[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] SDR Design Competition

From: David Bengtson
Subject: Re: [Discuss-gnuradio] SDR Design Competition
Date: Fri, 19 May 2006 23:14:43 -0400
User-agent: Thunderbird (Windows/20060308)

Tom Rondeau wrote:
Well put, Lamar. I just wanted to add a few things. I've worked with the SDR
Forum for a while now, so I decided to ask them about the concerns raised on
this discussion board.

First, to the IP, they are not trying to take IP rights away from the
inventors. Their intent is to reserve the right to publish the work. This is
basically the same as the copyright transfer we always give when publishing
papers for IEEE or to the SDR Forum Technical Conference. They have
recognized problems with the wording, as raised by a number of people
directly involved with the competition, and are right now waiting for the
new wording to clear this up. The SDRF is a non-profit organization and has
no interest in holding IP of this sort.
I've been to a couple of the SDR Forum conferences, and I've been impressed with the organization. I'm glad they are clearing this IP issue up. I was quite surprised to read the IP rights clause there, based on other publications I've seen from the SDR Forum.

The competition is meant to allow students to get hands-on work with SDRs,
techniques, and tools. The corporate sponsors have agreed to provide some
tools and equipment to facilitate this process. I understand the concerns
with the corporate sponsorship, but again, they do not receive the IP
created. They are mostly interested in building familiarity with their
products, getting their names out to the community for future development,
and hopefully build loyalty to their products.

I think this competition is a good idea, and I'm very curious to see the projects that get submitted to this. I've been out of school for a while , but the projects that are listed as examples strike me as quite difficult for a group of students, but I certainly willing to be proven wrong.

If nothing else, entering this contest would look great on a student's resume, and would really seal the deal when it comes to hiring. I'm not sure that it would really matter how well they placed, as long as there was something there that the students could talk about in an interview.

Without these tools made available to the competitors, it wouldn't really
fly. Unfortunately, most people do not look at the GNU Radio right now as a
professional tool whereas MATLAB is, not to mention the fact that MATLAB is
a simple CD installation, and I think all of the "Can't get gnuradio-core to
compile" emails on this listserv prove the difficulty in working with this

Matlab/Simulink makes for a decent toolset. While it is expensive, there is quite an ecosystem that has developed around it, something that Octave/Scilab/numPython/etc. aren't really able to offer. Looking at the other difficulties in this project, it certainly seems like a decent decision to eliminate as many software glitches as possible.

As someone who has taught communications classes (and used the GNU Radio for
in-class demonstrations!), I can tell you that most electrical engineers in
communications are not comfortable with programming, which I am highly
critical of as the future of communications is in software. That aside, many
of them are uncomfortable working in MATLAB in a Windows environment. When I
tell my students that they will have to program in this class, I loose a
number of them, and the rest groan. I can't imagine what would happen if I
told them they had to install Linux and the GNU Radio code if they wanted to
work on this stuff at home!

They are going to have to get used to programming something if they plan on working as an engineer. A significant amount of my work hours are spent staring into
1) The Matlab IDE
2) Agilent's ADS
3) Microsoft Excel

and I'd really like the time to learn Verilog. While Computer algebra packages are nice, if you plan on implementing something in hardware, you will need something that outputs something that you can build an IC from. Perhaps a topic for another discussion.

Bottom line, I love the GNU Radio project, but the world of SDR has a long
way to go. It started mostly, and still is largely, with military money. We
are just now seeing it as a commercial and hobbyist possibility. That means
that much of the work comes from the military and the big contractors.
Getting more students involved is nothing short of a good thing. If the GNU
Radio community wants to get involved, too, I think everyone would be happy.
This is about the future of SDR and communications, and I for one would like
to see as many students get their hands on this stuff as possible and
provide the necessary cooperation and competition to push it to new heights.

Thanks, and I'll get rid of my soapbox now.

