannot produce enough thrust to lift
itself off the ground, but it can very efficiently
accelerate spacecraft to speeds of 5-10 km/sec in up to
a year of continuous operation, after the spacecraft is
first sent into space by a conventional rocket.
• The point to remember is that Isp= F∆t/∆w, so for a fixed
amount of fuel, ∆w, high Isp implies a large product of
thrust, F, and burning time, ∆t, not just high thrust. Ion Engine on NASA’s Deep
* Ion engines currently under development have demonstrated 1 N (0.225 lb) of thrust. Space 1 Satellite
© Walter R. Dyer, 2004
Missile Autopilots And Control
• After booster burnout, the missile enters the terminal phase (also called the homing
phase or end game), where is guided by changing the direction, not the magnitude,
of its velocity vector*. Endoatmospheric missiles do this by rotating aerodynamic
control surfaces (usually tails, wings or canards). Exoatmospheric missiles use
divert thrusters for homing.
• In the next section, we discuss how missiles accurately control lateral acceleration
in the homing phase so as to hit their targets or pass as close to them as possible.
* In endo missiles, there is some slowdown due to atmospheric drag, but every effort is made to minimize it.
© Walter R. Dyer, 2004
1-10
Introduction To Missile Autopilots
• The purpose of an autopilot is to produce lateral missile ac
acceleration a in response to commanded acceleration ac.
• Exoatmospheric missiles produce acceleration t
perpendicular to their center lines using divert thrusters, Acceleration Command
a
which act through the missile’s CG.
• Endoatmospheric missiles create acceleration t
perpendicular to their centerlines using the aerodynamic Ideal Zero Order Autopilot
force from the angle of attack α between the missile’s (No Delay, No Accl Limit)
α t
V Higher Order Response
Thrust ac a
Saturation
CP Auto-
t
t pilot
F
© Walter R. Dyer, 2004
1-11
Missile Control
• Missiles use tail, wing, canard, or thrusters to maneuver and control their attitude.
Some missiles use a combination of these control systems.
– Thrusters are used exclusively to control exo missiles, and sometimes in combination
with other types of controls in endo missiles.
– In tail- and canard-controlled missiles, the control surfaces are placed as far as possible
from the CG to produce the maximum pitch/yaw moment. In wing control, surfaces are
located at or near the CG.
– In endoatmospheric missiles, fixed surfaces are often added to improve stability and, in
tail- and canard-controlled missiles, to increase forces at the CP for high lateral
acceleration.
• Tail control
– Allows uniform component mass distribution (seeker and
warhead (if any) front, avionics and power mid, control
surfaces and actuators aft).
a
– Most stable in supersonic flight (minimal roll problems
from downwash from fixed surfaces) G
CP C
– Lowest drag on control surfaces (with positive static V
margin).
– Highest angle of attack without control surfaces stalling
(with positive static margin).
– Initial motion in wrong direction (corrected with forward
thrusters) and slightly reduced lift from tails. © Walter R. Dyer, 2004
1-12
Missile Control (continued)
• Canard control
– Produces largest moment for given control surface area. a
– Slightly more efficient than tail control because initial
CP CG
motion is in right direction (if static margin is positive). V
– Canards can stall at relatively low angle of attack.
– Downwash from canards complicates roll control.
– Nonuniform component mass distribution (seeker and control
system must be at front).
• Wing control
– Minimizes angle of attack, reducing body drag and relaxing a
seeker FOV requirement.
– Allows angle of attack to be controlled directly, eliminating
the need for accelerometers in simple missiles. CP=CG
V
– Wings must be large compared to tails or canards, increasing
drag, servo motor size, power consumption, missile mass,
and volume requirements.
– Requirement to pivot wing at or near CG makes the missile
sensitive to CG and CP shifts. a
• Exoatmospheric missiles use divert and attitude control
thrusters to accelerate missile normal its center line and V
maintain attitude directly. No aerodynamic problems in
space. © Walter R. Dyer, 2004
1-13
Exoatmospheric Missiles
• Exoatmospheric missiles obey the same six Euler y
equations as endoatmospheric missiles. Making the a V x
α
same assumptions as before, the motion decouples into
two orthogonal planes described by the same linear γ ψ σ
differential equations, mVγ& = Fy and Iψ&& = M z .
• Outside the atmosphere these equations can be
mechanized directly using divert and attitude control
thrusters to provide Fy and Mz. There is no need for aerodynamic coefficients.
