From IGEP - ISEE Wiki
|A newer version of this information may be available: New software for IGEP RADAR LAMBDA|
This is the 1/3 chapter of the Getting Started with IGEP RADAR LAMBDA Tutorial Guide. In this first chapter, we will learn:
- Connect the radar to a PC via Ethernet.
- Boot the board with the default firmware.
- Make first measurements using web server demo application.
Upon completion, you will be ready to continue with chapter 2/3 that explains more advanced tasks.
- 1 Overview
- 2 Requirements
- 3 Getting started
- 4 Make your first measurement
- 5 Compare Measurements
- 6 What is "Real Time" section
- 7 What is "Technical" section
- 8 What is "Files" section
- 9 Managing Files
- 10 Next Steps
- 11 Radar Applications: Examples
- 12 Calibration
In this tutorial we are going to use the following materials:
- IGEP RADAR LAMBDA with its power supply
- Network cable
- PC with keyboard, mouse and monitor
Please follow the next steps as indicated. You could change later the way you use the radar but it is very important you follow these instructions at least once for the first time. With this exercise you will start to understand how the radar works and most of your questions will be solved in a few minutes.
Place the Radar on a stable surface
Place the radar over your desk and oriented with the ORION board (antenna) facing a wall of your office with nothing in between. A distance of about 3 meters between the radar and the wall would be good to make the first measurement.
You may now plug the Ethernet cable into the 10/100 Ethernet jack of the IGEPv2 board of the radar to get network access. The default firmware configures the Ethernet device with static ip address (192.168.2.232). Start your PC and plug the Ethernet cable to it.
Now, before you apply the source power to the radar, ensure the equipment is fixed over your desk and the cables do not move the radar. Once the radar is fixed facing to the wall, you can power it through any of its power connectors: J200 on IGEPv2 board or JP1 on ORION board. Remember you can power the radar through any of both connectors but never apply power to both at the same time. Here there is a diagram of an ideal measurement over a wall:
Wait few minutes
Once you apply power, please wait a few minutes while IGEPv2 board boots with the default firmware flashed on the internal memory. You will see LEDs from IGPv2 board blinking until system has booted, then you will see a fixed lighted green LED.
Run Web Browser on PC
Run a web browser on your PC with the ability to work with gnuplot4.6. We recommend the use of "Mozzilla FireFox" web browser that you can install and download for free at: http://www.mozilla.org/en-US/
Connect to Radar IP
Put the radar IP 192.168.2.232 on the web browser. If the radar has already started you will see the Home page of the web radar application demo, if not, please be patient and wait a little bit more until the radar starts. You will see a home page similar to this one:
Do not worry if your home page differs from the picture, there are some different format versions.
Make your first measurement
To see the time domain IF radar signal in pseudo-real time, you can click on any of the yellow marked icons: You should see a wave similar to the picture, this is a dominant sinus.
To see the frequency domain IF radar signal in pseudo-real time, you can click on any of the blue marked icons: You should see a signal similar to the picture, this is a big peak in the first samples.
We will learn later the meanning of these graphics.
Be aware the application demo is designed to work on an environment with only a single target scenario so it is very probably you may see quite different signals than the presented here due to multiple reflections in your room, but this is normal, and this is the power of the radar to detect multiple objects at the same time, we encourage you to develop your own application and process the IF signal information to detect not only one but several objects. See next guides if you want to learn more about this. To be sure your radar works well, we recommend you find an adequate environment without multiple reflections before to make any wrong conclusions.
What is Time domain signal
The time domain signal window shows the 2048 samples captured of the IF radar signal on the last measurement.
Y axis represents the signal voltage.
X axis is the sample number.
In the picture a measurement using the default modulation over a wall about 3m distance from radar is shown. Considering that the time to get 2048 samples with the default modulation is 3,471ms and we can see about 16 periods of the sinusoidal signal this means that the dominant sinus shown below is about 4,609KHz frequency. The frequency of the IF radar signal is proportional to the distance from the radar to the target (in this case the wall).
We will learn later on next chapters how to know the sampling rate and time to get 2048 samples.
Why I am getting this weird signal? It has no sense, my Radar does not works!
No. Please, be patient, this is completely normal, your radar works well. If you are getting a signal like this or even still more strange this means that you are detecting different objects. This signal is the result of adding different sinus waves with different frequencies resulting on an apparently noisy signal but in fact is a typical IF radar signal on a multitarget environment. Our main objective now is try to get a clear signal from a single target to start learning with the radar, so please try to slightly change the radar orientation in order to get only a single reflection, or if you are on a small room, try to go to a better scenario, on a wider room with few objects or simply outside close to a wall.
