Re: [Discuss-gnuradio] MIMO and USRP

[Matt Ettus]

john boon wrote:
> hi all,
>        i haven't had any reply till now on "MIMO with USRP". its very 
> unfortunate that big people like eric,jonathan and even matt didn't 
> reply to my query. A Board like USRP which started way back in 2005 
> (usrp1..which is mimo capable according to ieee spectrum 2006 ), still 
> dont have any reference code for MIMO. sorry to say, this is really slow 
> pace..work. I Understand I NEED to write and make  a way out  and for 
> that one has to go through mails lists archives. But my investigation 
> requires me to test different flavours of MIMO, i would have been happy 
> if there is some working code on which I NEED TO BUILD MANY FLAVOURS of 
> MIMO without really breaking down my head over writing hardware related 
> configurations program. Forget about C++, i dont think any python 
> reference code exists for MIMO. having so many standards like wimax and 
> LTE which have MIMO implementations, i think gnuradio forum people 
> should seriously think about MIMO implementations. Anyway..no way..out i 
> have started writing code..but i still stand on the word.."this is not 
> the way you support for open source hardware ----> this is to hardware 
> manufacturers"
>
> thanks
> john


John,

Normally I wouldn't respond to a post/flame like this, but in this case 
I'll make an exception.

On response, or lack thereof, to your original email -- there are a 
number of reasons that I did not respond to it.  First, your email was 
sent about 2 days before, and sometimes when you are busy it just takes 
longer than that to get to it.  Second, in addition to the mailing list, 
you sent copies of it to 10 people, at least 8 of whom are also on the 
GNU Radio mailing list.  You then sent it to the mailing list _again_ 
only 5 hours later, resulting in 3 copies of it for all of us.

I can't and don't speak, however, for the other "big" people, but I do 
know they are busy as well.  We all do tend to be more likely to answer 
questions when it looks like the author has followed these guidelines:

        http://gnuradio.org/trac/wiki/ReportingErrors


On the subject of there being no MIMO reference implementation -- you 
need to understand that GNU Radio is a library on which to build your 
applications.  We do have examples in the code which show how to receive 
multiple signals simultaneously, and how to transmit multiple signals 
simultaneously.

We cannot possibly have reference implementations of everything people 
would like to do with GNU Radio.  That is a little like saying that the 
glibc C library should have a reference implementation of a word 
processor.  Also, LTE, WiMAX, and other systems like that are _very_ 
complex, and a reference implementation could take many person-years to 
complete.  Us "big people" only have so many hours in a day.

I can tell you that many people have successfully built MIMO systems 
around GNU Radio and the USRP, so it is possible to do with the tools 
which we have provided.  Some of these implementations were done as part 
of Master's or PhD theses, and some have been done in the commercial 
sphere.  It would be nice if these implementations were contributed back 
to GNU Radio or to CGRAN (or even hosted elsewhere), but we cannot force 
anyone to do this.


As for a slow pace to the work, I would say that this is a volunteer 
project, and people work on what they are interested in.  None of the 
core people are working on a MIMO reference implementation.  Instead, we 
are working on the _core_ itself, so that we can provide the best 
possible platform for _you_ to build your application.


As for hardware support, the first thing to understand is that Eric 
Blossom and Johnathan Corgan do not supply hardware.  They receive no 
direct benefit from your purchase.  They have their own companies, 
(Blossom Research and Corgan Enterprises respectively), and do 
consulting and training around GNU Radio in order to earn a living.

If you do feel that the hardware support has been lacking, then I am 
sorry.  However, I do not promise to include MIMO reference 
implementations or immediate answers to every question about GNU Radio 
with your purchase.  This is a part of the reason why we can provide the 
hardware at the price we do.


Matt