• Exo missiles can fly at any angle of attack, α, without concern for aerodynamic forces, and an
exo seeker can point at a target continuously (σ ≈ψ). Just as in the endo case, missile
acceleration, Vγ& is referenced to V , so it is necessary to divide acceleration commands by cosγ.
• Radomes are unnecessary, eliminating the problems of radome error slope.
• The exoatmospheric missile’s airframe response does not change with flight conditions, so
open loop autopilots could be considered. However, an IMU is necessary for midcourse fly
out and position and rate feedback enhance thruster firing precision, improving both miss
distance and fuel economy. Hence gyro and accelerometer feedback are used in most designs.
• Although there is no atmospherically-induced roll, roll control is still desirable (though not
mandatory), because it prevents precession, reduces image smearing the on seeker’s focal
plane array, and decouples lateral missile motion from attitude.
• As divert and attitude control system (DACS) fuel burns, the missile’s CG shifts, causing the
divert thrusters to produce moment as well as the desired lateral ∆V. This is the largest source
of unwanted coupling in exo missiles, and the ACS must be sized to counter it. © Walter R. Dyer, 2004
Missile Guidance
VT
VM
M1 T1
Line of Sight (LOS)
T0
M0
Intercept: LOS Angle Remains Constant
(Constant Bearing Decreasing Range)
Miss Distance With
No missile .
MF Miss Distance With
No missile
. . VM
.
}
Acceleration Acceleration
TF TF
}
aM for
VM VT MF VT
intercept
aM for
intercept
M0 T0
M0 T0
Miss: LOS Angle Increasing With Time Miss: LOS Angle Decreasing With Time
© Walter R. Dyer, 2004
1-18
Zero-Effort Miss
• In deriving homing guidance laws, it is useful to know what the miss distance would
be if the missile and target did nothing but continue on their current courses (VM and
VT in the diagram below). This is called the zero-effort miss (ZEM). It is the miss
distance, calculated at any time, t, during the homing phase, assuming neither the
missile nor the target accelerates (in any direction) from time t until their closest
point of approach (CPA) at time tf.
• The logic behind ZEM is that if we know how far a missile will miss its target
without effort, an acceleration can be applied that will reduce that miss. A homing
guidance law can then be derived using ZEM by repeating the process: “Find ZEM,
compute missile acceleration to reduce ZEM”, …, throughout the homing phase.
• Previous charts showed that planar analyses can be used for y VT
VM
LOS motion of without loss of generality. The planar r T
diagram at the right shows a missile, M, and its target, T, σ y
Ref = Initial LOS
x
moving with velocities VM and VT respectively, with M
the x-axis along the initial LOS, as in the previous charts.
• Assume the range rate, r&, is consant. Then for any two times t and τ , the range is
r (τ ) = r (t ) + (τ − t )r&. If τ = t f = the time of closest point of approach (CPA), then
r (t f ) ≈ 0, and r (t ) ≅ −(t f − t )r&. Defining t go = (t f − t ), this can be written r ≅ − r&t go .
• Define the relative missile-target y-displacement as y=yT−yM, then the miss distance,
y(tf), assuming no accelerations take place during time interval tgo is the ZEM, Z.
Z = y + y&t go . © Walter R. Dyer, 2004
1-19
Zero-Effort Miss and LOS Turning Rate
• ZEM is closely related to the LOS turning rate, σ& . Ignorance of this relationship has
led some to assert they discovered new guidance laws using ZEM, which later turned
out to be simply existing laws with σ& converted to ZEM.
• The xy and rσ systems are rectangular and polar coordinate representations of the
same problem. We retain both because rectangular coordinates are usually easier to
work with, while missile seekers make measurements in polar coordinates.
• To derive the relationship between ZEM and σ& , recall y
VT
from the last chart that ZEM is given by Z = y + y& t go , so VM
from the diagram at the right, since σ is small, r (xT , yT )
σ y
x
y = r sin σ ≅ rσ . (xM , yM )
so Z = rt goσ& .
• Alternatively, Z = − r&t go
2
σ& .
© Walter R. Dyer, 2004
1-20
Proportional Navigation
• Russian mathematician Aleksandr Lyapunov (1857-1918) developed a stability theory
which follows from the fact that a quantity that is always positive and decreasing
approaches zero asymptotically. Lyapunov’s theorem states that for a system with
state vector x, if a scalar function V(x) can be found with the following two properties
for all nonzero x,
V( x ) > 0 and V& ( x ) < 0,
then the system is asymptotically stable (||x(t)||→ 0 asymptotically as t→∞).
