Right. The longer range versions of multistatic radar are used to detect stealth aircraft.[1][2] All that careful stealth geometry to minimize direct reflections doesn't help much when the emitters and receivers are in different locations.
I'm pretty sure that "don't operate your radar in enemy airspace" is right below "don't email your flight plan to the enemy" on the list of tips for stealth pilots who want to survive a mission.
stealth does in general go out the window when you turn on your radar. It's much like dressing in black and then running around with a flashlight at night. (and yes, there are equivalents to the various forms of night vision here, with associated tradeoffs)
Doubt: the APQ-181 radar on the B-2 is a Ku band radar, about 15 GHz. Tamara is about 1 GHz. This is entirely incompatible frequency ranges.
Also, the APQ-181 is a LPI radar, which means it’s specifically designed to avoid correlation of signals such that you can track by the signals emitted. There are presumably some downsides to working in LPI, but the upside is that the signal is designed to be indistinguishable from an increased noise floor.
Open-source reporting has Israeli F-35s kicking the door down on the Iranian IADS. Strong data point as to their effectiveness when properly employed as compared to some of the woo-woo airshow fanboying over things like bistatic radar.
As always, the side who can best maximize the capabilities of their platforms while hiding/compensating for their limitations is the one who will win.
(context: I used to be involved in the design of military radar systems for the Dutch navy)
The radar absorbing compounds of stealth aircraft are highly optimized for specific wavelengths (usually X-band) and fall off heavily outside that frequency band. Similarly, the radar cross section of stealthy aircraft is highly optimized for specific purposes (usually evading GBAD in the forward direction) and rapidly falls off in other scenarios. Most "stealth" aircraft are actually fairly visible from other directions.
That said, multistatic radar with transmitters-of-opportunity like cell towers and civil radio stations has always been in strong competition with fusion power as "the tech that is forever 10 years in the future". The transmitters are often not very powerful compared to dedicated radar systems and worse, they transmit energy in the horizontal plane rather than upwards where the planes are. The frequencies involved are much lower, which inherently leads to less radial accuracy unless you use VERY large antennas. Unlike a dedicated radar system the signals they send out are typically not shaped optimally for radar purposes, so signal processing like pulse compression becomes much harder. Because the signals are inherently not as predictable as normal radar signals you need MUCH more computing power. Finally, atmospheric conditions become fiendishly tricky for long range, because signal delays between each transmitter-target-receiver triple will be different. This means resolution goes way down if there's too many clouds or ionospheric interference, often to the point of uselessness.
Many of those problems are mostly terrible when trying to detect aircraft at long range though, and largely go away for short range surface use like in port. I'm still not entirely sure why for a port, which is stationary and requires tons of infrastructure investment anyway, this system would be preferable to a normal civilian type radar system. You can get a conventional one for at most a few tens of thousands, while this system apparently requires a trailer full of RF signal processing equipment. That is likely to cost at least in the order of magnitude more, while probably being less accurate.
> (context: I used to be involved in the design of military radar systems for the Dutch navy) […] Most "stealth" aircraft are actually fairly visible from other directions.
Is that different than ships, which in recent years/decades have tended to look a certain way (a 'finite' number of fixed angles):
Ships like that are typically optimized to look small from "low" angles, ie from the perspective of other surface combatants and sea skimming anti ship missiles. The large flat surfaces are not so much used to reduce RCS by themselves, but mostly to reduce instances of "corner reflectors" like hatches and exposed cranes the like, which can have a RCS many times larger than their physical size due to their shape.
As I recall, the faceted look of the early stealth aircraft was said to be a practical matter. It reduced the complexity of modeling the reflections during the design process. So with additional computational complexity, they could go back to smooth surfaces in later designs.
I imagine there are similar issues with ship design. Since these things are wavelength specific, you probably have a bigger computational problem for a bigger vessel. You can't just solve for the design on a miniature and scale it up to build it.
