CSC / SETI Institute Colloquium Series

SETI INSTITUTE, Mountain View, CA, USA

November 25th, 2009.

DEEP SPACE FLIGHT

& COMMUNICATIONS :
SETI, KLT & ASTRONAUTICS
IN A 2009 BOOK
Claudio Maccone
Member of the International Academy of Astronautics
Home Page: http://www.maccone.com/
1
E-mail: clmaccon@libero.it
Cost = $140.00 & Book Cover

2
Source, Sun and Spacecraft

3
Gravitational Lens of the Sun

Figure 1:
Basic geometry of the gravitational lens of the Sun:
the minimal focal length at 550 AU
and the FOCAL spacecraft position.
4
Gravitational Lens of the Sun

• The geometry of the Sun gravitational lens is easily

described: incoming electromagnetic waves (arriving, for
instance, from the center of the Galaxy) pass outside the Sun
and pass within a certain distance r of its center.
• Then a basic result following from General Relativity shows
that the corresponding deflection angle (r) at the distance r
from the Sun center is given by (Einstein, 1907):

4GM Sun
 r   2
.
c r

5
Gravitational Lens of the Sun

• Let’s set the following parameters for the Sun:

1. Assumed Mass of the Sun: 1.9889164628 . 1030 kg, that is
Sun = 132712439900 kg3s-2
2. Assumed Radius of the Sun: 696000 km
3. Sun Mean Density: 1408.316 kgm-3
4. Schwarzschild radius of the Sun: 2.953 km
One then finds the BASIC RESULT:
MINIMAL FOCAL DISTANCE OF THE SUN:
548.230 AU ~ 3.17 light days ~
13.86 times the Sun-to-Pluto distance.
6
Larger and larger Telescopes...
• Wherever in space there are intelligent creatures like us, they
will be driven to explore and understand our universe, just as we
do. We and they wish to see to the farthest depths of space with
the greatest clarity allowed by the laws of nature. To this end, we
build, at great expense, ever more powerful telescopes of all
kinds on Earth, and now in space.
• As each civilization becomes more knowledgeable, they will
recognize, as we now have recognized, that each civilization has
been given a single great gift: a lens of such power that no
reasonable technology could ever duplicate or surpass its power.
This lens is the civilization's star. In our case, our Sun.
• The gravity of each such star acts to bend space, and thus the
paths of any wave or particle, in the end creating an image just
7
as familiar lenses do.
Every Star IS a Telescope !!!

• This lens can produce images which would take perhaps

thousands of conventional telescopes to produce. It can
produce images of the finest detail of distant stars and
galaxies.
• Every civilization will discover this eventually, and surely
will make the exploitation of such a lens a very high priority
enterprise.
• One wonders how many such lenses are being used at this
moment in time to scan the universe, capturing a flood of
information about both the physical and biological realities
of our time.
• Frank Drake, 1999, from his Foreword to Claudio’s book.
8
This presentation is dedicated
to Professor Frank D. Drake
• This author came to learn about the Sun as a Gravitational
Lens for the first time at the Second Bioastronomy
Conference held at Balatonfuered, Hungary, in June 1987.
• Professor Frank D. Drake made the relevant presentation
there in a qualitative way i.e. without equations. This author
then decided to write down all equations and especially
study the space mission to 550 AU and beyond to 1000 AU.
• The result was the first author’s book in 1997 about the
FOCAL space mission to 550 AU. This book was awarded
the IAA Engineering Book Award in 1999.
• In 2000 a Proposal to ESA ensued, but was not funded.
• This presentation is dedicated to Prof. Frank D. Drake.
9
 Infinity of Focuses > 550 AU
• There is an infinity of focuses from 550 AU outward in any
direction. Thus, 550 AU is actually the minimal focal sphere.
• In the practice, we won’t have to stop a spacecraft just at 550
AU, but we just let it go.
• The further the spacecraft goes beyond 550 AU, the better it is.
In fact, radio waves impinging on the spacecraft at distances
higher than 550 AU will have to cross less and less dense layers
of the Solar Corona.
• The Solar Corona is difficult to model. Essentially, because of
the radialy decreasing electron density, the Solar Corona acts as
a divergent lens opposing the convergent lens of gravity.
• We shall study the Corona later. Just the Naked Sun for now.
10
 Gain of Any Star as a Lens
• The “gravity lens” concept means that the Sun (and any other
massive celestial body) is an antenna since it can increase the
intensity of the signal, by virtue of its deflection.
• We define the Gain associated to any star, Gstar, as the ratio
between the intensity of the signal in presence of the star
compared to the intensity of the signal without the star. It can
be proven that, along the focal axis, one has
rg 8  2 G M star 1
G star    4  2
 2

