UPGRADE YOUR BROWSER

We have detected your current browser version is not the latest one. Xilinx.com uses the latest web technologies to bring you the best online experience possible. Please upgrade to a Xilinx.com supported browser:Chrome, Firefox, Internet Explorer 11, Safari. Thank you!

cancel
Showing results for 
Search instead for 
Did you mean: 
Visitor sravan@243
Visitor
4,160 Views
Registered: ‎12-21-2011

Multiple delays identification

hi,

I need to find multiple delays between two signals.When i use cross correlation, i could identify the single delay between the signals but not multiple delays.While searching for the solution, i came to know that multiple delays can be identified by time-frequency analysis, for that we need to go for wigner distribuion(i.e, FFT of the autocorrelation) and then time-frequency cross correlation.I could not perform the time frequency cross correlation. Can anyone tell me how to analyse a signal in time-frequency domain, how to perform Time-Frequency cross correlation to identify the multiple delays

 

regards,

sravan

0 Kudos
4 Replies
Adventurer
Adventurer
4,135 Views
Registered: ‎01-28-2008

Re: Multiple delays identification

Hi Sravan, Do you mean the signal is a sum of a signal which has been delayed by different delay amounts? Sanka.
0 Kudos
Visitor sravan@243
Visitor
4,130 Views
Registered: ‎12-21-2011

Re: Multiple delays identification

hi sanka, the multiple delayed signal is generated by summing different signals thar are delayed by different delays.

I'm attaching the simulation model of the same

Now i need to find all the 3 delays between the multiple delayed signal and the reference signal

post.JPG
0 Kudos
Adventurer
Adventurer
4,114 Views
Registered: ‎01-28-2008

Re: Multiple delays identification

Hi Sravan,

 

I think you will be able to work out the 3 delays by sending a chirp pulse as you indicated in the diagram and performing auto correlation on the summed signal. This is because you don't have any frequency shifts on your individual delay sources. If you had delay + frequency offset on each component, then I would do the following:

  • send multiple pulses with a constant pulse width and pulse repetition interval (say 64)
  • compute auto correlation on each received pulse and arrange them as a matrix
  • compute FFT along the transpose direction of the matrix

You will find 3 peaks on this 3D map and you will be able to work out the frequency offset and the delay corresponding to each delay-frequency source.

 

Hope this helps.

 

Sanka 

0 Kudos
Visitor sravan@243
Visitor
4,110 Views
Registered: ‎12-21-2011

Re: Multiple delays identification

Hi sanka,

I have only time delay between the signals no frequency shifts. To identify single delay, i have performed cross correlation between reference and delayed signal and i got the result. The problem is to identify the multiple delays, i need to do time-frequency distribution of the signals(i.e, wigner distribution) and perform time-frequency cross correlation between the wigner distributions of the reference and delayed signals. As wigner distribution is nothing but FFT of Auto correlation of the signal, i performed the same and obtained both wigner distributions. Now the question is how to perform time-frequency cross correlation to identify the multiple delays.

I'm attaching my matlab model

post1.JPG
0 Kudos