Basic FMCW Radar Theory
You can check chapter 9 of the user manual to get more in detail about radar theory. Basically, a Radar with continuous wave linear frequency modulation (FMCW) follow the next equation:
This equation gives the relation between the target distance and its corresponding IF frequency (hzpm) where:
hzpm: "Herz per meter" is the ratio between IF signal frequency (in Herz) and target distance from radar (in meters)
BW: RF output frequency bandwidth in Hz
V: media electromagnetic wave propagation speed (3*108m/s by default).
T: FM linear Modulation sweep time in seconds.
In our example, as the default modulation has the next paramenters:
the resulting hzpm is 1271Hz/m
So, if the measured IF radar signal frequency is 4,609KHz it means the target is at a distance of 3,6m from the radar.
Considering the radar has an offset of about 0,35m (this is due to the microwave signal has this electrical lenght internally on the equipment before to reach the antenna) the real distance results in 3,25m
What is Frequency domain signal
The frequency domain signal is the result of applying the FFT to one ramp modulation interval of the 2048 sampled IF radar signal captured samples.
The radar application takes the samples of one single ramp and completes the rest with zeros (padding) completing 2048 samples, and it performs the FFT over these 2048 samples, resulting on the spectrum of the signal with 2048 points resolution.
Following our example, the above picture shows the FFT of the measurement. Y axis represents signal strenght and X axis represents the FFT sample. Note that only the 300 first samples are shown.
We can see a dominant peak centered about sample number 16. This peak should correspond to the measured dominant sinus of 4,609KHz on time domain graphic.
The sampling rate used by the equipment with the default modulation is 590Ksps. This means that each sample of the FFT represents 590000/2047=288,2Hz.
If the peak is centered about sample 16 it means the frequency represented is 16*288,2Hz=4,611KHz, thus corresponding with accuracy with the calculations made on time domain.
We can see other small peaks at about samples 40 and 50. These peaks are due to other reflections on the room at a longer distance. So here an example of the possibilities of the radar to detect several targets at different distances. But by the moment is better if we pay our attention on a single target scenario to keep learning about the radar signal and later we will be able to analyse more complex scenarios.
Now you can move the radar and make measurements over the wall at three different distances. For example, make a measurement with the radar at 2m distance and another measurement with the radar at 5m distance. You should get similar results as the following for 2m, 3m and 5m in order:
It can be clearly shown how the IF radar signal increases frequency when the target gets more and more distance from the equipment.
What is "Real Time" section
Now you can also make use of this web application demo feature.
If you click on "Real Time" icon you will be able to continuously monitor the speed and the distance of a single target as much time as you need.
In case of a wall at 4.9m distance you should get something like this:
Note that in this case you get always the same position (4.9meters) and speed (0Km/h) as our target is a no moving object (wall).
You have a real time bar indicating the target range (orange) and another real time bar indicating the speed (blue).
There is also a graphic where:
Left Y axis: is the target range in meters.
Right Y axis: is the target speed in Km/h
X axis is time
Each point of the graph is one combined position-speed measurement, position in orange color and speed in blue color.
You can see on next chapter examples of real traffic measurements where you can see how the radar track the vehicles measuring its position and distance.
What is "Technical" section
This is the most interesting feature of the web application demo. By clicking on "Technical" icon you will get access to the heart of the radar. You will be able to make measurements with the modulation you want, by simply prompting the parameters on the rectangles.
There are three sections:
SECTION-1: RADAR CALIBRATION
This feature is for advanced measurements and we do not recommend to use on this chapter. It will be explained later.
SECTION-2: RADAR MEASUREMENT
This feature is interesting if you want to use the web application demo as a console where you can configure the radar with different modulations and get the results prompted on the screen or saved on a text file.
You can see that the default parameters are: -F 3C0000 -s 50 -i 512 -S 80 -m 3. This is the same default modulation used on the Home Menu icons f(t) and F(jw). Now here you can remove these parameters and put what you want.
For example, put in this section the next commands:
-T 800 -m 20
and click on "RUN" button
-T 800 means you want a modulation with 0,8ms sweep time
-m 20 means that you want to make 20 consecutive measurements
You can check a complete explanation of each parameter on the user manual of the radar.
The result obtained over a wall at 4.8m distance is like this:
The screen shows the result of each of the 20 measurements done on 20 rows, one row per measurement:
1st column shows the measured target position (4.8m)
2nd column shows the measured target speed (in this case 0Km/h as the wall is no moving)
3rd column shows a value proportional to the measured signal strength.
4th column indicates the delta time between each measurement. In this example the radar did each measurement in only 7ms. This time includes the time to program the modulator, perform the ramp modulation, capture the IF radar signal with the ADC, process the data, and show the result.
Finally, the last column indicates the name of the captured file. IMPORTANT: This file will be generated and stored ONLY if you add the parameter -w
-w means that you want to write all the measurements on a file (there will be generated 20 files, one for each measurement, and each one with 2048 samples)
SECTION-3: RADAR CONTINUOUS MODE
This section is used to get the same graphics as the obtained on the Home Menu icons f(t) and F(jw), but in this case we can remove the default modulation and put the parameters we want. By checking on TIME or FFT we will obtain the time domain or the frequency domain graphic.
For example, remove the default modulation and put:
This will program the modulator with a 1ms modulation ramp. If you click the "RUN" button you will see in pseudo-real time the time domain or the frequency domain graph of the IF radar signal.
What is "Files" section
This is a feature of the web application demo that will help you to manage the captured files in your measurements. By clicking on "Files" icon you will be able to check all the captured files and you will have the possibility to:
-Browse: represent a time domain graphic of its contents.
-Figure. the same as before
-Download: Check the data contained on this file, this is a column with the 2048 samples captured by the ADC of the IF radar signal. You can select and copy to a text file for post-processing
-Remove: click here if you want to remove this file
Check IGEP RADAR LAMBDA user manual if you want to know more about the meanning of the file names.
IGEP RADAR LAMBDA offers several ways to manage its files. Here we present some of the most interesting but you can use others if you prefer:
Files with web demo
You can manage files with the web demo application by entering on its "Files" section.
If you click on Download you will see something like this:
This is the list of 2048 samples captured on this measurement.
You can use the mouse and select all or only one part of the data, copy and paste to a text file for later processing.
You can also click on "Edit" section of your browswer, then click on "Select All", then click on "Copy" and then paste all the data into a text file.
And there is a third option, by clicking on "File" section of your browser and simply "Save as" text file.
Files using WinSCP
You have a second option to manage files by using the WinSCPfree application.
Once installed you only have to connect to the radar and start managing files as with other explorer applications.You can edit, copy, paste, remove and rename files on an easy way.
Files using Linux
You have a third option to manage radar files.
You can do it using the console and using the standard Linux commands. Here are some useful links and basic commands:
cd to change directory path
rm to delete a file
mkdir to make a directory
cp to copy a file
Capture measurement results
Linux is an interesting environment to generate files of radar measurement results. Here we present some examples on how to save measurement results into a file:
You want to capture 5 consecutive measurement results into a file named measure5.txt. then you must prompt the next command line:
./radar -m 5 >measure5.txt
NOTE: Remember you must be in the "radar" directory and you must execute "radar_init.sh" at least once before start doing measurements.
Now you can check the measurement results in the generated measure5.txt file, and you will see something like this:
Position Speed Level Dtime filename
[m] +/-0.5 [Km/h] +/-3 . [ms] .
3.9 1 417 27 V70D99999A00.
3.9 1 414 8 V70D99999A01.
3.9 1 422 9 V70D99999A02.
3.9 1 423 9 V70D99999A03.
3.9 1 414 9 V70D99999A04.
Where each row shows the result of each of the 5 consecutive measurements, in this particular case it was measured a fixed target at a 3.9 meters distance from the radar.
NOTE that it takes about 9ms time to make each measurement.
NOTE also that the file names with the ADC captured data are generated but the files are not saved. They are only saved if you add the command -w
You want to capture 20 consecutive measurement results into a file named measure20.txt. In addition you want to use a modulation sweep time of 1,15ms and you also want to generate a file of the ADC captured data for each measurement. Then, in this case, you must prompt the next command line:
./radar -T 1150 -m 20 -w >measure20.txt
Now you can check the measurement results in the generated measure20.txt file, and you will see something like this:
Position Speed Level Dtime filename
[m] +/-0.5 [Km/h] +/-3 . [ms] .
4.9 -1 866 61 V70D99999A00.
4.9 -1 861 106 V70D99999A01.