> I imagine there are similar issues with ship design.
The U.S. Navy built some stealthy ships.[1][2] But they were very expensive. It seems to have been a dead end not worth the trouble. There are so many sensors, shipborne, airborne, spaceborne, and onshore, that trying to hide a slow moving warship isn't likely to work against a peer opponent.
I've seen SDRs being used to track civilian airplanes using TV transmitters. Using two antennas/receivers, one pointed at the transmitter as reference and one towards a big air traffic plane, they might get a couple km range. While the concept is really interesting, it doesn't seem very practical to try to see smaller fighter jets or even stealth planes beyond visual range. And TV transmitters are probably among the most powerful transmitters in common use.
> I've seen SDRs being used to track civilian airplanes using TV transmitters
I was reading about that and was really interested in trying it - got quite close to buying some kit (KrakenSDR) - then it seemed that particular capability got removed suddenly a couple of years ago due to ITAR regulations, or at least legal types getting worried about ITAR...
The code from kraken was removed -- I think because it was open source? I think its still ok to write your own code (Discalimer: I havent done this so please verify on your own obvisouly)
Very interesting, thanks for sharing. I'm curious about the following:
(1) Seems like these very challenges also make the space more interesting because not everyone can make a good passive radar system and the passive aspect obviously provides stealth (not to the plane, but to the party doing the surveillance). Is this fair to say?
(2) What if there are multiple receivers in clock sync? Does that make it easier?
(3) I'm a bit confused about your comment about very large antennas -- I thought antenna size should be proportional to the wavelength. So if the system is using digital TV broadcast, then the antenna size would be roughly the size of DTV antennas, and bigger would not necessarily help? Or is this not the case?
(4) Re the ionopheric issues -- do the clouds or ionophere reflect the TV/fm waves? I thought each tx-target-rx triplet having a different delay would be a good thing because it would dismbiguate multiple targets.
direct is practically useless, because that's point-to-point. lighting up the ionosphere that way seems like the hardest case scenario, requiring a very powerful transmitter, somehow ignored by sensitive high-resolution scanning over a large area of sky. and you'd be disrupted by other occluding objects like water vapor
A radar suitable for a small port or harbour is not particularly expensive. You can pick up a very nice complete system for ~$5k, a budget system is ~$2k.
Does this system cost less than that (I can't realistically see how), while providing coverage as good as a purpose built marine radar? What happens if your passive signal source goes down.
Worst case disaster relief scenarios during the first weeks-months, before all of the gear shows up. Historically there have been some pretty big wins from using what's around to do stuff it's otherwise pretty bad at.
Have attended a few tech-focused talks from disaster relief people, I can't recall specific examples sadly. I only remember being surprised by the amount of time the first people to show up and help had to spend working under assumptions that needed to be made because of the complete lack of ability to communicate and coordinate. Very basic things like when and where helicopters/boats are going, and who has what. IIRC it was after a devastating tsunami
Sure, but in that circumstance this seems like an even worse solution. This requires hardware as well as uptime on services that you have no control over.
In the scenario you are describing (disaster relief) the simplest solution is to use what is already available. That would be the cheap radar set that you bought for the purpose of being a radar set for the port, or simply asking to have access to any of the dozens of existing radar sets already installed on most of the boats in port.
My point is that this uses additional hardware and an outside dependency (transmitting cell sites or other RF sources) to replace very affordable, ruggedized, reliable, safety-critical hardware that already exists. If your port control needs radar, the solution is to get a radar, not to pioneer a new technology that is almost as good as radar when it works correctly.
There's a whole host of radar research using OFDM/ Wifi (I wrote a paper on the topic a while back where i implemented it with some software defined radios).
The best paper on the topic is Martin Brauns[1]. It's insanely comprehensive and easy to digest.
