Jon this is such a great series! I’ve got a Phaser and this is the perfect way to get familiar with the hardware. I am very much looking forward to the final video. I suspect you are super busy getting ready for the May Hands On Phased Array Workshop. I wish I could attend but cannot.
Any chance you can try doing a series using Plasma Antenna? I know It's difficult to get info about them, that's why we should push on civilian side the knowledge regards these superior antennas. I would love to start seeing open projects that use them
Hey Jon, I’ve noticed your python GitHub examples (which are super instructive BTW) run quite a bit slower on my stock phaser board than what appears in your videos. Have you perhaps overclocked your pi 4 a bit? I am running the latest downloaded ADI phaser image and updated software.
No I'm not running the programs on the Rpi. I prefer to run, and edit them, on my desktop computer. So just connect Pluto (via USB) to your desktop computer, and ethernet from computer to Rpi). It will run much better, especially for processing intensive things like CFAR. Also, for technical support on Phaser/Pluto (and much faster replies!), please use this forum: ez.analog.com/adieducation/university-program/
hi, is it possible to use already developed radar sensors from car industry like "adaptive cruise control radar" for drone tracking or other purposes , with developing code for it.
@@Enderbro3300 i was reading technical documentation from OEM parts manufacturer BOSCH it is radar working in 77GHz but they don't specify what type is it FMCW or CW. YT channel @Thesignalpath was doing dissection of this radars.
They are indeed FMCW and can discriminate between stationary and moving objects as well as figure out speed differences for emergency braking. It obviously also needs the steering wheel angle.
Possibly. Most cars that have AEB or adaptive cruise control use this kind of radar. The first generations just provided a distance to target and target relative speed, I think. The newer ones provide a list of targets, angular position, relative speeds. Pretty sure none of them can detect a small drone and cannot be easily used outside of a car. Bosch has a list of products 'off the highway' or something similar but everything is 'request a quote'.
Nice video, are their default settings for different target sizes, which I could change depending on the target of choice, and then fine tune due to the clutter?
Yes, I think @livy-kk7ob is right. It seems like people sort this out on the radar range. Though I'm sure they have much better methods than my slider bars! Maybe some others that have experience with this can give more background on how it is done in the real world?
Hi This is great stuff. I'm really enjoying it. One suggestion: You can do that loop much faster with something like: half_width = num_guard_cells + num_ref_cells kernel = np.zeros(1+2*half_width) kernel[:num_ref_cells] = 1 kernel[-num_ref_cells:] = 1 kernel /= 2*number_ref_cells cfar_values = np.convolve(X_k, kernel, 'valid') edges = np.ones(half_width)*cfar_values.min() cfar_values = np.hstack((edges, cfar_values, edges)) # Pull out targets only as before I didn't test the code but I think it is OK. I might have the indexing a little off and I didn't use the masked array. Just a suggestion. Again, good stuff.
Great thorough, but easy to understand coverage of these topics. One thing that might have been interesting to see in a demo is results of multiple targets are visible to the radar beam. Realize keeping the experimentation/demonstration setup has advantages. Thinking of use case of detecting bicycles, or cars.
Yes, but the squelch adapts to the data. In real radars there is also filter and range-gate blanking for signal mixing birdies and reflection anomalies.
The Phasers are hard to keep in stock, but post your question at the link below, and ADI will try to help you find one: ez.analog.com/adieducation/university-program/
Jon this is such a great series! I’ve got a Phaser and this is the perfect way to get familiar with the hardware. I am very much looking forward to the final video. I suspect you are super busy getting ready for the May Hands On Phased Array Workshop. I wish I could attend but cannot.
Where is the part:6 of this series.
(Multi Chirp range Doppler)
ruclips.net/video/xnaaD9Um3K8/видео.htmlsi=gjiNoWx1SqAWuG-9
Any chance you can try doing a series using Plasma Antenna? I know It's difficult to get info about them, that's why we should push on civilian side the knowledge regards these superior antennas. I would love to start seeing open projects that use them
Hi. May i ask you what will be the expected range of this project when it comes to size of a consumer drone?