 c 

The gain is constant along the focal axis but is wavelength-

dependent. There, rg is the Schwarzschild radius of the star.
11
 Off Axis Gain for any Star

• Off axis, when the spacecraft is at a distance  from the focal

axis, and when it is at a distance z from the star, the gain can be
proved to be:
rg  2  2 rg 
Gstar  ,  , z   4  2
 J 02 
   z 
 
where J0(x) is the Bessel function of order zero and argument x
• The total gain for the combined (Star + receiving antenna)
system is:
GTotal ( )  Gstar ( )  Gantenna ( )

and so it increases with the CUBE of the frequency (next slide).

12
Total Gain of Star+Spacecraft

• If a spacecraft (S/C) has an antenna with radius rantenna and

efficiency kantenna, the spacecraft antenna gain is given by
A physical  k antenna 4  2 rantenna
2
 k antenna
Gantenna (  )  4  2

 2
and is proportional to the inverse of 2.
• The total gain for the combined (Star+S/C receiving antenna)
system is proportional to the inverse of 3 and is:
32  4 G M star rantenna
2
 k antenna
GTotal (  )  G star (  )  Gantenna (  ) 
3

13
Sun comes BEFORE interstellar!
• Sun’s Focus Comes FIRST.
• Interstellar Target Comes SECOND.
1) The Sun’s gravity focus is MUCH CLOSER than the target star,
actually hundreds or thousand of times closer according to the
target star (for  Cen it is 253 times closer). That is, 1000 AU,
where the “true” focus is found by taking the CORONA into
account, is 253 times closer than the target star system,  Cen.
2) BEFORE any interstellar probe is launched towards a nearby star,
we need a highly magnified radio-map of whatever lies around
that star. This can be achieved only by sending a probe to the
opposite direction to let the Sun magnify!
3) It is much CHEAPER to reach 550 AU or 1000 AU than hundreds
14
of AU, and it takes so much time less!
ASTRODYNAMICS: Max Speed

SOLAR SYSTEM EJECTION PROBLEM :

• Find the sequence of optimal FLYBYS within the
solar system such that…
• THE SUN FLYBY IS THE LAST ONE and…
• the Focal probe LEAVES the solar system at the
MAXIMUM POSSIBLE SPEED…
• and in the direction OPPOSITE to the target star.
• Conclusion: ONE TARGET – ONE MISSION !!!
• This problem was first considered by K. Ehricke in
1972. Today’s Astrodynamics go for better solutions!
15
 The Solar Corona
• So far, we have been concerned with the Naked Star
Gravitational Lens, due to a spherical distribution of Mass.
• However, above the surface of the Sun, the Corona extends
into space across distances that are comparable to the Sun
radius, and the coronal effects may only complicate the
physical picture of the Sun as a gravitational lens.
• How does the Sun Corona affect the electromagnetic waves
convergence? How does the Sun Corona affect the
electromagnetic waves gain?
• We used the Baumbach-Allen CORONA MODEL in Chapters
8 and 9 of our book, as the JPL expert John L. West did in
1999 in Acta Astronautica, Vol. 44, (1999), pp. 99-107.
16
 Electrons in the Corona
“pushing the Focus Out” 1100