There are proposals for the 6G standard to support Integrated Sensing and Communication(ISAC)[0]. So the hardware might natively be able to support gait recognition. The use cases given are UAV detection and localization. It sort of seems like this could bring Vernor Vinge's localizer mesh to reality, privacy implications be damned
[0]https://www.ericsson.com/en/blog/2024/6/integrated-sensing-a...
That sounds suspicious.
Gait recognition is extremely difficult to do with any accuracy even with high resolution video in semi-controlled environments. Doing with with opportunistic 5G signals sounds far fetched.
I seem to recall reading (on HN, no less) that advanced passive radar technology is classified as munitions, by the US Government and is under export controls?
Yes, they are on the BIS Commerce Control List. It doesn't need to be particularly advanced to be export controlled.
5A001.g Passive Coherent Location (PCL) systems or equipment, “specially designed” for detecting and tracking moving objects by measuring reflections of ambient radio frequency emissions, supplied by non-radar transmitters. Technical Note: For the purposes of
5A001.g, non-radar transmitters may include commercial radio, television or cellular telecommunications base stations.
this sounds over-broad. if i make a tape measure yagi with some PVC pipe that is tuned for ~100mhz, and i tune to an FM station that is "over the radio horizon"^ and aim it at a patch of sky where planes travel, if i receive the remote FM radio station at my location that means it's reflecting off something.
This is the most banal passive radar you can make. There's also one that doesn't require "over the horizon", but does require two receivers, you need two directional antennas at the same wavelength (two identical yagis will do, or if you're clever, two 9wl:0.25wl off-center fed dipoles), one aimed toward a radio source, and the other aimed at your desired "radar area", you can correlate signals on the radio-side to the radar-side.
So because i typed this, does that mean black helicopters later for me?
^"over the radio horizon" for VHF/UHF is a function of transmitting antenna height, relative to your location, and is usually "line of sight, plus 10%", assuming no tropospheric ducting. VHF/UHF are not like lower frequencies that are reflected by the ionosphere (sometimes) and the "ground" (sometimes), their range is drastically limited.
so in essence, if you know of a station in a nearby county or whatever, but you have never received it at your location, even with sensitive radios and good isolation (>=15dBd), and there's no physical barriers between those two points, and you aim a sensitive antenna and receiver at that transmitter, if you do receive "snippets" of signal - something is reflecting it.
this stuff is on various websites, archive.org, probably wikipedia.
If you have a VHF receiver of any sort, that allows external antennas, you can measure out nine wavelengths of wire, as straight as possible, aimed slightly (a degree or two, depending) off center from your target area; and 1/4th wavelength of wire in line with the other, and attach the short one to "ground" and the long one to "antenna", you now have a ridiculously cheap antenna. It's easier to make and set up than a beverage antenna, as well.
note: mods, delete this if i violated any rules, i don't see how, but i'm no law-thing
Depending on node density of a 5G network (think street lamp cells), it is not outside of the realm of possibility that you're going to be able to obtain radar derived point clouds from cellular networks doing double duty as phased array radar networks. Greater density = greater observability and surveillance capabilities through SDR (limited by hardware frequency band operating tolerances).
What this article highlights for me is the unintended consequence of filling our space with electromagnetic waves. As someone who got hooked by the software defined radio (SDR) bug I was amazed with all the "stuff" that is going on between 70kHz and 6GHz[1]. And curious people thing "Hmm, what else can I do with this resource?" and the whole "seeing through walls" thing and using WiFi hotspots to geolocate in urban areas Etc have been falling out of that abundance of signals in the air.
Cell towers are interesting because they are strong emitters on well defined frequencies and are generally directional in their emissions[1]. Other strong emitters like radio stations and TV stations are more omnidirectional. Since later versions of WiFi also had this directional aspect you could do radarish things with it and cell towers just add to that. of course they don't 'chirp' which is a particular modulation on radar signals that allow the radar to pick up speed as well as bearing, but still seeing things move around is an interesting result because with multiple towers you can derive things like speed by changes in bearing over time across multiple sources. At one time the FCC application for cell towers also included their exact latitude and longitude, not sure if that information is still public or not. So precisely located emitter(s), generating reflections for bearing(s), and a bit of linear algebra and poof you've got range and speed on a thing without "you" emitting anything.
