radar signal processing

[OS]

-

 

[Menelaos]

I'll give it a try, but it's not easy to explain since i'm translating from dutch and even in dutch it's hard to explain .

Typically a radar sends out pulses, which are reflected by an object and recieved by the radar again. If you look at the signals in time, you would get a big pulse at t=0 and a (much) smaller pulse at t=Dt. Without any noise/distortion you could just measure the time Dt and calculate the distance. In practice of course there is a lot of noise on the recieved signal, but both pulses are still in there, even if there completely "covered" by the noise.
The important thing about noise is that it is fully random (general assumption). This comes in handy when you want to extract the pulses from the signal. This is done by using the autocorrelation - function (translation?) of the signal. It takes two points in time on a signal and compares the values at those points in time. Noise is assumed to have no correlation.
So the only correlation inside your signal comes from the pulses transmitted by the radar and reflected of the object. Altough the second pulse will be very small, it is still a pulse, with a shape similar to the original pulse. When you calculate the autocorrelation of the signal, you are not bothered by the noise (because it has no correlation). This autocorrelation will be maxed for a time - shift of Dt, which will be exactly the time between pulses.

Hope this makes sence,

Menel.

 

[highwire]

Yeh, lots of circuitry to pull radar targets out of the mud. In fact, if one eyeballs the un-processed signal on a radar PPI or A scan, it is not dificult to disern sigs well below the noise level. The target has persistence at a certain range position, the noise does not. The eye and the brain do the processing.

If anyone needs to be impressed by techniques though, it is hard to beat the statistical processing that an "ordinary" GPS uses. That method allows the sig to be detected and used from very very deep in the noise.

Now i've got to dig up that info again. BTW, A GPS receiver IS a passive radar. RADAR = RAdio Ranging And Detecting.

 

[Mark R]

I've wondered for quite some time exactly how a GPS unit manages to detect the signals, which are a) so weak [similar strength to a cell phone at 1,000 miles] and b) all transmitted, not just on the same band, but on the same frequency.

Do you have a link or reference to such techniques?

 

[OS]

-

 

[highwire]

Mark R -
Yes. The center of much of this is the "Kalman filter". It is a big conceptual thing, a "must know" in sig processing and control. It is implemented in the GPS and apparently, nearly every other high tech control or signal processing system now. I found some links, but not the paper I read a few years ago.

Much of this is not circuitry now, but digital signal processing using algorithms based on Kalman and others.

Kalman page at UNC

Chapter 1 of Stochastic Models, Estimation, and Control

And a page of GPS tech links

BTW, stochastic (steh ka' stick) is greek for 'good at guessing'. But I didn't want to guess, so I looked it up.

Some other GPS magic is 'pseudo random coding'.

 

[nau_baig@yahoo.]

how can i model an EM signal transmitted in a particular direction in MATLAB