Forum: GNU Radio Google Summer of Code 2013

Posted by Martin Braun (CEL) (Guest)
on 2013-02-12 10:26
(Received via mailing list)
Hi everyone!

Admittedly, the time between the pre-announcement and this email is
pretty short, but well, Google seems to predict myself better than I can
predict Google.

So, the Google Summer of Code (GSoC) has been announced.
Most people around here would like to see GNU Radio participate,
including myself, but that depends on YOU.
We need your ideas for projects, and we need mentors.

For the impatient, here's a tl;dr:

1) Think about cool projects that can be done by motivated students
   in ~3 months and post them here:
http://gnuradio.org/redmine/projects/gnuradio/wiki/GSoC
   Don't be shy! This is the brainstorming phase. Anything goes.
   Anyone can post ideas (as long as they're GNU Radio-related).
2) Can you mentor a project. Be honest, you can, can't you!
   If so, put your name below a project you'd like to mentor, or simply
   sign up without a specific project. (Right now, 'signing up' means
   telling us, hey, I'd like to mentor. The GSoC page doesn't have a
   sign-up feature yet).



OK, now more detail:

What is the GSoC and why is it good for us?
===========================================
In GSoC, Google will fund a student for 3 months to contribute code to
a free software project. This project could be GNU Radio! Ideally, this
means we will get fresh code into our repositories, and motivated young
students to join the project, and stay on.
In 2012, we were quite successful in that respect. This just isn't
something we can miss out on.


How does this work?
===================
The details are on
https://www.google-melange.com/gsoc/events/google/gsoc2013.
The quick version is:
1) We create a great list of ideas for projects (this is important!)
2) A couple of us sign up as mentors (the more, the better)
3) Starting April, students can contact mentors about projects and
   prepare their applications.
4) If we're lucky, a couple of students are assigned projects. Then they
   start working on them.
5) The mentors grade the students twice; once at half-term, once at the
   end.
6) Ideally, at the end, we have cool new code and new contributors!


How about coding a .... ?
==========================
Excellent idea! Please post this on the wiki:
http://gnuradio.org/redmine/projects/gnuradio/wiki/GSoC

We'll be going over the list towards the end of the idea-collecting
phase. At the end, projects must:
* be possible in the GSoC timeframe (~3 months) and
* advance the GNU Radio project in some way.

They don't necessarily have to be something for the core GNU Radio
codebase; out-of-tree modules for GNU Radio are fine, too.

Of course, it should also concur with Google's own rules and
requirements (check our wiki page and the GSoC FAQ from last year).

In 2012, we had lots of ideas that never made it into GSoC projects.
Perhaps they're still valid, and could therefore be suggested again. All
the old ideas have been left on the page as an inspiration.


I'm a student and would like to participate, what do I do?
==========================================================
Read the FAQ for all the paperwork etc. As far as we're concerned, you
should:
- Think about projects and post ideas of your own
- Be involved on the mailing list. If we know your name, that *does*
  increase your chances. Active contributor's applications simply are
  more convincing than "You've never heard of me, but I really want to
  contribute", even if you're sincere. Of course, don't let that stop if
  you if you really want to participate!

Note that we can't guarantee anyone a slot right now.
Google has it's own process, and we have to stick to that.

Who can mentor? And what does that mean?
========================================
Pretty much anyone who would like to!
You don't have to be associated with academia or anything like that.
You *should* have some time available for the mentoring, of course.

Mentoring is a great way to do something good and advance the GNU Radio
project without actually having to do the hard work :)
Perhaps you have some expertise you'd like to bring into the project, or
you just like tutoring students--in any case, mentoring might be
something you'd enjoy.

Put briefly, a mentor supervises a student's work during the GSoC
duration and grades the work (one evaluation at mid-term and one at the
end).
Google has put up a great mentoring manual, which gives some insight
into what mentoring means: http://en.flossmanuals.net/GSoCMentoring/

Speaking out of personal experience, I mentored the filter design
project last year. It was great fun, working across continents was not a
problem and the actual time spent wasn't too extreme.
I'll be happy to answer questions about mentoring on this list (if
they're not covered in Google's FAQs etc.)