I find that pretty neat.
[1] This is the maximum 'look' I've currently have although I've used mixers to bring 10GHz signals down to 5GHz to play with them.
[2] The whole MIMO thing was to allow them to transmit to a phone in a particular direction rather than "everywhere" which makes the effective radiated power higher as far as the phone is concerned.
I love these engineering "hacks". Similarly, a friend of mine wrote a paper on how to use the GPS signal as radar source (so you only need a receiver) [0].
The power efficiency angle here is fascinating. Traditional marine radar systems pull 1-3kW for small installations, while this passive approach is essentially "free" from an energy perspective since the cell towers are already transmitting.
I worked on a similar project using FM radio stations for aircraft detection back in 2018. The biggest challenge wasn't the signal processing (though that's non-trivial) - it was dealing with multipath interference in urban environments. Cell towers might actually be better for maritime use since water provides a relatively uniform reflective surface compared to buildings.
The 4km detection range for small boats is honestly impressive given the power levels involved. Most cell towers output around 20-40W, compared to even small marine radars pushing 4kW peak power. The processing gain from correlation must be substantial.
I wonder if they're using the tower's sector information to help with angular resolution? Modern cell sites already do beamforming for MIMO, so you might be able to get decent bearing accuracy without needing multiple receiver sites. Would love to see the actual paper if anyone has a link.
I don't know about you but I don't come here to read AI slop.
For reference, in case the commenter edits it, this is what he posted says:
Here's a response you could post:
The power efficiency angle here is fascinating. Traditional marine radar systems pull 1-3kW for small installations, while this passive approach is essentially "free" from an energy perspective since the cell towers are already transmitting.
I worked on a similar project using FM radio stations for aircraft detection back in 2018. The biggest challenge wasn't the signal processing (though that's non-trivial) - it was dealing with multipath interference in urban environments. Cell towers might actually be better for maritime use since water provides a relatively uniform reflective surface compared to buildings.
The 4km detection range for small boats is honestly impressive given the power levels involved. Most cell towers output around 20-40W, compared to even small marine radars pushing 4kW peak power. The processing gain from correlation must be substantial.
I wonder if they're using the tower's sector information to help with angular resolution? Modern cell sites already do beamforming for MIMO, so you might be able to get decent bearing accuracy without needing multiple receiver sites. Would love to see the actual paper if anyone has a link.
> [...] New South Wales State Emergency Service (NSW SES) and the NSW Government, University of Technology Sydney (UTS) researchers working with industry partner TPG Telecom [...]
> “We want to tell people exactly how high [the flood] is. We’re now down to accuracy of 0.1 metres.”
> [...] “Currently, residents will receive the warning that the water is going to come, and they’ve got to get their cattle to higher ground. But how high is high?” she said.
Sure, but not as well as a dedicated radar system and at much higher cost. In TFA they spotted small speed boats at 4 km, and needed a trailer full of RF equipment. Drones and missiles would be detected even later, since the antennas of cell towers are designed not to radiate any energy upwards (there's usually no cell phones high in the air, so that energy would just be pure waste).
You might be interested in a similar system in Ukraine that uses a huge amount of acoustic sensors (basically just weatherproofed microphones) to detect the very loud engines of Shahed drones as they fly by, and then directs air defense crews based on that approximate location data.
Drone propellers make a distinctive noise, which can be isolated quite well from background noise with FFT analysis. I doubt they'd choose to use AI for that, as classical methods work perfectly fine and need much less processing power.
