Carl R. Stevenson wrote:
wrote in message

No, robots do *not* listen before transmitting which is against the
regs and is the crux of the problem.

Brian,

Most of the "robots" *do* (at least make an attempt to) listen before
transmitting ... however the vagaries of propagation and the highly dynamic
nature of usage in the HF bands cause a serious "hidden terminal" problem
and that results in interference (it's unintentional, but still there - and
it happens more with "robots" because their "listen before talk" is not as
effective as a human sending "Is the frequency in use" and being
appropriately patient before blasting away

I realize that the things are advertsied as "listening before
transmitting" but from any practical point of view they simply don't to
any even half-decent extent, they're deaf as rocks. It's just
advertising hype.

A human operator causing QRM is either lousy operating practice or an
accident, a robot blindly causing QRM via it's inherent design is
illegal. One solution might be to come up with a robot which tunes
around it's frequency before transmitting. There's one for you
code-writers to chew on.

Solving the hidden terminal problem on HF for automated stations is a
difficult nut to crack ... in addition to the propagation issues and the
dynamics of usage, there are so many modes that a "robot" would have to
sense/detect/recognize to optimize the "clear channel assessment" and it
would have to do it quasi-continuously ...

Depends on how "optimum" is good enough to get the basic job done well
enough within the Part 97 hobby spectrum. I don't see why it would be
necessary for the algorithms to actually decode any of the modulations,
all they need to do is tune up & down 500Hz or so from their center
freq and sum the total level of signal activity/energy over some
reasonable short period of time vs. some threshold and make the
decision to transmit or not. Nor do I see why they should have to
listen continuously either. Once an acceptably intelligent robot
decides the freq is "clear enough" it's *his* and the devil can take
the hindmost. Which is exactly what human ops do.

As far as optimization goes how many times have you watched a
development effort die because the engineers hung too much gold plate
on it?

I'm not saying that it's a
permanently insoluble problem, but for now the mechanisms aren't up to the
level that's needed.

Mmmm . . I dunno . . I'm working on a project right now which is giving
me a good look at what neural network technology can do these days and
it's pretty impressive and it's not just academic pushups, it's fully
commercialized. I can easily visualize even simplistic implementations
of NNT giving robot stations ears which actually work to a useful
extent. But somebody has to actually quit talking about it and actually
DO it which is the ultimate tough nut to crack. Like actually getting
spread spectrum running on any ham band . . heh. Sorry, had to do that
for old time's sake Carl!

My working group, IEEE P802.22, (http://www.ieee802.org/22) is working on
"cognitive radio," but in response to the FCC's NPRM on license-exempt
devices using geographically unused TV channels ... this situation makes the
"incumbent detection/avoidance/protection" a more soluble problem because
there are a limited number of incumbents, they are high power transmitters
at generally fixed, stable locations, they use the same standards (NTSC,
which will be going away, and ATSC the new digital TV standard), the
spectral characteristics of their transmissions have "features" that are
easily detectable (the NTSC carriers or the DTV "pilot carrier"), etc.

.. . OK . . makes sense here . .

However the "detect and avoid" problem becomes much more difficult in an
environment with many lower powered stations that come and go, whose
locations vary, and who use a wide variety of different modulation
techniques ... again, these problems will likely be solved in the future,
but we're not there yet.

Agreed.

In the meantime in ham radio however we have the current flap over the
ARRL proposal to deal with. My bet is that in whatever any final form
the ARRL comes up with and submits the FCC will toss it back at us to
muddle through because the general public has no stake whatsoever, for
instance, in Pactor stations being rude.

73,
Carl - wk3c

w3rv