MB

--
Karlsruhe Institute of Technology (KIT)
Communications Engineering Lab (CEL)

Dipl.-Ing. Martin Braun
Research Associate

Kaiserstraße 12
Building 05.01
76131 Karlsruhe

Phone: +49 721 608-43790
Fax: +49 721 608-46071
www.cel.kit.edu

KIT -- University of the State of Baden-Württemberg and
National Laboratory of the Helmholtz Association
Posted by Martin Braun (CEL) (Guest)
on 2013-02-12 12:45
(Received via mailing list)
On Tue, Feb 12, 2013 at 10:24:56AM +0100, Martin Braun (CEL) wrote:
> OK, now more detail:
>
> [stuff on GSoC]

Faculty and GSoC
================
One thing I'd like to add: Participants must be students, of course.

This implies that people will be doing GSoC while enrolled in some kind
of university degree.
I believe the whole thing is timed to sync with summer break at US
universities.  Anywhere else, it might conflict with other university
stuff. In this case, I'd like to ask universities to consider supporting
students participating in GSoC. In some cases this can be difficult, but
say you're an EE department, and your student is writing signal 
processing
code--surely there must be some way to support your student, e.g. by
rewarding her or him with credits.

MB

--
Karlsruhe Institute of Technology (KIT)
Communications Engineering Lab (CEL)

Dipl.-Ing. Martin Braun
Research Associate

Kaiserstraße 12
Building 05.01
76131 Karlsruhe

Phone: +49 721 608-43790
Fax: +49 721 608-46071
www.cel.kit.edu

KIT -- University of the State of Baden-Württemberg and
National Laboratory of the Helmholtz Association
Posted by Martin Braun (CEL) (Guest)
on 2013-02-27 18:41
(Received via mailing list)
Ideas, ideas, ideas!

Everyone, we need ideas.
The GSoC application deadline is coming closer, and we still are lacking
some ideas.

So what cool feature of GNU Radio are *you* missing?

Head over here, and write them down:
http://gnuradio.org/redmine/projects/gnuradio/wiki/GSoC

At this point, any random idea is OK. Also, you're not signing up for
anything if you post an idea.

MB

--
Karlsruhe Institute of Technology (KIT)
Communications Engineering Lab (CEL)

Dipl.-Ing. Martin Braun
Research Associate

Kaiserstraße 12
Building 05.01
76131 Karlsruhe

Phone: +49 721 608-43790
Fax: +49 721 608-46071
www.cel.kit.edu

KIT -- University of the State of Baden-Württemberg and
National Laboratory of the Helmholtz Association
Posted by Dan CaJacob (Guest)
on 2013-02-27 19:52
(Received via mailing list)
Really great documentation would be nice.  I know that it has been
improving, but maybe GSOC is an excuse for a sprint?  Ideally, the
documenter is someone who really knows DSP and what is going on behind 
the
curtains.

References to examples in the documentation would help out immensely.  I
often resort to hunting down a block's source code or other blocks and
examples that use the block to understand it.

I have always wanted an auto-magically generated list of references to
other blocks/examples that use the block being documented.  I think this
would be easy to do in python.  When building the docs, a script could
search for instances of each block in other blocks and examples, then
insert them as references in the block's documentation.

As someone with little DSP backrground, I would like to be able to see 
some
great documentation when I open up a block in GRC or introspect it in
python.

Very Respectfully,

Dan CaJacob

-- This email contains sensitive proprietary and confidential 
information.
The technical data contained herein is/may be controlled under the U.S.
International Traffic in Arms Regulations (ITAR) and may not be exported 
to
a Foreign Person, either in the U.S. or abroad, without the proper
authorization by the U.S. Department of State. --


On Wed, Feb 27, 2013 at 12:39 PM, Martin Braun (CEL)
Posted by Johnathan Corgan (Guest)
on 2013-02-27 19:58
(Received via mailing list)
On Wed, Feb 27, 2013 at 10:49 AM, Dan CaJacob <dan.cajacob@gmail.com> 
wrote:


> I have always wanted an auto-magically generated list of references to
> other blocks/examples that use the block being documented.  I think this
> would be easy to do in python.  When building the docs, a script could
> search for instances of each block in other blocks and examples, then
> insert them as references in the block's documentation.
>

Great idea.  Let us know if you do any work towards this--I'd be happy 
to
review and see how to integrate it into the main GNU Radio software 
base.

