# In case of using a Doppler radar, do we have to use IQ demodulation to find the direction of a target?

I want to know the way to find the velocity of a target by using a doppler radar.

In summary,

1. send rf signal to a target
2. receive the reflected signal from the target.
3. conduct the IQ demodulation.
4. compute complex FFT to find the doppler frequency.

So,

• If we want to know the direction of a target, or a velocity toward or away from the radar, we must use the IQ demodulation technic to find a negative velocity, right?

If it is right,

• What does The IQ demodulation mean?

The reflected radar signal is just a rf signal.

• Do we need to modulate a radar signal before transmitting it to a target?

Radar designer here: It sounds like you’re talking about pulse-Doppler (PD) radar systems. For PD radars, the process is essentially as you described:

1. Generate a waveform (typically at IF) and then convert to RF
2. Transmit the waveform at RF
3. Receive the waveform at RF, and eventually mix it down to IF.
4. Apply IQ demodulation (for digital receivers a Hilbert transformer is typically used but that’s not an important detail). In this step, the received signal is added with a version that is 90 degrees out of phase with itself multiplied by the square root of -1, or the imaginary number. In general, you’ll have the generic signal s(t) = I(t) +jQ(t), where j is the symbol for the imaginary number. You can accomplish IQ demodulation in analog hardware, or you can use a digital receiver to accomplish it as well when you sample (via a Hilbert transform)
5. Collect the individual pulses and “stack” them in an array
6. Perform pulse compression and Doppler processing. Doppler processing is accomplished by computing the DFT across each pulse.

Since each waveform has phase which is recoverable with IQ data, the cross pulse DFT basically allows you to see the rate of change of that phase information, which is naturally the Doppler frequency. By being complex, yes you can tell whether a target has a negative or positive velocity, which is naturally very useful and pretty crucial for most tracking radars.

As for your question about needing to modulate the waveform: if by modulation you mean something other than a pure sinusoid, yes generally that’s a good idea for pulsed Doppler radars, as you’ll need good pulse compression. LFMs (linear frequency modulated) signals are a popular choice due to their nice properties, though other modulation schemes exist (NLFM, PSK, FSK, etc). If you just use a pure sinusoid, you effectively get no pulse compression, and you’ll have a pretty poor estimate of where targets may be, and the performance would be awful if you have several closely spaced targets.

• Thank you for your answer. In step 4, IQ demodulation is conducted to find the negative or positive velocity of a target, right? Because we could know the directional velocity of a target toward or away from a radar only if the complex DFT technic is computed. I am not clear about that..... – Sinecosine Feb 20 '19 at 4:31
• Yes, for pulse-Doppler radar IQ demodulation allows you to preserve the phase of the signal. By knowing the phase, the frequency spectrum is unique rather then conjugate symmetric. This allows you to differentiate between positive and negative frequencies without ambiguity. As another user pointed out though, a DFT is inherently complex. It’s performing a DFT on an IQ signal that gives you the unique frequency content you’re looking for – matthewjpollard Feb 20 '19 at 11:18

If we want to know the direction of a target, or a velocity toward or away from the radar, we must use the IQ demodulation technic to find a negative velocity, right?

No, whether you do that on the original passband signal, or on an intermediate frequency or in baseband after IQ demodulation doesn't matter, mathematically all these three are equivalent.

It's that one for of processing is often easier than the other. For pure Doppler radars, you typically don't do IQ demodulation but mix RX with the TX signal; but that's an excursion into Radar basics that would lead us out of scope quickly.

What does The IQ demodulation mean?

There's multiple names for this: Quadrature mixer, IQ demodulator. This is really radio engineering basics, so I'd recommend finding an entry-level textbook for a digital communications course.

Basically, you take the real-valued, hence 0-symmetric RF spectrum and mix it down to complex baseband, i.e. an asymmetric signal concentrated around 0 Hz.

Do we need to modulate a radar signal before transmitting it to a target?

There's no such thing as an unmodulated signal, so not quite sure what you mean.

• Thank you for your answer and more clearly editing my question, Actually, I know a little about fundamentals of the IQ, but I am not clear that how doppler radar could extract the negative velocity of a target. I heard that we can find the negative velocity of a target only if the IQ signal should be acquired to calculate the COMPLEX DFT. Because only the COMPLEX DFT could find the directional velocity toward or away from a radar. After the COMPLEX DFT, we could use the spectrum amplitude values between N/2 and N point to find the negative velocity of the target. Thank you... – Sinecosine Feb 20 '19 at 0:14
• The DFT is complex. There's no "special" complex DFT. Negative relative velocity simply results in a Doppler frequency shift with an opposite signal, so again, this is just entry-level radio communications textbook stuff. – Marcus Müller Feb 20 '19 at 10:38
• I meant the real DFT and the complex DFT. You might know it. Anyway, thank you for your comment. – Sinecosine Feb 20 '19 at 13:04
• There's no real DFT. All DFTs are complex. – Marcus Müller Feb 20 '19 at 13:28
• edaboard.com/… – Sinecosine Feb 20 '19 at 13:44