1050
FocusGravityPlasma( b SunRadius 59.514GHz
 )
AU 1000

FocusGravityPlasma( b SunRadius 122.3GHz

 )
950
AU

FocusGravityPlasma( b SunRadius CMBpeak) 900

AU

FocusGravityPlasma( b SunRadius 300GHz) 850

AU
800
FocusGravityPlasma( b SunRadius 500GHz)
AU
750
FocusGravityPlasma( b SunRadius 1000GHz
 )
AU
700
17  1   2
( b SunRadius)
30 rg
650
AU

600

550
1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 17
b
TWO TETHERED ANTENNAE

18
INTERFEROMETRY in space

19
 EXAMPLES: Three Targets

• We now provide THREE EXAMPLES of

Targets for three different FOCAL missions:
• 1) The Galactic Black Hole (i.e. a FOCAL
mission for Astrophysics and Cosmology).
• 2) The Alpha Centauri system of three stars,
Alpha Cen A, B and C (Proxima) at 4.37 ly
(basically a FOCAL mission to radio-explore the
first target for any really interstellar mission).
• 3) Any Extrasolar Planet, for instance the
20
Earth-size recently discovered Gliese 581 e.
#1 Target: Galactic Black Hole

21
 #1 Target: Galactic Black Hole
• From the similarity of the two triangles one
gets:
Minimal Tether Length 2 rSchwarzschild of Galactic Black Hole

550 AU 32,000 light years

• Hence the MINIMUM TETHER LENGTH :

Minimum Tether Length = 1.6 km.

• IT CAN BE DONE !!!

22
#2 Target: Alpha Cen A, B, C.

• Real and Apparent Orbits of Alpha Cen B wrt A:

23
#2 Target: Alpha Cen A, B, C.
• From the similarity of the two triangles one
gets:
 22 * arcsec  6
550 * AU * tan   4.388 * 10 km
 2 
• Hence the MINIMUM TETHER LENGTH :

Minimum Tether Length ~ 1 Million km.

• THIS TETHER CANNOT BE MADE !!!

24
#2 Target: Alpha Cen A, B, C.

• But we can change the trajectory to a

CONICAL HELIX, like the profile of the
Guggenheim Museum in New York City:

25
#3 Target: An Extrasolar Planet
• For instance, consider Gliese 581 e (or Gl 581 e), the
fourth extrasolar planet just found around Gliese 581.
• At a minimum of 1.9 Earth masses, it is the smallest
extrasolar planet discovered around a normal star,
and the closest in mass to Earth. See Wikipedia site.

26
 #3 Target: An Extrasolar Planet
• The LINEAR RESOLUTION provided by FOCAL at
distance z from the Sun and observing frequency  is:
c2 1
RObject  d SunObject resolution  d SunObject 2
.
2 GM Sun z
• Here we know the object distance from the Hipparcos
Star Catalogue, and we know the Focal distance z.
• Thus, we can select the observing frequency (one
or many) at will: THE HIGHER THE FREQUENCY,
THE BETTER IMAGES OF THIS EXOPLANET WE
27
GET !!!
NASA Interstellar Probe (ISP)

Interstellar Probe
Exploring the Interstellar
Medium and the
Boundaries
of the Heliosphere

                                                   
• Interstellar Probe
                                                   
• Exploring the
                                                  
Interstellar
Medium and the
  Boundaries of the
Heliosphere

28
ESA Interstellar Probe

29
Trajectory of NASA/ESA Probes

•Arrival of interstellar
hydrogen, helium, and
oxygen atoms as seen in
the IBEX-Lo sky maps.
The Sun's gravitation
deflects the interstellar
wind away from its
original arrival direction,
i.e. coming from the nose
of the heliosphere. (Image
by the University of New
Hampshire and Boston
University)