• Lyapunov’s theorem enables the derivation of a very effective homing guidance law.
• Let the state variables be y and y& , and choose V ( y, y& ) = Z 2 = ( y + y& t go ) 2 . Then V > 0*,
and V& is V& = 2 ZZ& = 2 Z [ y& − y& + (t − t ) &y&] = 2 Zt &y&,
f go
where &y& = &y&T − &y&M = aT − a M is the relative missile - target lateral acceleration.
• If we assume zero effort, aT = a M = 0, V& = 0, and the missile probably misses the
target. To make || x || = y 2 + y& 2 → 0 requires a missile acceleration, a M , that satisfies
V& = 2 Zt go ( aT − a M ) < 0 with aT remaining zero. That is, V& = 2 Zt go ( − a M ) < 0. Using
Z = rt goσ& , from the previous chart, this becomes V& = −2 rt go
2
σ& a M < 0. Dividing by
− 2 rt go
2
, this simplifies to σ& a M > 0.
• A missile acceleration, a M , that satisfies this inequality throughout the homing phase will
reduce || x || = y 2 + y& 2 , and therefore Z , asymptotically to zero.
* V = Z 2 > 0 provided Z ≠ 0. If Z = 0 , there is nothing to solve. Intercept occurs with zero effort. © Walter R. Dyer, 2004
1-21
Proportional Navigation (continued)
• A sufficient condition for the inequality, σ& aM > 0, of the previous chart to hold is
a M = K σ& , with K > 0 .
• This simple guidance law is called proportional navigation (PN)*. It uses only one
variable, the LOS turning rate, σ& , yet is remarkably capable of intercepti ng a wide
variety of maneuvering and non - maneuvering targets. Most missile homing
guidance laws rely on σ& alone or in combinatio n with other variables to intercept
targets with great success * *.
• Dividing by VM , PN can also expressed as γ& M = Nσ& . This is the form used in the
autopilots section to analyze radome effects.
• The change of Z = y + y & t go is found using its derivative , Z& = y& − y& + &y&t go = &y&t go ,
where &y& = aT − a M . Assuming aT = 0 as before, this becomes Z& = − a M t go = − Kt goσ& .
& − Kt goσ&
Z
Hence, = 2 = K , or dZ = K dt , or, assuming constant K , ln Z = − K ln(t f − t ).
Z − t go r&σ& r&t go Z r& t f − t K K r&
− −
Z (t ) ⎛⎜ t f − t ⎞⎟ r& ⎛⎜ t ⎞⎟ r&
• Evaluating this from t=0 to t=tf , = = 1−
Z (0) ⎜⎝ t f ⎟⎠ ⎜ t ⎟
⎝ f ⎠
• Let Λ −
K be a positive, dimensionless constant, then the convergence of ZEM using PN
r& Λ
is controlled by Λ as Z (t ) ⎛⎜ t ⎞⎟
= 1−
Z (0) ⎜⎝ t f ⎟⎠
* Note that K need not be constant here, just positive.
* * In deriving PN it was assumed that r& was constant t o simplify t he math. However, this restrictio n is not necessary, as the
derivation of appendix E shows. In fact, PN works very well with vary ing r& , as long as r& < 0. © Walter R. Dyer, 2004
1-22
Proportional Navigation (continued)
• Using K = − Λ r& from the last chart in the PN law
a M = K σ& ,
a M = − Λr&σ& .
• This is the usual form of the PN law. The plots of
Λ
Z (t ) ⎛ ⎞
ρ (t ) = = ⎜⎜ 1 − t ⎟⎟ vs τ = t shown at the
Z ( 0) ⎝ t f ⎠ tf
right suggest use of large Λ for rapid convergence,
but missile acceleration must also be considered.
Acceleration is given below as a function of Z 0 .
Z Λ Z ΛZ 0 (1 − t / t f ) Λ
a M = − Λr&σ& = − Λr& = 2 =
− rt go t go
& 2
(t f − t ) 2
Λ −2
(t f − t ) Λ−2 ΛZ 0 ⎛ ⎞
t ⎟
= ΛZ 0 = ⎜ 1 −
t Λf t 2f ⎜⎝ t f ⎟⎠
• A plot of the normalized PN acceleration,
Λ- 2
a M t 2f ⎛ ⎞
= Λ ⎜⎜ 1 − t ⎟⎟ vs τ = t is shown at the right.