4.9 -1 839 97 V70D99999A02.
4.9 -1 846 91 V70D99999A03.
4.9 -1 833 92 V70D99999A04.
4.9 -1 840 141 V70D99999A05.
4.9 -1 864 92 V70D99999A06.
4.9 -1 847 116 V70D99999A07.
4.9 -1 836 92 V70D99999A08.
4.9 -0 816 220 V70D99999A09.
4.9 -1 818 92 V70D99999A10.
4.9 -1 825 139 V70D99999A11.
4.9 -1 820 91 V70D99999A12.
4.9 -1 825 91 V70D99999A13.
4.9 0 800 166 V70D99999A14.
4.9 0 802 90 V70D99999A15.
4.9 0 802 92 V70D99999A16.
4.9 0 807 142 V70D99999A17.
4.9 -0 805 90 V70D99999A18.
4.9 -0 810 110 V70D99999A19.
Each of the 20 rows of the measure20.txt file shows the measurement result in range and speed, in this particular case it was measured a fixed target placed at 4.9 meters distance from the radar.
NOTE that it takes about 100ms time to make each measurement. This is more than previous example due to now the equipment has to write one file after each measurement.
NOTE also that the file names with the ADC captured data are now generated and saved because we have used parameter -w. You should find the 20 generated files in "radar" directory. If you look at one of these files you will see something like this:
completing the 2048 samples of each measurement capture
The ADC captured files can be used to make your own data processing algorithms.
You want to capture 7 consecutive measurement results into a file named measure7.txt. In addition you also want to use a modulation sweep time of 0,9ms and generate a file of the ADC captured data for each measurement. In order to identify your measurements you want to name these files with 12345. Then, in this case, you must prompt the next command line:
./radar -T 900 -D 12345 -m 7 -w >measure7.txt
Now you can check the measurement results in the generated measure7.txt file. You will see 7 rows with the position and speed measured:
Position Speed Level Dtime filename
[m] +/-0.5 [Km/h] +/-3 . [ms] .
4.9 -1 632 48 V58D12345A00.
4.9 -1 630 92 V58D12345A01.
4.9 -1 629 91 V58D12345A02.
4.9 -0 645 92 V58D12345A03.
4.9 1 619 91 V58D12345A04.
4.9 1 587 149 V58D12345A05.
4.8 -0 611 91 V58D12345A06.
If you look in the "radar" directory you will find the 7 data captured files containing the 2048 samples measured by the ADC on each measurement, and these files include the desired extension name: V58D12345A00.txt, V58D12345A01.txt, etc.
The first three digits of the file name are automatically generated depending on the modulation sweep time. You can check the user manual fore more details.
Congratulations! You have made your first measurements and started to understand how the radar works. You also know most of the features of the radar application demo.
Now you are able to do the next steps with the radar trying to adapt it to your specific application.
You can check if your application is similar to any of those related below and have a first approach on how you should configure your IGEP RADAR LAMBDA and which is the set up you should use to make your measurements.
Radar Applications: Examples
This section will show you some real measurements and set-ups where IGEP RADAR LAMBDA can be successfuly used. We are accumulating experiences and we will be updating our results obtained in different field of applications.
Security and Traffic applications show how to configure the radar in order to obtain good results. Each particular application need to configure the radar with the most optimum parameters, we show you here some of the possible configurations that can help you to start with a reference point for your particular needs.
We also refer to other radar applications that you may be interested to explore: Automotive, Speed displays, UAV altimeters, Tank Level Gauging, Vibration, Smart Cities and Sports.
Security is one of the field applications of radar technology. IGEP RADAR LAMBDA has been used for these purposes, we want to show you here some of the results.
Measurements of a person approaching by walking slowly the radar were made during 7 seconds, capturing 1.000 measurement files.
The modulation used was: -T 1600 -m 1000 -w
This modulation means:
-T 1600 ===> program modulator with a 1,6ms modulation sweep time
-m 1000 ===> make 1000 consecutive measurements
-w write ===> and save the captured data into text files
The 1.000 measurement files were captured with the radar and processed later with matlab on a PC.
FFT was made for each measurement obtaining the next 2D plot with matlab:
2D plot of 1.000 consecutive frequency domain radar signals detecting a person
Y axis: represents twice the distance
X axis: represents the measurementfile number (so it can be considered as time)
It can be noted that the person is detected from about 18m in the first measurements and is moving towards the radar until 14m distance in the last measurements.