Johnathan
Posted by Philip Balister (Guest)
on 2013-02-27 20:05
(Received via mailing list)
On 02/27/2013 10:49 AM, Dan CaJacob wrote:
> Really great documentation would be nice.  I know that it has been
> improving, but maybe GSOC is an excuse for a sprint?  Ideally, the
> documenter is someone who really knows DSP and what is going on behind the
> curtains.

GSoC is for code related work, not documentation work.

However, a proposal to write some code that would improve usability
should be OK.

Philip
Posted by Carles Fernandez (Guest)
on 2013-02-27 20:14
(Received via mailing list)
Hi there,

I would like to ask you about the participation in GSoC 2013 of the
GNSS-SDR project under the umbrella of GNU Radio. Last year I served as 
a
mentor and we all (the student, other developers and I) had a nice
experience adding Galileo capabilities to the GNSS software receiver. We
enjoyed it, benefited from it, and I would love to participate again in
this year's edition, but I personally feel that only GNSS-SDR took
advantage from the work done in GSoC 2012, having no impact for the
majority of GNU Radio users.

GNSS-SDR is an open source Global Navigation Satellite System software
receiver written in C++ that uses GNU Radio's scheduler, block hierarchy
and some processing blocks. In that sense, we are *users* of GNU Radio 
but
we do not contribute directly to the project. Since I'm aware that the
slots are limited, I understand that developing other features for 
general
GNU Radio usage (e.g. the Great Documentation proposal) can have a wider
impact and more people can benefit from it.

Otherwise, if you feel that having third-party applications based on GNU
Radio is positive to the project, I will immediately rewrite the
description at http://gnuradio.org/redmine/projects/gnuradio/wiki/GSoC 
with
new (and hopefully cool) proposals for this summer :-)

Best regards,
Carles
Posted by Johnathan Corgan (Guest)
on 2013-02-27 20:21
(Received via mailing list)
On Wed, Feb 27, 2013 at 11:12 AM, Carles Fernandez <
carles.fernandez@gmail.com> wrote:


> ...but I personally feel that only GNSS-SDR took advantage from the work
> done in GSoC 2012, having no impact for the majority of GNU Radio users.
>

As we discussed at the GNU Radio conference last year, it would be very
useful to go through the GNSS-SDR code to identify signal processing 
blocks
you had to create, but weren't directly part of the GPS/Galileo
implementation code.  These are good candidates for things that are 
missing
from GNU Radio and would of be of use to the wider community.

Johnathan
Posted by Martin Braun (CEL) (Guest)
on 2013-02-28 11:24
(Received via mailing list)
On Wed, Feb 27, 2013 at 08:12:46PM +0100, Carles Fernandez wrote:
> I would like to ask you about the participation in GSoC 2013 of the GNSS-SDR
> project under the umbrella of GNU Radio. Last year I served as a mentor and we
> all (the student, other developers and I) had a nice experience adding Galileo
> capabilities to the GNSS software receiver. We enjoyed it, benefited from it,
> and I would love to participate again in this year's edition, but I personally
> feel that only GNSS-SDR took advantage from the work done in GSoC 2012, having
> no impact for the majority of GNU Radio users.

Just a follow-up, in case there's people wondering if their idea/project
is OK to go on the wiki page: GNSS-SDR is a successful and (kind of)
independent project, so they will participate on their own.

This does *not* mean that all the GSoC projects have to go straight into
gnuradio/master. An extension to GNU Radio as an out-of-tree module is
OK, too (most of the projects on CGRAN would qualify).

I've updated the GSoC page with some more ideas:
http://gnuradio.org/redmine/projects/gnuradio/wiki/GSoC
(The first mentor has also already signed up, shoutouts to Jens!)

Perhaps this gives people an idea about potential ideas.