30
Incoming Interstellar Wind
•Arrival of
interstellar
hydrogen,
helium, and
oxygen atoms
as seen in the
IBEX-Lo sky
maps. The Sun's
gravitation
deflects the
interstellar wind
away from its
original arrival
direction, i.e.
coming from
the nose of the
heliosphere.
31
CMB-looking NASA/ESA Probe
• The Cosmic Microwave Background (CMB) is (nearly)
uniformly distributed all over the celestial sphere.
• This means that the FOCAL probe will MAGNIFY
the CMB no matter which outward direction from Sun
is selected. So, that applies to ALL FOCAL MISSIONS.
• The only important data is the frequency of the CMB peak
(Planck black-body distribution at T=2.725 K) at 160 GHz.
• At this frequency the real focus of the Sun, i.e. taking the
Corona into account, is “pushed out” by the electrons in the
Corona from 550 AU to a real distance of 763 AU.
• 763 AU AWAY FORM THE SUN MUST BE 32
REACHED TO HAVE THE CMB MAGNIFIED.
CMB MAGNIFIED by FOCAL
• Suppose that FOCAL reached 763 AU.
• This means that the FOCAL probe will MAGNIFY
the CMB but… how much?
• This author estimated the angular resolution of the
CMB magnified by the Sun’s Gravity Lens to be about 5
x 10^(-8) arcsec (see page 146 of his book) at the peak
frequency of 160 GHz.
• This is about A BILLION TIMES better than any CMB
probe launched so far (COBE (1989), WMAP (2001),
PLANCK (2009).
• This should bring the COSMOLOGISTS to SUPPORT 33
the FOCAL space mission to 763 AU.
 SETI INSTITUTE
Mountain View, California, USA, November 25th, 2009

KLT: Principal Component Analysis

in the Hilbert Space spanned by the
Eigenfunctions of the Autocorrelation
of the: “Noise plus a possible ET signal”.
Claudio Maccone
Member of the International Academy of Astronautics
E-mail: clmaccon@libero.it
Home page: http://www.maccone.com/
34
 What is the KLT ?

• Example (a Newtonian analogy):

consider a solid object, like a BOOK,
x
described by its INERTIA MATRIX.
• Then there exist only one special
y
reference frame where the Inertia
Matrix is DIAGONAL. This is the x
 I'xxx I0xy I0xz 
reference frame spanned by the  x 
I '   I0y I 'xxyy x
I0x 
EIGENVECTORS of the Inertia
z  0x
I0xx I 'xxzz 
 Iz

matrix.

35
 KLT mathematics
• If X t  is a stochastic process (= input to the radio
telescope) it can be expanded into an infinite series

X t    Z n n t  0t T
• then n 1
 n t  are orthonormalized functions of the time:
T

  t   t  dt  
0
m n mn

Z n are random variables, not changing in time, with

the property
Z m Z n  n  mn

In conclusion, the KLT separates the

radiotelescope input (= noise + signal(s))
into UNCORRELATED components. 36
 KLT mathematics
T

 X t  X t 
0
1 2 n t1  dt1  n n t2 

• This is the INTEGRAL EQUATION yielding the KLT

eigenfunctions n t  and corresponding eigenvalues n .
• The kernel of this integral equation is the autocorrelation.
• This yields the best basis in the Hilbert space describing the
(signal+noise). The KLT adapts itself to the shape of the
radiotelescope input (signal+noise) by selecting, as a reference
frame, the one spanned by the eigenfunctions of the
autocorrelation. And this is turns out to be just a LINEAR
transformation of coordinates in the Hilbert space. Thus, 37 the
KLT is an easily INVERTIBLE LINEAR TRANSFORMATION.
 KLT filtering
• There is no degeneracy (i.e. each eigenvalue
corresponds to just one eigenfuction only).

• The eigenvalues turn out to be the variances of the

random variables Z n , that is  Z2n  n  Z n2  0.

• Since Z n  0 we can SORT in descending order of

magnitude both the eigenvalues and the corresponding
eigenfunctions. Then, if we decide to consider only the
first few eigenfunctions as the “bulk” of the signal, and to
apply the inverse KLT, that’s what KLT filtering is: we just
declare the taken-off part as “noise”!

