How to find distance between usrp and the object reflcting its rf wave using 2901 and labview 2014

[Tawahaa38]

Hi 

I want to find the distance between my usrp and the object from which its 2.1 GHz wave would reflect back that come in its way. please guide me how would i do this in labview.

I am thinking of making a radar that sends a pulse and receive it and find the time thus finding its distance. but the unable to do it in labview. I just need a  guidence that would help me make this labview VI.

Regards,

[muffin.vi]

Hi Tawahaa38,

How far away is the object that will reflect the signal? A TDOA setup is very difficult unless the object is rather far away. 

[Tawahaa38]

the distance would vary from 1 meter to 30 meters. Maximum distance is 30 meters

[muffin.vi]

Tawahaa38,

How are you planning on measuring that distance? The time it takes for the pulse to make it back to the SDR? Or are you measuring something else, like phase? 

[muffin.vi]

And what kind of resolution are you trying to achieve? 

[Tawahaa38]

Yes,  time it takes for the pulse to make it back to the SDR.

time of transmission of wave and time at which wave is received after reflecting these two times would help me in finding distance usinf formula 2S=v*t.
I need some hint how can i measure this time accurately.

[Tawahaa38]

what do you mean by resolution could yo please explain?

what i am doing is I am monitoring the magnitude of the reflected wave and now i need the distance with this magnitude to differentiate that how much magnitude I would get on how much distance.

[muffin.vi]

Tawahaa38,

There is a BIG difference between measuring the time between when a pulse is sent and recieved, and the magnitude of that pulse. 

Let's look at time first. The speed of light is approx 3e8 ms/s. If you are 30m away from an object, that means it will take approximately 2e-7 seconds between when the pulse is sent and when it is received. If you sample at 5 MS/s, you will sample once during the flight time of the signal. However, you will have terrible resolution--+/- 60m! The faster you sample, the better resolution you will have. For instance, if you sample at 300 MS/s, you will have a resolution of +/-2m. In other words, you'll have 30 samples during the flight time of the signal. This is the problem with measuring time of flight to determine distance. You have to sample VERY fast in order to accurately measure the distance. The FPGA on a USRP RIO would help quite a bit here, but a 2901 won't be able to sample anywhere near fast enough to get an accurate measurement. 

Measuring magnitude is a bit different. It's a lot easier to measure the magnitude of a signal because it is continuous. However, a USRP is not a calibrated instrument, so the accuracy of the measurement is not good. 

I am not an expert in this area, so there may be a way to measure distance using something besides time of flight or signal magnitude. But I don't think doing TDOA will work with a 2901. 

[Tawahaa38]

muffin.vi

Thankz for your reply I get your point.

So do have Rio usrp at our university which we can utilize that have a sampling rate upto 200 MS/s.

So could you please help me how can i measure this signal fight time in labview. Is there a specific module that can help me in finding this signal flight time.

[DuncanW.]

What muffin.vi is trying to explain is that you cannot accurately measure the signal flight time to the accuracy you would need to differentiate between an object that is 10 meters away and an object that is 30 meters away without an extremely fast clock. Even at 200 MHz, you will not get very good resolution, which means your precision will be bad. The reason you would use a RIO here is if you need to process signals or act on these signals (control, etc.) very quickly. What you need is a faster clock.

What is your application? If it is possible to use a signal that doesn't travel as quickly, like a sound wave, the flight time would be much easier to measure over such short distances.

Duncan W.

[Tawahaa38]

I am making a long range metal detector that can detect a metal from a maximum of 30 meters.

and i know about this clock problem but i am trying to implement a proof of concept of long range metal detector with bad precision.
now i am thinking of finding the flight of signal by using bpsk modulation.

[DuncanW.]

That's an interesting project. BPSK could help you preserve the quality of your signal, but I don't believe it will allow you to get around the clock problem. Is there a reason you need to know the flight of the signal in order to detect a metal? It seems to me that you can just threshold the magnitude of the return signal.

If you don't care that your flight time will be heavily quantized, then you can use a counter that is hardware-triggered to start counting by the sent signal and hardware-triggered to stop counting by the received signal.