FMCW Radar using FSK Superimposed LFM ramp

Question

While reading the datasheet of the ADF4159 I came upon a section describing how to create an FMCW ramp with FSK superimposed to "enable unambiguous (distance and velocity) multitarget detection". While this sounds like a wonderful radar technique, I do not understand how it works. There does not seem to be a lot of online documentation on it.

Datasheet: https://www.analog.com/media/en/technical-documentation/data-sheets/ADF4159.pdf

Page of the section: 29

I am especially confused with the picture they provide of the frequency domain:

What are the dashed lines supposed represent? Why does each step seem so "uneven" with each other? What exactly is the smaller dimension arrows/line measuring?

I already understand the concepts that drive FMCW radar using a linear sawtooth or triangular ramp, so could anybody explain how LFM-FSK radar works, possibly in a visual way? How exactly does it interact with the target? What does it look like on the frequency and/or time domain? And is this method of measurement worthwhile (in terms of complexity vs functionality) compared to other alternatives?

I simply seek knowledge here, I don't plan to try to detect missiles anytime soon :)

Answer 1

First thing first: Once you start talking about stepped FM waveforms with the goal of achieving LFM-like range resolutions you are introducing a more complex receiver along with other associated disadvantages. Stepped-LFM waveforms require that you process each pulse individually, and thus the receiver must reconfigure itself for every pulse which are at different frequencies. Having said that, you must transmit multiple pulses (think multiple PRIs) in order to reap the range resolution benefits. Contrast this to an LFM system that has the large enough instantaneous bandwidth where it can achieve high range resolution with only a single pulse.

This documentation implies that you use the stepped-FM portions of the waveform to achieve high range resolution and the FSK processing to perform Doppler filtering.

Specifically, with this waveform you can process the strictly stepped-FM portions (the solid black lines) as mentioned before to achieve the range resolution required. You can then pass the FSK portions (dotted lines) through a Doppler filter bank, after appropriate mixing and integration, to yield the target velocity.

As you can see, the receiver for such a system sounds pretty complex! In the past, hardware was not up to par to provide the instantaneous bandwidth and computational power that we have today, so great effort was made to design these complex receivers to achieve the desired performance.

Thankfully, hardware has come a long way even since this datasheet was published and we now have access to high instantaneous bandwidth systems which allow for simpler receiver designs. Modern DAC/ADC and embedded processor technology has enabled multi-GHz bandwith waveform generation and processing. We can easily generate a multi-hundred-MHz LFM pulse and process it digitally via a matched filter using a rather simple receiver design.