discuss-gnuradio
[Top][All Lists]
Advanced

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

Re: [Discuss-gnuradio] synchronization problem


From: CEL
Subject: Re: [Discuss-gnuradio] synchronization problem
Date: Tue, 27 Mar 2018 09:20:50 +0000

Hi!
On Tue, 2018-03-27 at 16:28 +0900, 김무연 wrote:
>  As you mentioned I used costas loop block connected to usrp sources but it 
> didn't work as i expected

What does that mean? What did you expect, what did it do instead, how
did you parameterize your costas loop? Plots/Graphs!

>  I want to the receiver to get cosine wave form that the transmitters sen

I understood that.

> My professor said to me that because it is not synchronized, the received 
> signal looks like a barcode

What does "look like a barcode" mean? In which visualization?

> I discribe my experiment
> Like you said, I don't send information 
> My goal is beamforing 

Aha!

> transmitters ,2 usrps connected with mimo cable,

So, USRP N210?

>  send the same cosine signal and the receiver ,1 usrp, receives the signal 
> and it sends 1-bit feedback to transmitters 

> The receiver measures the power of the received signal and saves the biggest 
> power 

So, a kind of a max hold envelope detector? For that, frequency
synchronization wouldn't matter at all. A cosine with 1kHz has exactly
the same envelope as one with 3 kHz. The only thing I could imagine if
that if your control (see below) has a bandwidth large relative to the
frequency of your oscillation, you get into trouble, because you're
"following" the power variations of the carrier. But I think we can
savely rule that one out.

> And it compares the biggest power with the power of recent received signal 
> continually 
> If the power of recent received signal is bigger than the biggest power, it 
> sends bit 1 as a feedback 
> if the power of recent received signal is not bigger than the biggest power, 
> it sends bit 0 as feedback 
> If transmitters get bit 1, they change the phase of signal 
> If transmitters get bit 0, they maintain the phase of signal 
> I want to the usrps to find the best phase for beamforming automatically 

That constitutes kind of a binary control loop.
I think you should probably be keeping track of what your phase
adjustment does – you probably just constantly jump around in far too
large steps to ever sensibly converge. Don't forget the beam pattern of
a phased 2-array is symmetric and not unambigous in itself – I don't
think a 1-bit quantized purely differentiating control loop fits your
problem at all, but I'm, to be completely honest, not enough of an
expert for discrete state systems to be able to make qualified
statements on control loop design myself.

Have you, before employing your automated adjustment algorithm,
actually verified your beamforming works as expected, for example by
manually adjusting phase, and looking for the direction of receive
power maxima? Have you designed your beamforming control loop with a
reasonable bandwidth? Why the 1-bit quantization? Seems like an
additional hurdle one would avoid for a start?

Best regards,
Marcus

Attachment: smime.p7s
Description: S/MIME cryptographic signature


reply via email to

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