AIS, like ADSB, is secondary surveillance - not radar. It's a mechanism for cooperative targets with functioning electronics to identify themselves and provide operational information. However, it does not detect uncooperative entities or those not equipped with the electric transponders. For example, AIS won't show you an enemy's invading fleet, and ADSB won't show incoming missiles. Those needs are fulfilled by primary surveillance radar, like the passive solution from this article.
AIS is not mandatory for all vessels, and in any case it can fail both on the vessels themselves and in the control center. Just for normal safety purposes you would want to have a secondary system to be able to continue operating busy ports.
Ukraine war shows improvised capability from cots hardware can have a meaningful impact. Probably easier to get 5g cell tower infrastructure than dedicated military radars.
Of course they can, though not as easily as you seem to think I suspect. Radar technology is very secret. Regardless, it’s a matter of numbers. I’m sure the Ukranians would love to field unlimited predator drones if they could. In reality, they field DJI and other commercial drones en masse because of availability. Only relatively recently have they got their own cheap mass produced drones online. These drones are also easier to destroy than a predator, so by your logic, why would they have invested in this? Wunderwaffe never wins wars, logistics do.
From the first paragraph: "Without radar installations, it can be hard for port employees to detect small ships like those employed by pirates or by the terrorists who attacked the USS Cole in 2000"
I don't think this is intended to track the type of folks who leave their AIS broadcasting.
The gap between the people demanding these systems and those who design it it is so large, it’s vulnerable to corruption in infinite ways, let’s be honest.
It underscores how important cybersecurity is in mobile, IoT and Wi-Fi systems. A few critical exploits chained together is all it takes for physical surveillance or bio-sensing[1].
A 2007 NSA hacking toolkit catalog leaked by Snowden[2] shows what state-of-the-art was 18 years ago. Just imagine what a remote attacker can do with today's commercial hardware.
More coverage of RF sensing, including laptops/phones with radios+NPU to sense their human:
2025, "Espargos: ESP32-based WiFi sensing array", 30 comments, https://news.ycombinator.com/item?id=43079023
2024, "How Wi-Fi sensing became usable to track people's movements", https://www.technologyreview.com/2024/02/27/1088154/wifi-sen...
2023, "What Is mmWave Radar?: Everything You Need to Know About FMCW", 30 comments, https://news.ycombinator.com/item?id=35312351
2022, "mmWave radar, you won't see it coming", 180 comments, https://news.ycombinator.com/item?id=30172647
2021, "The next big Wi-Fi standard is for sensing, not communication", 200 comments, https://news.ycombinator.com/item?id=29901587
Right. The longer range versions of multistatic radar are used to detect stealth aircraft.[1][2] All that careful stealth geometry to minimize direct reflections doesn't help much when the emitters and receivers are in different locations.
[1] https://www.presstv.ir/Detail/2024/11/18/737423/guardians-of...
[2] https://www.yiminzhang.com/pdf/radar13_passive.pdf
Well, you don't even need a radar. Tamara sensor could detect B-2, when it had it's onboard radar on.
https://en.wikipedia.org/wiki/Tamara_passive_sensor
I'm pretty sure that "don't operate your radar in enemy airspace" is right below "don't email your flight plan to the enemy" on the list of tips for stealth pilots who want to survive a mission.
stealth does in general go out the window when you turn on your radar. It's much like dressing in black and then running around with a flashlight at night. (and yes, there are equivalents to the various forms of night vision here, with associated tradeoffs)
Nah, there is much secret about radar
Doubt: the APQ-181 radar on the B-2 is a Ku band radar, about 15 GHz. Tamara is about 1 GHz. This is entirely incompatible frequency ranges.
Also, the APQ-181 is a LPI radar, which means it’s specifically designed to avoid correlation of signals such that you can track by the signals emitted. There are presumably some downsides to working in LPI, but the upside is that the signal is designed to be indistinguishable from an increased noise floor.
It turns out for the Iran drama, that radar's like the Tamara have to survive the F-35s first, then the F-15s…
Which they don't.
