Re: [Discuss-gnuradio] GSOC 2016 :Implement optimized, standardized channel codes

[Jan Krämer]

Hi Vlad,

nice job on the simulations. Looks really good. I suggest you put the results and a (short) explanation in your proposal. Remember monday is the first day to submit it officially.

Also although I messed up the first time as well and had to sent the email twice, try to remember to include the GNURadio list when replying. The more mentors see the mails, the better for you!

Cheers,

Jan

2016-03-11 19:13 GMT+01:00 Влад Василенко <address@hidden>:

Hi Jan,

I would suggest the implementation of a pseudo-random interleavers, namely randon and s-random. It is more difficult to implement, but have a better performance. Used in turbo codes, they interleaver purpose is to offer each uncorrelated or random encoder version information, whereby each encoder parity bits are independent. The degree of independence of the parity bits is essentially a function of the type of length / depth of the interleaver.

The attached files contain BER graphics, depending on various parameters.

Figure 1 shows the effect of N input block size N on BER to Log Map decoding algorithm, 8 decoding iterations polynomial generators g0 = 7, g1 = 5, data rate in the Turbo code R = 1/2. At 1000 and 10000 symbols used in the transmitted pseudo-random interleaver block, and at 169 - the block.

Figure 2 shows the effect inte interleaving on BER to Log Map decoding algorithm, 8 decoding iterations polynomial generators g0 = 7, g1 = 5, data rate in the Turbo code R = 1/3. At 1000 and 10000 symbols used in the transmitted pseudo-random interleaver block, and at 169 - the block.

Figure 3 shows the effect of limiting the length of the encoder and polynomial generator on the characteristics of Turbo code.

These graphs were obtained by me in the course of the simulation.

I have implemented a system for simulating the operation of the turbo code. I implements the encoder, the decoder decoding with different algorithms(MAP-algorithm, LOG-MAP algorithm, MAX-LOG-MAP algorithm). It is also sold various types of interleavers (block, diagonal, random, s-random).

Best Regards,

Vlad Vasylenko

2016-03-08 9:26 GMT+02:00 Jan Krämer <address@hidden>:

Hey Vlad,

glad to hear that you're interested in GSoC and GNURadio. For starters I would like to refer you to the Email that Martin Braun send to all applicants. http://lists.gnu.org/archive/html/discuss-gnuradio/2016-03/msg00161.html

My advice is to now get familiar with GNURadio [0], VOLK [1], and the FEC-API [2] (the GNURadio Forward Error Correction Framework).

Also one note, if you want to implement various types of interleavers other than the standard ones, you might have to convince

 us that they will actually be an improvement in your proposal.

Looking forward to reading the proposal,

Jan

[0] https://gnuradio.org/redmine/projects/gnuradio/wiki/Guided_Tutorials
[1] http://libvolk.org/
[2] https://gnuradio.org/doc/doxygen/page_fec.html

2016-03-07 12:41 GMT+01:00 Влад Василенко <address@hidden>:

I have a master's degree in computer systems and networks Chernihiv National Technological University. Currently I’m PhD student  at Institute of Telecommunications and Global Information Space of the Academy of Sciences of Ukraine.

I am interested in your project - Implement optimized, standardized channel codes, Mentor(s) Jan Krämer, Sebastian Koslowski.

I want to do the implementation of LTE Turbo code, because for a long time doing their research. My thesis topic is based on their research. I have the copyright certificate and an article in this area.  I would like to propose another implementation of various types of interleavers. (Depending on the size of the block interleaver type affects BER).

I have skills in programming languages C/C++, Matlab.

I will appreciate any response of any form.

Regards

Vlad Vasylenko

_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio