Chaos harnessed for encryption / Fluctuations and Order research

EE Times, Aug 9, 1993, p. 31 reports that "MIT's Research Lab of
Electronics is creating new signal processor designs, based on chaos
theory, that could open up a simple route to secure communications.

The new designs use a recent discovery called synchronized chaos to
transform a meaningful signal into what only seems to be random
noise., A similarly constructed receiver responds to the noisy signal,
sychronizing its own chaotic behaviour to extract the message.  The
MIT design requires only eight op-amps and is based on the Lorenz
attractor, which generates a simple three-dimensional chaotic system."

There's more, this is just a pointer.  Their current encryption system
is analog, not digital, and encrypts analog signals like audio; I
don't know if this is a fundamental design property or not.  They claim
it's not super-great encryption, just cheap and interesting.

Wired Sep/Oct 93 also reports (p.118) a Sep 9-12 conference on "Fluctuations
and Order" at Los Alamos National Labs' Center for Nonlinear Studies.
"The labs are gathering a couple dozen researchers who have realized
they can induce order into systems by using noise and randomness.  As
one abstract says, `The addition of noise to certain types of driven
systems can paradoxically cause a signal to become clearer.'"  These
seem related, to me.

	John

-----BEGIN PGP SIGNED MESSAGE-----

A recent _Scientific American_ had a brief piece on inducing order in
chaotic systems; I don't have it handy, but apparently it's fairly
simple to induce order in some nonlinear systems.

I'm no chaotician, but it seems that if you want to synchronize two
chaos generators at different sites, you must a) use the same initial
values and b) use the same mechanism to induce order. Granted that
small changes in a) or b) can change the system greatly, this doesn't
seem all that different from conventional synchronized encryption
systems.

(I'm happy to note that much of this work is being done at Georgia
Tech, my alma mater. It's great to be a fuzzy bee!)

- -Paul

- -- 
Paul Robichaux, KD4JZG     | "Crypto-anarchy means never having to say
perobich@ingr.com          |  you're sorry." - Tim May (tcmay@netcom.com)
Intergraph Federal Systems | Be a cryptography user- ask me how.


-----BEGIN PGP SIGNATURE-----
Version: 2.3a

iQCVAgUBLGk15SA78To+806NAQFLyQQAoZkg0VNeLCpfyqBtIDOsXcZQtBt0lo/Z
gOSS8p1Q2hSYAaO6NgGAgZ3dsVBSaGVpoGxMoIGlzbjNbJ72BEIRxiz2Itt3ul/s
DGbCIvqU8omph0msq8s2a3FBAnwE/yHfCbSHBPqmqRL29Bif7SpNh5qAc5JpEXBT
IjrsgcVa83I=
=4Mbz
-----END PGP SIGNATURE-----

This is how I see the situation:

 Neural Nets : Computers  ::  Chaotic analog encryption : DES

The chaotic encryption work depends on a secret algorithm, no less.  If
you want a system which works, do it digitally.  If you want to play and
get papers accepted to the next hot-topic-of-the-day conference, go play
with some op amps.  If you want to play, there's an article in SciAm
this month on building a chaotic "encryption" machine.  It probably
provides acceptable security if you use triple DES on signals prior to
sending them.