• The Galileo mission by NASA-JPL used the KLT… 38

“Classical” KLT vs. FFT

KLT FFT
Works well for both wide Rigorously true for narrow
and narrow band signals band signals only

Works for both stationary Works OK for stationary

and non-stationary input input stochastic processes
stochastic processes only
Is defined for any finite Is plagued by the
time interval “windowing” problems

Needs high computational Fast algorithm FFT

burden: no “fast” KLT
39
NO Classical KLT in the 1990s

• If N is the size of the autocorrelation matrix ( N may equal

millions ore more in real SETI), the number of calculations
requested to find the KLT is of the order of N square, while
the same number for the FFT is much less: just N ln(N).

• This COMPUTATIONAL BURDEN prevented all SETI

scientists from replacing the FFT by the KLT until 2007:
1) François Biraud et al. at Nançay in the years after 1983.
2) Bob Dixon et al. the Ohio State SETI Program after 1985.
3) Stelio Montebugnoli et al. at Medicina, Italy, after 1990.
40
BAM (=Bordered Autocorrelation
Method) to EASILY find the KLT
• BAM is acronym for “BORDERED Autocorrelation Method”.
• The new key idea is to regard the autocorrelation as a NEW
FUNCTION OF KLT FINAL INSTANT, T. That is, to add
one more row and one more column (= bordering) to the
autocorrelation matrix for each new positive, increasing T.
• For STATIONARY processes, this amounts to the matrix:

41
2007 Breakthrough in the KLT
• In the winter of 2006-7, the SETI-Italia Group at Medicina,
(Stelio Montebugnoli, Francesco Schillirò, Salvo Pluchino
and Claudio Maccone) discovered a way to CIRCUMVENT
2
that big obstacle of the KLT N computational burden.
• The idea is to use the BAM to exploit the dependence on
the final instant T in both sides of the relationship, pointed
out back in 1994 by Maccone in his KLT book, p.12, eq.
(1.13). This is called “Mercer’s Theorem” in the theory of
integral equations since discovered by James Mercer (1909).

 T


n 1
n T    X2 t  dt
 42
0
BAM (Bordered Autocorrelation
Method) to EASILY find the KLT
• Differentiating both sides wrt T yields the FINAL
VARIANCE THEOREM (Maccone, Proc. of Science, 2007):

n T 

n 1 T
2
  X t .

• Let us now confine ourselves to STATIONARY processes

X(t) and to the FIRST (“dominant”) EIGENVALUE only:

n T  1 T 

n 1 T

T
2
   CONSTANT _ wrt _ T .

• The SETI-Italia Team found that the Fourier transform of

43
this CONSTANT is the FREQUENCY of the ET SIGNAL .
 KLT & Special Relativity

• In Relativity TWO DIFFERENTLY-ELAPSING TIMES exist:

• 1) The COORDINATE TIME t, elapsing FAST since it is the
time of those who move slowly wrt the speed of light c.
• 2) The PROPER TIME  elapsing SLOWLY since it is the
time of those on spaceships moving with a high v/c ratio.
• 3) It can be proven that, if the coordinate-speed v(t) of the
moving spaceship is known, the two times are related by:

t
2
v w
 t    1  2 dw
0 c
44
Gaussian NOISE in Space is like

Standard Brownian Motion B(t).

• Consider the standard Brownian Motion, B(t), with B(0)=0 :
x2
1 
f B t  x   e 2t
with f B 0  x    x .
2 t
• Its mean value is zero: Bt   0.
• Its standard deviation is the horizontal parabola  B t    t .