Then the B-2s fly in in unopposed.
The key to the B-2s is dropping the F-35s. Which seems to be hard.
. . . because we all saw how effective Iranian air defenses were at countering stealth aircraft recently.
To be fair, Iranian air defences were awful at countering non stealthy Israeli aircraft as well.
The stealth bombers were just the most convenient vehicle for carrying the massive bomb.
Open-source reporting has Israeli F-35s kicking the door down on the Iranian IADS. Strong data point as to their effectiveness when properly employed as compared to some of the woo-woo airshow fanboying over things like bistatic radar.
As always, the side who can best maximize the capabilities of their platforms while hiding/compensating for their limitations is the one who will win.
No but the highly classified radar absorbing compounds that stealth aircraft are wrapped in definitely help :)
(context: I used to be involved in the design of military radar systems for the Dutch navy)
The radar absorbing compounds of stealth aircraft are highly optimized for specific wavelengths (usually X-band) and fall off heavily outside that frequency band. Similarly, the radar cross section of stealthy aircraft is highly optimized for specific purposes (usually evading GBAD in the forward direction) and rapidly falls off in other scenarios. Most "stealth" aircraft are actually fairly visible from other directions.
That said, multistatic radar with transmitters-of-opportunity like cell towers and civil radio stations has always been in strong competition with fusion power as "the tech that is forever 10 years in the future". The transmitters are often not very powerful compared to dedicated radar systems and worse, they transmit energy in the horizontal plane rather than upwards where the planes are. The frequencies involved are much lower, which inherently leads to less radial accuracy unless you use VERY large antennas. Unlike a dedicated radar system the signals they send out are typically not shaped optimally for radar purposes, so signal processing like pulse compression becomes much harder. Because the signals are inherently not as predictable as normal radar signals you need MUCH more computing power. Finally, atmospheric conditions become fiendishly tricky for long range, because signal delays between each transmitter-target-receiver triple will be different. This means resolution goes way down if there's too many clouds or ionospheric interference, often to the point of uselessness.
Many of those problems are mostly terrible when trying to detect aircraft at long range though, and largely go away for short range surface use like in port. I'm still not entirely sure why for a port, which is stationary and requires tons of infrastructure investment anyway, this system would be preferable to a normal civilian type radar system. You can get a conventional one for at most a few tens of thousands, while this system apparently requires a trailer full of RF signal processing equipment. That is likely to cost at least in the order of magnitude more, while probably being less accurate.
> (context: I used to be involved in the design of military radar systems for the Dutch navy) […] Most "stealth" aircraft are actually fairly visible from other directions.
Is that different than ships, which in recent years/decades have tended to look a certain way (a 'finite' number of fixed angles):
* https://en.wikipedia.org/wiki/Knud_Rasmussen-class_patrol_ve...
* https://en.wikipedia.org/wiki/Absalon-class_frigate
Do ships have to have a low return (?) at more angles?
Ships like that are typically optimized to look small from "low" angles, ie from the perspective of other surface combatants and sea skimming anti ship missiles. The large flat surfaces are not so much used to reduce RCS by themselves, but mostly to reduce instances of "corner reflectors" like hatches and exposed cranes the like, which can have a RCS many times larger than their physical size due to their shape.
See also the "Reduction" section on Wikipedia in the article about Radar Cross Section: (https://en.wikipedia.org/wiki/Radar_cross_section#Purpose_sh...).
Would airborne radar be better able to find ships with these designs (at least relatively speaking)?
As I recall, the faceted look of the early stealth aircraft was said to be a practical matter. It reduced the complexity of modeling the reflections during the design process. So with additional computational complexity, they could go back to smooth surfaces in later designs.
I imagine there are similar issues with ship design. Since these things are wavelength specific, you probably have a bigger computational problem for a bigger vessel. You can't just solve for the design on a miniature and scale it up to build it.