Z0 ⎝ tf ⎠ tf
• Both large and small Λ demand high missile
acceleration. To avoid exceeding acceleration
limits, intermediate values are needed. Therefore,
most missiles use
3 ≤ Λ ≤ 10 . © Walter R. Dyer, 2004
Missile Seekers
Semi-active
Seeker
Passive Seeker
Active
Command
Seeker
Guidance
}
– One Hz=one cycle per second, hence Hz has dimension sec-1. 1
– Radar waves travel at the speed of light, c=3(108) meters/sec, so frequency f and wavelength λ are
related by λf=c.
– The formulas above apply to all portions of the electromagnetic spectrum (IR, viz, etc), not just radar.
• Radars transmit electromagnetic waves and track the LOS angle σ via the energy reflected
back from targets using a parabolic dish antenna or a phased array. Their angular accuracy is
approximately their beam width θ=λ/D, where D is the diameter of the dish or array. σ& for
homing is then found by filtering (usually Kalman filtering).
• Radar can also measure a target’s range r by the time ∆t it takes a transmitted pulse to reach
the target and return. Since the round trip distance is 2r=c∆t, the range is r=c∆t/2.
– Recall that for proportional navigation, the missile commands an acceleration a c = − Λ r&σ& . The
range measurements can be used to estimate r& or it can be measured directly by doppler radar.
– Optimum homing laws use an estimate of time-to-go, tgo, which, as we saw, is closely approximated
by t go = − r / r&. Hence, both r measurement and r& estimates are needed for optimum homing.
© Walter R. Dyer, 2004
1-28
The Radar Equation
• Consider a radar transmitting power Ptr uniformly in all directions (isotropic radiation). A
sphere of radius r centered at the transmitter has area 4πr2, so the power density J0 through its
surface is J0=Ptr/(4πr2). With an antenna of gain* G, the power density is Jr=GJ0, hence,
with an antenna, the power density through the sphere (where the energy is focused) will be
Ptr G
. Jr =
4π r 2
• A target r meters away will act like a reflector of some area σ m2, called the radar cross
section** (RCS), echoing a fraction of the transmitted power back to the radar receiver. The
power density at the receiver will be σ Ptr G σ
J rec = J r =
4π r 2 ( 4π r 2 ) 2
• If the radar uses a receiving antenna of effective area A1, it will measure power
Prec=A1Jrec=A1PtrGσ/(4πr2)2. Like most radars, missile radar seekers are monostatic (i.e., they
use the same antenna to transmit and receive). It can be shown that the gain of an antenna
with effective area A1 is G=4πA1/λ 2 , so A1=Gλ2/(4π), and the received power is
G 2 λ2σ
Prec = P
( 4π ) r
3 4 tr
• This is the radar equation. Note that the power received by the missile’s seeker is inversely
proportional to the fourth power of the missile-target range r.
* Gain is achieved by focusing the radar energy through only a portion of the sphere. The ratio of a radar’s power with an antenna to its
isotropically radiated power is defined as the antenna’s gain.
** Typical RCS: Missiles 0.5-2 m2, fighter aircraft 2-6 m2, jumbo jet 100 m2, car 100 m2, man 1 m2, insect .00001 m2 © Walter R. Dyer, 2004
1-29
Radar Bands
• Radar wavelengths and frequencies are shown below. Most missile systems
operate in the C, X, or Ka band. Higher frequencies are preferred because they
allow better angular resolution and the components are smaller, however,
atmospheric attenuation restricts the use of some bands above X. Ka and W are
attractive because of reduced attenuation in portions of these bands.
• Improvements in inexpensive, light weight components is still needed above Ka
band.
