"Radar Detector" is a rather general descriptor. If I understand your intended application, you want a radar speed detector. It is rather simple to accomplish radar-doppler speed detection, once you have a doppler-radar oscillator/detector module - not as difficult to come by as you might think. These are usually die-cast aluminum modules with a small horn antenna and oscillator cavity with a gunn-diode oscillator and a detector diode mounted in precise locations in the cavity. The beauty of these is that all the RF work is done for you, and all you have to do is provide precise power to the module and processing for the audio-frequency doppler output. Once you have the module, building the speed detector is well within the capabilities of students with a modest knowledge of electronic design and construction or with a helpful instructor.
We built a radar speed detector more than three decades ago to measure the speed of trains and autos during passby noise measurements. We purchased the modules from Amperex for about $60 or $80 (I don't remember exactly, but it was well below $100 per unit.
The tricky part is processing the doppler speed signal in a reliable manner. The output of the module is the IF (Intermediate Frequency) resulting from the built-in detector diode mixing the strong output from the gunn-oscillator an inch or so away with the much weaker (-100 dB or smaller?), doppler-shifted return signal from the moving target. Separation of the audio-frequency speed signal from the microwave frequency is simple using an RC filter due to the extremely large ratio between the two. This signal is handed to you on a platter from the module, but it is generally not very clean due to multiple signal reflectors and large fluctuations in return signal strength. For street traffic as the target, the signal resembles a heavily over-modulated AM carrier with lots of phase shifting of the carrier. From this, you have to extract the dominant frequency that indicates the speed of the dominant target within view of the antenna.
We fabricated a ten-inch long, rectangular horn antenna out of brass sheet and angles soldered to an X-band flange that matched the module antenna flange and trashed the 1-inch long, die-cast antenna. This gave about 20 dB gain and greatly improved directionality. You may not have to do that if you are close to a known target.
The basic modules provide no AGC, limiting, or other signal processing, since that is up to the designer of the end product. We tried implementing AGC using the LM170 (I think), but never got it working as intended. I think it would be much easier to implement AGC with chips available today. We ended up just using a limiting amplifier, a constant-charge dispensing frequency-to-voltage converter, and an analog meter readout. Works great! We can detect the speed of a subway train hundreds of feet away, out of sight around the bend of a tunnel.
In our ignorance, tuning up the device was a pain and we were initially discouraged with the very low sensitivity we were getting (being only able to detect an isolated car very nearby). But, as we twiddled the tuning plug (a screw sticking into the cavity), suddenly we were getting solid speed indications on autos a couple of blocks away.
Our project building a vehicle speed detector was several decades ago, and I know nothing about what is available today, or how much support you will get from manufacturers. However, I would expect modules to be much cheaper today with all the radar intrusion detectors on the market. These use the same type of modular oscillator/detector modules and they also use the audio-frequency doppler signal to sense motion of an intruder. I think the modules in radar intrusion detectors would be ideal for speed detection under controlled conditions. The signal processing in intrusion detectors will be useless for speed detection, since the actual frequency of the IF signal is of no value for the intrusion detection function. All they want to do is sense amplitude above a threshold and close a time-delay relay. If you shop around, you may be able to buy a handful of used radar intrusion detectors for a song. Just be sure they are not passive IR motion sensors.
The frequency of the IF output is directly proportional to the speed of the target, however there is no direction information. This is because the detector responds identically to return signals above or below the oscillator frequency. You can calculate the IF frequency from the microwave wavelength and the speed of the target. My vague recollection is that, for the X-band modules we were using (the frequency is built in to the device by the dimensions of the cavity), the signal was about 30 Hz per MPH, but that's a guess. It is, however, in that ballpark and is, therefore, easy to deal with using ordinary audio-frequency construction techniques.
This could be a fascinating and very educational project for students interested in electronics, and it would be well within their capabilities.
Go for it!
awright