MB

--
Karlsruhe Institute of Technology (KIT)
Communications Engineering Lab (CEL)

Dipl.-Ing. Martin Braun
Research Associate

Kaiserstraße 12
Building 05.01
76131 Karlsruhe

Phone: +49 721 608-43790
Fax: +49 721 608-46071
www.cel.kit.edu

KIT -- University of the State of Baden-Württemberg and
National Laboratory of the Helmholtz Association
Posted by Alex Zhang (Guest)
on 2013-02-28 18:39
(Received via mailing list)
For me, I am very interested in that is there any open project to 
improve
the data rate and packet loss for the GNURadio based OFDM?
Some guys claimed that the LTE standards are implemented with USRP and 
PC
by SIMD programming. But their code are not open.

If GNURadio based OFDM can support high data rate as LTE, the PC based
access point or LTE base station can be used for a wide range of 
research,
especially the real large scale network.


On Wed, Feb 27, 2013 at 11:39 AM, Martin Braun (CEL)
<martin.braun@kit.edu>wrote:

>
> Research Associate
> National Laboratory of the Helmholtz Association
>
> _______________________________________________
> Discuss-gnuradio mailing list
> Discuss-gnuradio@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>
>


--

Alex,
*Dreams can come true  just believe.*
Posted by Martin Braun (CEL) (Guest)
on 2013-02-28 18:52
(Received via mailing list)
On Thu, Feb 28, 2013 at 11:38:40AM -0600, Alex Zhang wrote:
> For me, I am very interested in that is there any open project to improve the
> data rate and packet loss for the GNURadio based OFDM?
>
> Some guys claimed that the LTE standards are implemented with USRP and PC by
> SIMD programming. But their code are not open.
>
> If GNURadio based OFDM can support high data rate as LTE, the PC based access
> point or LTE base station can be used for a wide range of research, especially
> the real large scale network.

Hi Alex,

perhaps you might be interested in two ongoing projects:

https://github.com/benreynwar/gnuradio/tree/ofdm
is a rehaul of the current OFDM implementation.

https://github.com/kit-cel/gr-lte
is an LTE receiver for the broadcast channels, although pretty early in
it's development.

MB

--
Karlsruhe Institute of Technology (KIT)
Communications Engineering Lab (CEL)

Dipl.-Ing. Martin Braun
Research Associate

Kaiserstraße 12
Building 05.01
76131 Karlsruhe

Phone: +49 721 608-43790
Fax: +49 721 608-46071
www.cel.kit.edu

KIT -- University of the State of Baden-Württemberg and
National Laboratory of the Helmholtz Association
Posted by Alex Zhang (Guest)
on 2013-02-28 18:56
(Received via mailing list)
It's cool! Thanks!


On Thu, Feb 28, 2013 at 11:51 AM, Martin Braun (CEL)
<martin.braun@kit.edu>wrote:

> access
>
> Dipl.-Ing. Martin Braun
> KIT -- University of the State of Baden-Wrttemberg and
> National Laboratory of the Helmholtz Association
>
> _______________________________________________
> Discuss-gnuradio mailing list
> Discuss-gnuradio@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
>
>


--

Alex,
*Dreams can come true  just believe.*
Posted by Dan CaJacob (Guest)
on 2013-03-01 16:43
(Received via mailing list)
Will do.  I'll take a crack at it this weekend.  I suppose the tricky 
bit
not be implementing this for source-based build and install, but for a
package-based install where source is not left behind.  Maybe in that 
case,
the doc example/reference links can point to the public repo.

Very Respectfully,

Dan CaJacob

-- This email contains sensitive proprietary and confidential 
information.
The technical data contained herein is/may be controlled under the U.S.
International Traffic in Arms Regulations (ITAR) and may not be exported 
to
a Foreign Person, either in the U.S. or abroad, without the proper
authorization by the U.S. Department of State. --


On Wed, Feb 27, 2013 at 1:56 PM, Johnathan Corgan
Please log in before posting. Registration is free and takes only a minute.
Existing account (Switch to SSL-encrypted connection)
NEW: Do you have a Google/GoogleMail or Yahoo account? No registration required!
Log in with Google account | Log in with Yahoo account
No account? Register here.