1000
10
0.1
mm
UHF L S C X Ku K Ka W © Walter R. Dyer, 2004
1-30
Semi-active Missile Systems
• In the semi-active system shown, let the radar-target distance be r1 and the
missile-target distance r2, and call the ground radar illuminator and seeker
r2
antenna gains G1 and G2 respectively. G2
P G
• As in the active seeker case, the power density at the target is J r = tr 21 . r1
The power density reflected back the the missile seeker is
1
4 π r1
σ Ptr G 1σ
J rec = J r1 = G1
4π r22 ( 4π r12 )( 4π r22 )
• If the seeker uses a receiving antenna of effective area A, it will measure power
Prec=AJrec=APtrG2σ/(4πr12)(4πr22). From the formula for antenna gain, A=G2λ2/4π, the received power is
G 1G 2 λ 2σ
Prec = Ptr
( 4 π ) r1 r2
3 2 2
• Semi-active missile systems are attractive due to their long acquisition range and because a missile’s mass
and cost are lower without a radar transmitter; also a ground radar can be much larger and more powerful
than a missile-borne radar. For example, if a ground radar antenna’s diameter is 10 times that of a missile
seeker’s antenna, and transmitts100 times the power, the G1Ptr product is 10,000 times larger than the
GPtr product obtainable with an active seeker aboard the missile. Before the advent of semiconductor
electronics, most homing missiles were semi-active.
• But semi-active systems are often less effective than active systems. Three reasons for this are:
– As the missile closes on the target, the power reflected back to its seeker increases faster with an
active missile seeker than with the semiactive (r2−4, vs r2−2, assuming r1 is nearly constant).
– Multiple targets are difficult to handle with semiactive missile systems. Separate frequencies or
codes are needed to control each target. Semi-active ground systems become less effective with each
additional target they must track, until they finally become overwhelmed.
– The powerful illuminator is easy to detect and jam or target. © Walter R. Dyer, 2004
Passive Missile Seekers
fl r
• Images of a 1 m x 2 m conical target are shown below for a seeker with 1.4 degree
FOV, a 25 cm aperture, and a 256 x 256 FPA at λ=10 µm.
One Pixel
r > 20 km
2-3 Pixels
r = 10-20 km
8 Pixels
r = 5 km
© Walter R. Dyer, 2004
1-33
Radar vs IR Seeker Detectors
• An x-band radar seeker with λ=3 cm, a 25 cm aperture, and fl=15 cm produces a
resolution limited spot 2.44(λ/D)fl=4.4 cm in diameter, while a 10 µm LWIR seeker
with the same parameters produces a 15 µm spot.
• With such a small spot, the sensor for an IR seeker usually consists of an array
(called a focal plane array(FPA)) of many (up to several million) small detectors,
called pixels (picture elements), which can form a two-dimensional image of a target
if it is close enough.
• Radar seekers usually have only one detecting element, and do not image targets in
this way.
• A seeker’s field of view is the angular width it can see. The field of view of a radar
or passive seeker is FOV=1.22 λ/D. Note that a radar seeker’s FOV is the same as
its beam width.
• The FOV of a single pixel of a FPA is called the instantaneous field of view (IFOV).
The FOV of a passive seeker is its instantaneous field of view multiplied by the
number of pixels in a row or column of the FPA. Thus for a 265x256 FPA
operating at λ=9 µm with a 20 cm aperture, IFOV=1.22(9)(10-6)/0.2=5.5µr, and the
field of view is 256(5.5)(10-6)=14 mr=0.81°.
© Walter R. Dyer, 2004
1-34
Focal Plane Arrays And Read Out Electronics
• A focal plane array (FPA) is an array of 4 to 1,000,000 detectors, called pixels, which are sensitive to a
portion of the IR spectrum. Each pixel is 20-30 µm in size in missile seekers.
• FPAs consist of two parts: a pixel array (UV, Visible, or IR) and a read out electronics (ROE) chip,
bonded to the pixel array by indium bumps, which also provide electrical contact.
• The ROE converts the large number of individual pixel outputs into a pulse train, reducing the number of
output wires from tens of thousands to under 100.
• Frame rate is the number of times the entire array is read out each second. Frame rates of 10-100 Hz are
typical in missile seekers. The frame rate must be short enough to avoid smearing the target image and
long enough to allow adequate energy collection before readout. Readout A/D speed also limits the frame
rate (e.g, a 512 x 512 FPA operating at 50 Hz with 16 bit words must transfer over 200 Mb/sec)
• All the pixels of a seeker’s FPA observe the FOV for some fraction of each frame, called the integration
time. During the integration time, the photoelectrons are collected in small capacitors behind each pixel in
the ROE, called wells. The storage capacity of the wells limits the dynamic range of FPAs to about
10,000-100,000 and also limits the S/N.
• An array of microlenses can be bonded to the FPA to increase its sensitivity when the fill factor is small. It
also improves nuclear hardness because the microlenses do not refract high energy radiation. However,
microlenses increase cost.
Readout Integration time
Detector Integrated
Array Circuit
IR L
Frame time
t
Input/Output
Pads