Tom Rondeau
Virginia Tech

-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of Lamar
Sent: Thursday, May 18, 2006 6:16 PM
To: John Gilmore; address@hidden
Subject: RE: [Discuss-gnuradio] SDR Design Competition

[You know, I might get flamed for this, but here goes....]

From: address@hidden on behalf of John
It looks like an incredible amount of work, under really picky and
idiotic rules, solving problems so challenging that there *isn't* any
commercial gear that does it, at any price.  For an unknown and
probably tiny reward.  And to hand it all over to somebody else to

Looking over the rules, FAQ, and phases, this looks  pretty normal, having
dealt with engineering academia before (for PARI, and during my own senior
year 17 years ago).

The purpose of these challenges is education of the students in practices
found in the industry.  Whether those practices are correct or not is not
even relevant, as industry practices are what they are.    The student is
getting the use of the development tools of the sponsors for free during the
scope of the challenge; it is an extremely good educational opportunity from
the educational point of view, and provides valuable people networking for
the students involved.  This sort of challenge mirrors the processes by
which industrial electrical engineering is actually done, not how we (myself
included) wishes it were done.

The rules and projects are ordinary in terms of actually engineering
industry practice, in my experience.  Reading through the sample challenges,
and understanding that Matlab, Simulink, and all of Xilinx's tools will be
made available, I don't see any of the challenges that would be too
difficult for a team of bright engineering juniors and seniors.

To give you an example, here at PARI we just finished a two semester
mechanical engineering project with NC A&T University.  In one year, the
students developed programs, techniques, skills, and processes to measure,
model, and change the balance of our 26 meter radio telescopes, each of
which weight over 300 tons.  The one is bottom heavy, and the other top
heavy.  For obvious reasons they began with the bottom heavy dish, and
measured torques, calculated moments, centers of mass, and weights, and then
recommended not only how much weight to remove, but the ideal (using finite
element analysis) weights to remove.  The upper axis had 2,200 pounds of
lead counterweight removed, and the lower axis around 6,000 pounds.  Oh, and
the students had to design the fixtures to remove the weights, and actually
help remove weights.

There were five students, and they completed all of the modelling and 50% of
the physical work (they hadn't counted on rust, for instance, on the three
inch bolts (not length; diameter!) holding the weights to the structure). But the bottom heavy dish is now much less bottom heavy (we wanted to keep
it stable, and not try to perfectly balance it; but they could have made the
balance perfect).

A team of five students can accomplish amazing things.
(They won't accept work that has been released under a public license,
such as the GPL or even the BSD license.  If you spend two years
writing this stuff, they will *own* it at the end, and you won't even
be able to keep working with or evolving your own software or
hardware.  And you won't be paid for any of this.)

And just what is wrong with any of this?  The sponsors provide tools; the
students use them for no charge, and get recognition, valuable experience,
and a great time.  That is, unless you want a blob of
Matlab/Simulink/Xilinx-centric code running around that requires those
tools.  The scoring is weighted towards those solutions that use the
sponsors' products (this is normal industry practice, too; put up the money,
and you make the rules).

You know, if the GNUradio Project wanted to sponsor such a contest and
provide USRP's for each team, then the GNUradio project can set the rules of
license.  I would love seeing that, actually.  By the way, I personally own
two USRP's; our technical director owns one; and PARI owns four.  So I
certainly believe in what the GNUradio project is about; but seeing an
antagonistic attitutde towards a normal educational senior project baffles

If we are going to change engineering practice, we have to get students
using the tools. To get them using the tools incentives have to be
provided.  Although, there are quite a few students using the GNUradio
toolkit already.

Not directed at John, but to Al: none of these students or their faculty
deserve the epithet 'suckers.'
Lamar Owen
Director of Information Technology
Pisgah Astronomical Research Institute
1 PARI Drive
Rosman, NC  28772

Discuss-gnuradio mailing list

Discuss-gnuradio mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]