Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Problem with correlation between two signals

Status
Not open for further replies.
jakjoud

jakjoud

Full Member level 2
Joined
Feb 6, 2005
Messages
129
Helped
10
Reputation
20
Reaction score
1
Trophy points
1,298
Location
Marrakech
Activity points
954
Hi everybody,
I've got a serious problem, here is it: I have two signals u and v, I wanna measure the lateness ( time to fly) Δt, where v(t)=exp(jβ)u(t+Δt), so normally I've just to make a correlation between them and then search the max of this function f and f(Δt) = fmax, but it's not true I don't know why? am I wrong
So please could you help me!
 

Re: correlation problem

Hi,
exp(jβ) is just a constant and the complex correlation of v(t) and u(t) should give Δt. I find no problem in cocept but the issue may be with the implementation of correlation.
brmadhukar
 

Re: correlation problem

Correlation function may have set of local extremums each of them is less than absolytely extremum. You should look for that extremum but not locals.
 

correlation problem

in fact the problem could not be with the extremum, 'cause I'm working with pulses which means that the signals are finite.
And the amplemantation of correlation does'nt present a problem too, I'm working on MATLAB so it's just a function to use xcorr;
here is another real problem, in correlation of the two signals may I do xcorr(u,v) or xcorr(u,conj(v))?
thanks a lot
 

Re: correlation problem

Hi, jakjoud

I'm not sure what you mean.
You said
"search the max of this function f and f(Δt) = fmax, but it's not true"

Maybe you can seperate input to I/Q, and make correlation by oneself. Then take a (abs(I+Q*j))^2. You can see the peak.

At last, I should remind you, the peak appear position is different base on using "conv" or "xcorr".
"xcorr" will extend the shorter vetor length to match longer one first.
"conv" makes the result length to "input1 + input2 +1". You still can use "conv" to implement "xcorr". ^_^


dreamshower
 

correlation problem

hi,
for me it's all over, it was just a problem of periode, it means that I must write that Δtr=Δtm-n*T where Δtr is the Δt wanted Δtm is the Δt measured and T is the number of samples in signals n is a factor appropriated, ok thanks for everybody.

let your heart still be
 

Re: correlation problem

i need book about it
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top