The next is a 3D plot of the same data, where Z axis is the FFT module level:
3D plot of 1.000 consecutive "radar images" of a person
It can be shown that the person can be clearly tracked over the signal noise level (blue color).
20m range can be acceptable in most security systems but it is possible to apply CFAR processing techniques in order to increase the range over 100m. In addition, some more directive antennas could be developed to improve this parameter, and also developing a coherent system. Please contact radar support if you are interested on a development for an specific application.
Radar Traffic Control Systems is another interesting field of application for radar.
IGEP RADAR LAMBDA has been used on this kind of applications, we would like to show you some of the measurements performed.
The next picture is a graph obtained making use of "Real Time" feature of the web radar application demo. It was obtained placing the radar on on a tripod close to a road on one side, with the radar oriented on a 15 to 30 meters field of view, as shown on below pictures:
Real measurements of 4 vehicles tracked in range (orange points) and speed (blue points)
Left Y Axis: Range in meters
Right Y Axis: speed in Km/h
X Axis: Time (hh:mm:ss)
The first vehicle was tracked from 15 to 25m at a speed of 50Km/h
The second vehicle was tracked from 18 to 33m at a speed of 59Km/h
A third vehicle was tracked from 20 to 28m at a speed of 52Km/h
Finally the fourth vehicle was tracked from 12 to 29m at a speed of 40Km/h
The bars indicate the range and speed in real time while the vehicles are being detected and cannot be shown in this picture but the graph reflects the history of the measurements.
Set up used to make the measurements: Radar fixed on a tripod oriented on one side of the road
The modulation used to make this measurement was:
-T 5000 -l 80 -r
-r ===> it means we want to make consecutive measurements without limitation
-T 5000 ===> we apply a modulation sweep time of 5ms
-l 80 ===> this is the signal level threshold. If the radar detects a signal over this level it will start making measurements until level returns below this threshold, when the radar will stop taking measurements. Each time it detects a signal over this threshold it will present the results. This parameter is very important in order we obtain only desired signals and not noise that would provide unreadable results.
There is not a fixed rule to obtain this threshold value because it depends on a lot of things, mainly of the modulation applied, the environment and the type of measurement, so the best is to make first some trials to adjust this value on an empirical way until the system works properly in your specific application and you obtain readable results.
You have to put this modulation parameters on the command block of the "Real Time" window and press enter.
If you want to make faster measurements there is a second procedure to follow:
open the console and execute these instructions:
./radar -r -T 5000 -x 3000 -X 90000 -l 80 -H
radar.init ===> it is necessary only the first time in order to initialise parameters of the radar application demo.
-H ===> this parameter opens a socket in order the web server application can go faster reading results of the measurements.
-x 3000 ===> this is a limitation on the processing range, in order to avoid false measurements due to close objects, closer than 3m.
-X 90000 ===> the same as before but applying for far objects, beyond 90m.
Then, you must open the socket on the web server by prompting this address:
Now the radar is ready to make faster measurements and get more than one measurement per vehicle.
Note that each application needs to optimise its own parameters. You can tune the parameters in order to adapt better to your application. For example you may be will get better results if you use a different modulation sweep time, may be -T 6000 is better for your application. Just try and get the best configuration for you, this is one of the purposes of the radar evaluation board.
The next graph is the result to orient the radar to another direction, thus increasing the detection range up to 60m.
Note that an important thing to consider in this kind of measurements is the correct orientation of the radar. This is something you also must empirically optimise for your specific application.
Measurements obtained orienting the radar to get a wider range, up to 60m
Picture of the set up used: radar fixed on a tripod, oriented to de road and laptop to monitor the data in real time
Radar technology can also be used to make non contact remote measurements of building vibrations with high precision level.
This is a very interesting application on the construction area.
IMPORTANT: An special software has been developed to this specific application and cannot be done with IGEP RADAR LAMBDA radar application demo. If you want to use this feature you must contact with radar support and we will study your case and make you an offer for an special software.
The next picture show the construction that was monitored. It is a 344 m length viaduct, located in Sant Boi, over the Llobregat river. High speed trains pass over this bridge and the purpose of the measurements is to detect the vibration of the construction while trains are passing:
Measurements set up: IGEP RADAR LAMBDA with special software is placed on the ground facing the top of the viaduct at a distance of 15m
The next figure show the graphic obtained at the time the train was passing over the bridge. This graphic can be shown in real time while train is passing over the bridge. The data can be also captured for further processing. The graphic represents the relative position of the top of the viaduct over the ground. Yes, the structure moves 2mm, will you take the next high speed train? Do not worry, this is normal and you do not have to worry about it, this vibration response has a normal profile.