4:52 Looks like the false alarms that remain after CFAR tend to be ephemeral and could be filtered out by some kind of temporal averaging.
Hey Jon, I’ve noticed your python GitHub examples (which are super instructive BTW) run quite a bit slower on my stock phaser board than what appears in your videos. Have you perhaps overclocked your pi 4 a bit? I am running the latest downloaded ADI phaser image and updated software.
No I'm not running the programs on the Rpi. I prefer to run, and edit them, on my desktop computer. So just connect Pluto (via USB) to your desktop computer, and ethernet from computer to Rpi). It will run much better, especially for processing intensive things like CFAR. Also, for technical support on Phaser/Pluto (and much faster replies!), please use this forum: ez.analog.com/adieducation/university-program/
hi, is it possible to use already developed radar sensors from car industry like "adaptive cruise control radar" for drone tracking or other purposes , with developing code for it.
Are those actually radar? Or are they lidar? I know that a lot of marketing uses lidar, radar, and ultrasonic interchangeably
@@Enderbro3300 i was reading technical documentation from OEM parts manufacturer BOSCH it is radar working in 77GHz but they don't specify what type is it FMCW or CW.
YT channel @Thesignalpath was doing dissection of this radars.
They are indeed FMCW and can discriminate between stationary and moving objects as well as figure out speed differences for emergency braking. It obviously also needs the steering wheel angle.
Temic (continental) makes 77ghz radars (not lidar), quite a lot of cars have them. I think from 2009.
Possibly. Most cars that have AEB or adaptive cruise control use this kind of radar. The first generations just provided a distance to target and target relative speed, I think. The newer ones provide a list of targets, angular position, relative speeds. Pretty sure none of them can detect a small drone and cannot be easily used outside of a car. Bosch has a list of products 'off the highway' or something similar but everything is 'request a quote'.
Nice video, are their default settings for different target sizes, which I could change depending on the target of choice, and then fine tune due to the clutter?
It looks like you need to set up a test environment yourself to debug these parameters.
@@livy-kk7ob I am finding this series of videos quite exciting, I am looking forward to my own setup later in the year. God bless.
Yes, I think @livy-kk7ob is right. It seems like people sort this out on the radar range. Though I'm sure they have much better methods than my slider bars! Maybe some others that have experience with this can give more background on how it is done in the real world?
Amazing stuff, this is what the Internet was made for! Incredibly interesting and valuable, thanks for sharing your learning process with us
I just watched all five videos in the series. Cool stuff!
This guy is super smart and must have gone to an awesome college.
Hi This is great stuff. I'm really enjoying it.
One suggestion: You can do that loop much faster with something like:
half_width = num_guard_cells + num_ref_cells
kernel = np.zeros(1+2*half_width)
kernel[:num_ref_cells] = 1
kernel[-num_ref_cells:] = 1
kernel /= 2*number_ref_cells
cfar_values = np.convolve(X_k, kernel, 'valid')
edges = np.ones(half_width)*cfar_values.min()
cfar_values = np.hstack((edges, cfar_values, edges))
# Pull out targets only as before
I didn't test the code but I think it is OK. I might have the indexing a little off and I didn't use the masked array.
Just a suggestion. Again, good stuff.
Such a gem you are!
Great thorough, but easy to understand coverage of these topics.
One thing that might have been interesting to see in a demo is results of multiple targets are visible to the radar beam. Realize keeping the experimentation/demonstration setup has advantages. Thinking of use case of detecting bicycles, or cars.
+1
nice
Really enjoying this series, I'm looking forward to the next video!
So, basically you are adjusting the squelch???
Yes, but the squelch adapts to the data. In real radars there is also filter and range-gate blanking for signal mixing birdies and reflection anomalies.
do you also perform concultancy work on radar?
any information about max range?
and also price
Jon or anyone else.. is there a kit we can buy?
The Phasers are hard to keep in stock, but post your question at the link below, and ADI will try to help you find one: ez.analog.com/adieducation/university-program/