45
KLT of Standard Brownian Motion

• The KLT of Standard Brownian Mot. can be found analytically:

 
2   2n  1 
Bt    Z n n t    Z n  sin  t .
n 1 n 1 T  2T 
• Thus, the eigenfunctions of the KLT of Standard Brownian
Motion are just SINES, i.e. this KLT is the same as Fourier !
• The relevant eigenvalues are given by the sequence:

2
4T
n  2 , n  1, 2, ... .
 2n  1
2
46
RELATIVISTIC spaceships need
a Relativistic RADIO THEORY
• It means that the Brownian Motion received on
Earth from a relativistic spaceship depends on the
PROPER time B() according to the proper-time
formula

 t v 2
w 
B   B t   B  1  2 dw .

n
 0 c 
 
• This makes us wonder whether a TIME-RESCALED Theory
of Brownian Motion might possibly be created…
• The answer is “YES”, and mathematicians call it the Theory
of White-Noise Integrals: integrals where the integrand is the
Brownian Motion itself times a time-rescaling function f(t).47
Brownian Motion RESCALED
IN TIME by an arbitrary f(t)
• Any TIME-RESCALED Brownian Motion, X(t),
may be re-written as a “White Noise Integral”:

dBt 
Since White Noise  W t   , then :
n
dt
t
X t    f w dBw 
0

dBw
t t
  f w dw   f w W w dw.
0
dw 0 48
Brownian Motion RESCALED
IN TIME by an arbitrary f(t)
• On the other hand, it may also be proven that this
White-Noise Integral is the same as the following
TIME-RESCALED Brownian Motion, i.e. Brownian
Motion where the time does not elapse uniformly:
t
X t    f w dBw 
n

 B  f 2 wdw  .
 t

 0 
49
 RELATIVISTIC RADIO LINK :
RELATING v(t) AND f(t) .
• We can now understand the relationship between the
relativistic speed of the moving spaceship, v(t), and the time-
rescaling function f(t) appearing in the white-noise integral:

v w v t  2
t 2 2
t
 1  2 dw   f w dw , that is 1  2  f t 
n
2

0
c 0 c

4
v 
2
t 
and finally f t   1  2 or v(t )  c 1  f t .
4

c
50
 4 Relativistic SPEED PROFILES

2
v
1) Constant speed : vt   vconst then f t   4 1  const
2
 constant.
c
t1
1
-
gt  gt 2
 4
2) Hyperbolic Motion : vt   then f t   1     .
gt
2
  c  
1  
 c 
c2 1 c 1
3)Asymptotic Hyperb. Motion : vt   c 1  2 2 then f t   .
g t g t
1
2 H
 t 2 H 1  t 2
4) Decelerated Motion : vt   c 1   2 H 1  then f t   .
1 51
T  H
T 2
KLT of the TIME-RESCALED B(t)

• In his 1984 paper entitled “Eigenfunctions and Energy for

Time-Rescaled Gaussian Processes” (Bollettino dell’Unione
Matematica Italiana, Series 6, Vol. 3 (1988) pp. 213-217), this
author was able to solve the KLT integral equation
analytically for ALL THE TIME-RESCALED GAUSSIAN
PROCESSES X(t).
• He proved that the KLT eigenfunctions are time-rescaled
BESSEL FUNCTIONS OF THE FIRST KIND J(t).
• He also proved that the KLT eigenvalues are the ZEROS of
certain linear combinations of such Bessel functions and their
derivatives wrt time.
• These results are proven in Appendix G of his 2009 book.52
CONCLUSION: relativistic KLT
CODES are needed now !
• Chapters 11 thru 13, and the Appendixes H, I, and J of the
author’s book, contain the mathematically different forms of
Bessel functions of the first kind that actually are the best
basis in the Hilbert space to detect relativistic signals.
• The time is now ripe to write down COMPUTER CODES
capable of computing the KLT eigenfunctions for different
cases of high-speed spaceflight. The time-rescaled BESSEL
FUNCTIONS OF THE FIRST KIND J(t) are the key tool
to this end.
• Good luck and good RELATIVISTIC LINKS  !

53
Deep Space Flight
& Communications

54
Thanks !

55