The advantage of this kind of measurements is that depending on the vibration wave profile engineers can detect structure problems, in the same way a doctor can check if your hearth is OK by lookintg at an EKG.
Vibration wave obtained on the measurement. It can be clearly shown the high precision of IGEP RADAR LAMBDA using special data processing
Y Axis: Relative movement between viaduct top and ground in millimeters
X Axis. Time in seconds
We can see from the graph that IGEP RADAR LAMBDA has a precision better than 100 micrometers in this kind of application
2D graph representing frequency domain measurements.
Y Axis: Represents the sample number of the FFT of the captured data so, it is the Frequency axis
X Axis: represents the Captured file number, so it is the Time axis
This is another way to present the results, by representing all the measurements FFT. This graph was obtained processing the captured data with matlab on a PC, they are not plotted by the radar equipment. It can be seen the peak due to the viaduct top reflection, and in the mid measurements it can be noted the vibration during the time the train is passing.
The figure below is a 3D representation of the same: FFT module of all the measurements. Look at the vibration in the central area.
3D graph representing frequency domain measurements.
Tank Level Gauge
Radar technology can also be used to make non-contact precise tank level measurements.
IGEP RADAR LAMBDA can be used on this kind of application when pressure and temperature conditions are not extreme and no chemical corrosive material is surrounding the equipment. You only have to install the radar at the top of the tank with the antenna facing the material surface contained on the tank and start measuring range.
In case of extreme environmental conditions it would be needed to develop special antenna that can support high pressures and temperatures. ISEE engineering team is able to develop such kind of devices, please contact with radar support if you want we study your personal case.
Horn antenna to be used on extreme environmental conditions
The following applications require the development of an specific hardware and/or software to convert IGEP RADAR LAMBDA into a product. Please do not hesitate to contact with radar supportif you want we study your personal case:
Radar technology is beign more and more used on automotive safety systems. IGEP RADAR LAMBDA can be an interesting first step in order to evaluate the feseability of such systems and the reference design for developing more specialised products.
Speed Information Display
IGEP RADAR LAMBDA can be used in combination of a LED Display to develop Speed Information Displays:
IGEP RADAR LAMBDA can be used as short range altimeter for Unmanned Aerial Vehicles.
The estimated range of detection is from 1 to 100meters so, helping on take off and landing of the UAV thanks to the centimeters precission of K-Band radar technology in comparison with the meters precision of traditional 4GHz altimeters used traditionally on avionics.
Radar technology is widely used on marine detection systems. Traditional radars use pulsed modulations requiring high RF power levels, affecting power consumption and security of tripulation. FMCW Radar systems are bing more and more used in this application due to its higher precission and its drastically reduced power consumption.
Smart Cities: Intelligent Lighting Systems
Radar technology can help to develop intelligent systems for cost and energy savings on Smart Cities systems applications. One example is the Intelligent Lighting System consisting on activation of a row of street lights only when the radar detects a person approaching the area, thus saving lot of illuminating hours with nobody in the street. IGEP RADAR LAMBDA could be integrated on this kind of energy efficient systems.
Radar technolgy can be also used in several sports where ball speed or trajectory is an important source of information to improve performance. Archers, Tennis and Golf players are becoming more and more dependants on this kind of information in order to improve their techniques.
Radar calibration can be done using different techniques.
The only way to compare the performance of different radars is by defining an specific test set up in order each radar can be compared in the same conditions.
One of the most widely used techniques is based on the use of special trihedral reflectors.
Trihedral reflectors are easy to build and stable representing a good reference to be sure the compared radars have the same target reflection characteristics.
The next pictures show an example of how to make this kind of measurements.
First of all you have to find an scenario free of reflections that could disturb the measurements. Ideally you need a free space area of more than 300 meters.
Then you put the radar on a stable basis and start doing measurements of the signal reflected by the thriedral reflector placed in front of the radar. You can start by placing the reflector at 10 meters distance far away from the radar and after that you move it in steps of 5 more meters far away until the radar gets a signal in the limits of the detection. A threshold of 10dB over the noise signal can be used as a reference.
IGEP RADAR LAMBDA detected a 200mm trihedral reflector placed at 70 meters distance