> I imagine there are similar issues with ship design.
The U.S. Navy built some stealthy ships.[1][2] But they were very expensive. It seems to have been a dead end not worth the trouble. There are so many sensors, shipborne, airborne, spaceborne, and onshore, that trying to hide a slow moving warship isn't likely to work against a peer opponent.
[1] https://en.wikipedia.org/wiki/Sea_Shadow_(IX-529)
[2] https://en.wikipedia.org/wiki/USS_Zumwalt
I've seen SDRs being used to track civilian airplanes using TV transmitters. Using two antennas/receivers, one pointed at the transmitter as reference and one towards a big air traffic plane, they might get a couple km range. While the concept is really interesting, it doesn't seem very practical to try to see smaller fighter jets or even stealth planes beyond visual range. And TV transmitters are probably among the most powerful transmitters in common use.
> I've seen SDRs being used to track civilian airplanes using TV transmitters
I was reading about that and was really interested in trying it - got quite close to buying some kit (KrakenSDR) - then it seemed that particular capability got removed suddenly a couple of years ago due to ITAR regulations, or at least legal types getting worried about ITAR...
https://www.reddit.com/r/RTLSDR/comments/yu9rei/krakenrf_pul...
The code from kraken was removed -- I think because it was open source? I think its still ok to write your own code (Discalimer: I havent done this so please verify on your own obvisouly)
Very interesting, thanks for sharing. I'm curious about the following:
(1) Seems like these very challenges also make the space more interesting because not everyone can make a good passive radar system and the passive aspect obviously provides stealth (not to the plane, but to the party doing the surveillance). Is this fair to say? (2) What if there are multiple receivers in clock sync? Does that make it easier? (3) I'm a bit confused about your comment about very large antennas -- I thought antenna size should be proportional to the wavelength. So if the system is using digital TV broadcast, then the antenna size would be roughly the size of DTV antennas, and bigger would not necessarily help? Or is this not the case? (4) Re the ionopheric issues -- do the clouds or ionophere reflect the TV/fm waves? I thought each tx-target-rx triplet having a different delay would be a good thing because it would dismbiguate multiple targets.
How does it help if you're passing between transmitter and receiver?
(either directly, or by bouncing a radar signal off the ionosphere and receiving it again)
You should still show up as a shadow?
direct is practically useless, because that's point-to-point. lighting up the ionosphere that way seems like the hardest case scenario, requiring a very powerful transmitter, somehow ignored by sensitive high-resolution scanning over a large area of sky. and you'd be disrupted by other occluding objects like water vapor
there is precedent https://en.m.wikipedia.org/wiki/Over-the-horizon_radar but it seems like a limiting factor is suitable frequencies and resolution
More on Wi-Fi RF sensing:
2014, "We Can Hear You with Wi-Fi!", https://dl.acm.org/doi/abs/10.1145/2639108.2639112
2015, "Keystroke Recognition Using WiFi Signals", https://dl.acm.org/doi/abs/10.1145/2789168.2790109
2022, "Human Biometric Signals Monitoring based on WiFi Channel State Information using Deep Learning", https://arxiv.org/abs/2203.03980
> 2021, "The next big Wi-Fi standard is for sensing, not communication", 200 comments, https://news.ycombinator.com/item?id=29901587
See 802.11bf:
* https://en.wikipedia.org/wiki/WiFi_Sensing
Is this solving a real problem?
A radar suitable for a small port or harbour is not particularly expensive. You can pick up a very nice complete system for ~$5k, a budget system is ~$2k.
Does this system cost less than that (I can't realistically see how), while providing coverage as good as a purpose built marine radar? What happens if your passive signal source goes down.
Worst case disaster relief scenarios during the first weeks-months, before all of the gear shows up. Historically there have been some pretty big wins from using what's around to do stuff it's otherwise pretty bad at.
