Proposal for GSoC on gr-gsm

GNU Radio
1

Hi all,

As Martin B. noticed in the mail list yesterday, I haven’t open my
proposal here.

Here is the link:

If you have any suggestions, please don’t be hesitate to contact me.

Thanks to Sylvain M. for his help on reviewing my first draft.
Thanks to Piot Krysik for his suggestions on gr-gsm.

Best wishes,
Zhenhua

2

Hi Zhenhua,

gr-gsm should be a good project for you and for the community. I wish
you success.

You may want to look at the slow frequency hopping implementation of
airprobe that is placed here:
https://github.com/BogdanDIA/airprobe-hopping.
It uses PFB to split a 20mhz band received with USRP-N210 into GSM
channels and then use all of them according to the hopping algorithm
decoded on the Immediate Assignment message and SI 1. It is not a
perfect implementation as airprobe is not perfect too, that is why I
think this will be a good project.

The hopping patch has been done two years ago and has API for
gnuradio3.6 era. I did not update it yet on 3.7. Here is small paper I
wrote at the moment of creating a patch. It may be of help to you.

Bogdan

On Tuesday, March 11, 2014 5:40 AM, zhenhua han [email protected]
wrote:

Hi all,

As Martin B. noticed in the mail list yesterday, I haven’t open my
proposal here.

Here is the link:GitHub - hzhua/grgsm_proposal: Proposal for GSoC 2014 on gr-gsm

If you have any suggestions, please don’t be hesitate to contact me.

Thanks toSylvain M. for his help on reviewing my first draft.
Thanks toPiot Krysik for his suggestions on gr-gsm.

Best wishes,
Zhenhua

3

On 03/11/2014 04:40 AM, zhenhua han wrote:

Thanks to Sylvain M. for his help on reviewing my first draft.
Thanks to Piot Krysik for his suggestions on gr-gsm.

Hey,

congrats on publishing the first proposal! It looks quite good.

It’s unclear if you have good algorithms to implement your deliverables.
A few hints:

There’s also some code for this somewhere… can’t find it right now.

  • A CEL student wrote a highly optimized MLSE, specifically for GSM, a
    while back (a long while back, it still uses autotools). Still, the
    code is very good, and there’s a lot to be taken out of this. He tried
    out a lot of compiler switches, too, so check out lib/Makefile.am.

It assumes frame synchronization (i.e. post-SCCH sync), but you just
drop in a frame, and it returns the equalized result.

MB

4

Thank you, Martin.
I will check these and extend my proposal with more details.

Best wishes,
Zhenhua

2014-03-11 17:00 GMT+08:00 Martin B. [email protected]: