Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

UWB (Impulse) radar advantages over Pulsed radar

  1. Jul 24, 2012 #1

    I was wondering what the advantages are of using an Ultra Wide-Band (UWB) radar system (also known as impulse radar) where you transmit an extremely small pulse (sub-nanosecond), and hence a very large bandwidth, over conventional radar that use continuous waves or larger pulse widths.

    I am interested in the case of distance measurement between two UWB transceivers using time-of-flight estimation, as opposed to a probing radar that detects echoes from objects it is probing (like millimeter-wave radar).

    So consider a transmitter transmitting a gaussian pulse with a 300ps pulse width, then a receiver retransmitting it after a fixed, known delay, so the round trip time can be calculated by clocking the time at the reception of the pulse by the original transceiver.

    I know accurate timers exist for getting an accurate distance estimation, even at the fast speed of light, but I don't see what the difference is between transmitting an extremely small pulse with a high bandwidth, as opposed to detecting the leading edge of a longer narrowband pulse, for example if you were pulse-code modulating a narrowband signal.

    UWB radar systems are alot more difficult to make due to the high bandwidth, so they must offer some advantages for them to even exist.

  2. jcsd
  3. Jul 24, 2012 #2
    There already exists distance measuring sets using laser time of flight and it is a lot simpler and accurate also, you aim it at something reflective and it measures the time of flight even for short distances of say 20 feet or less which would be a signal path time of around 40 nanoseconds so the laser must have a pulse width of around 1 nanosecond already which would give an accuracy of less than a foot, and the ones I saw were more like 1/8 th of an inch at 200 odd feet. That would imply a pulse width of a couple hundred picoseconds. Sounds to me like your proposed measurement system would be a lot more complex since the laser measurement system is in a hand held device that looks more like a thermal imaging tool.
  4. Jul 25, 2012 #3
    Thanks, I'm aware of the laser positioning systems out there, but they are unsuitable for all applications, and UWB may offer more advantages in some cases. But I'm assuming the principle is exactly the same, and so how are range and accuracy requirements related to the required pulse width?

    Say your requirement is to measure over a range of 1m to 20m, with a resolution of 1mm (Which I know laser systems can easily acheive - whether thats TOF or otherwise), then what is the maximum your pulse width is allowed to be based on these numbers? (Is there a formula relating them?)

    This is where I'm struggling:

    Minimum round trip time is 2*[1 metre (minimum range)] / [speed of light] = 6.67ns.

    Time resolution required for distance resolution of 1mm = [0.001] / [speed of ight] = 3.33ps. This can be acheived by oversampling with a less accurate/expensive timer.

    So say you have to transmit the entire pulse before you can receive the return pulse, then if your pulse width is 6ns, you will have completely transmitted it before you receive its rising edge (6.67ns later at the minimum distance). Then if you can measure the rising edge with an accuracy of 3.33ps, you will have met your accuracy and range requirements.

    Therefore, it seems to me that the pulse width merely determines the minimum range of operation, while the accuracy is determined by the accuracy of your timer and the amount of oversampling used, which are two different things.

    This is why I don't understand the implication that a smaller pulse width equals a higher accuracy/ distance resolution. I can't see how this is true, what am I missing?

    Considering the case of a UWB radio system, shorter pulsewidth = considerably more noise, so therefore, why not just use longer pulse width, and have less noise if you can still get the required resolution?
  5. Jul 25, 2012 #4
    Look up 'over the horizon radar', a cold war technology. They used pulse widths very long so a signal would go out and it was a low frequency, say 20 mhz, and the PRF was like 1/30th of a second, which gave a range of 186,242/30, about 6000 miles in that case. But if the PRF was faster, the return pulse would come back too many times and interfere with the outgoing pulses, so they had to have proper timing of the PRF (pulse recurrence frequency). The accuracy was related to the pulse width because, say you are looking at an aircraft, say a 747 which is something like the size of a football field. If you want to be able to get a radar visualization of that beast you need to have the pulse width a lot shorter than the size of the craft, if you have a relatively long pulse width the return would be muddled, the front of the craft and the back of the craft would come back together, there wouldn't be a way to tell the difference between the front of the craft and the back, you need to visualize the field front as it leaves the transmitter, it wouldn't matter if it was 100 mhz or 1 billion hertz if the pulse width was too long the front and back of the pulse would come back to the receiver and you can't tell the difference.

    Overall resolution is related to the carrier frequency because you can only at best transmit one cycle and if the frequency is one mhz, a wavelength of 300 meters, even if you could only transmit one cycle you can't get better than 300 meters of resolution with that wavelength. If you send out 300 mhz, a wavelength of 1 meter, that is the max resolution even with one cycle of send pulse. So to get 1 mm resolution you need 300,000 mhz minimum frequency otherwise the front of an object and the back of the object would be sending signals back mixed together and can't recover information smaller than the wavelength of the sending pulse. So 1 mm resolution would require a single cycle pulse of 300 ghz, not sure if you can actually even generate a single cycle of that high a frequency. With light, you are in the hundreds of thousands of ghz so you don't have to send out, say, a single photon, you would send out a signal with many wavelengths together and the resolution would then be the composite size of the entire wavefront, say your sending out light at 1 micron, an IR frequency, and you send out 1000 cycles, then you would get your 1 mm accuracy. The size of the wavefront in that case would be about 1 mm long, which would be a pulse width of 1 picosecond. 1 meter at c is about 1 nanosecond so 1/1000 of a meter, 1 millimeter, would be 1 picosecond.

    Don't know enough about the electronics to say if they can arrange a pulse width of 1 picosecond though. 1 cycle of 1 micron would be about 1 femtosecond long. I'm pretty sure there are laser systems that can give something like that cycle time but they are not going to be suitable for field use, more like laboratory specialty items for probing states of matter and such. There are actually sub femtosecond laser probes out there but you have to use higher frequencies, say 100 nm, where one photon would be 100 attoseconds long in time.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook