Theory

Radar pulsing causes a phenomenon called aliasing, which occurs when the Doppler frequency created by reflector motion exceeds the pulse repetition frequency (PRF).^[1]

This concept is related to range ambiguity resolution.

Doppler frequency shift is introduced onto reflected signals used by radar.

Operation

When the Doppler frequency shift exceeds the PRF, the frequency is reduced. This limitation is called the Nyquist sampling rate. This introduces a modulo operation onto the apparent frequency of the reflected signal.

The ambiguous velocity is as follows.

${\displaystyle Ambiguous\ Velocity=-0.5\left({\frac {Doppler\ Frequency\times C}{Transmit\ Frequency}}\right)}$

Frequency is folded for high speed targets where radial velocity produces a frequency shift above the Nyquist frequency. The true speed of the target may be folded by a modulo operation produced by the sampling process.

${\displaystyle True\ Velocity=Ambiguous\ Velocity+0.5\times N\left({\frac {PRF\times C}{Transmit\ Frequency}}\right)}$

${\displaystyle N=Integer\ Between\pm \left({\frac {0.5\times Bandwidth}{PRF}}\right)}$

The Nyquist frequency will also change when the PRF is changed.

This is explained best using an example with 2 different PRF, although real systems use a different method.

In the example, PRF A can detect true speed up to 600MPH and PRF B can detect true speed up to 500MPH.

The apparent speed for PRF A falls in the 200MPH filter, and the apparent speed for PRF B falls in the 400MPH filter. This combination places the true target speed at 1,400MPH (2x6+2 or 2x5+4). This can be seen graphically when range intervals are stacked end-to-end as shown below.

"A" represents target speed possibilities for PRF A, and "B" represents target speed possibilities for PRF B.

This frequency ambiguity resolution signal processing technique determines true velocity.

Ambiguity resolution is typically implemented with a look-up table. This could also be implemented as a convolution function, where the spectrum amplitudes from one PRF are applied to spectrum samples from the other PRF as a sliding window function.^[2]