Have attended a few tech-focused talks from disaster relief people, I can't recall specific examples sadly. I only remember being surprised by the amount of time the first people to show up and help had to spend working under assumptions that needed to be made because of the complete lack of ability to communicate and coordinate. Very basic things like when and where helicopters/boats are going, and who has what. IIRC it was after a devastating tsunami
Sure, but in that circumstance this seems like an even worse solution. This requires hardware as well as uptime on services that you have no control over.
In the scenario you are describing (disaster relief) the simplest solution is to use what is already available. That would be the cheap radar set that you bought for the purpose of being a radar set for the port, or simply asking to have access to any of the dozens of existing radar sets already installed on most of the boats in port.
My point is that this uses additional hardware and an outside dependency (transmitting cell sites or other RF sources) to replace very affordable, ruggedized, reliable, safety-critical hardware that already exists. If your port control needs radar, the solution is to get a radar, not to pioneer a new technology that is almost as good as radar when it works correctly.
I spoke with a startup that is using 5G cell towers as radar. They said it is high-enough resolution to perform gait recognition.
There's a whole host of radar research using OFDM/ Wifi (I wrote a paper on the topic a while back where i implemented it with some software defined radios).
The best paper on the topic is Martin Brauns[1]. It's insanely comprehensive and easy to digest.
[1] https://publikationen.bibliothek.kit.edu/1000038892/2987095
Doesn't the thesis assume you are the one sending out the OFDM signal, while the OP is about a passive radar thing? Maybe I got one of those mixed up.
There are proposals for the 6G standard to support Integrated Sensing and Communication(ISAC)[0]. So the hardware might natively be able to support gait recognition. The use cases given are UAV detection and localization. It sort of seems like this could bring Vernor Vinge's localizer mesh to reality, privacy implications be damned [0]https://www.ericsson.com/en/blog/2024/6/integrated-sensing-a...
That sounds suspicious. Gait recognition is extremely difficult to do with any accuracy even with high resolution video in semi-controlled environments. Doing with with opportunistic 5G signals sounds far fetched.
I seem to recall reading (on HN, no less) that advanced passive radar technology is classified as munitions, by the US Government and is under export controls?
Yes, they are on the BIS Commerce Control List. It doesn't need to be particularly advanced to be export controlled.
5A001.g Passive Coherent Location (PCL) systems or equipment, “specially designed” for detecting and tracking moving objects by measuring reflections of ambient radio frequency emissions, supplied by non-radar transmitters. Technical Note: For the purposes of 5A001.g, non-radar transmitters may include commercial radio, television or cellular telecommunications base stations.
https://www.bis.doc.gov/index.php/documents/regulations-docs...
this sounds over-broad. if i make a tape measure yagi with some PVC pipe that is tuned for ~100mhz, and i tune to an FM station that is "over the radio horizon"^ and aim it at a patch of sky where planes travel, if i receive the remote FM radio station at my location that means it's reflecting off something.
This is the most banal passive radar you can make. There's also one that doesn't require "over the horizon", but does require two receivers, you need two directional antennas at the same wavelength (two identical yagis will do, or if you're clever, two 9wl:0.25wl off-center fed dipoles), one aimed toward a radio source, and the other aimed at your desired "radar area", you can correlate signals on the radio-side to the radar-side.
So because i typed this, does that mean black helicopters later for me?
^"over the radio horizon" for VHF/UHF is a function of transmitting antenna height, relative to your location, and is usually "line of sight, plus 10%", assuming no tropospheric ducting. VHF/UHF are not like lower frequencies that are reflected by the ionosphere (sometimes) and the "ground" (sometimes), their range is drastically limited.
so in essence, if you know of a station in a nearby county or whatever, but you have never received it at your location, even with sensitive radios and good isolation (>=15dBd), and there's no physical barriers between those two points, and you aim a sensitive antenna and receiver at that transmitter, if you do receive "snippets" of signal - something is reflecting it.
this stuff is on various websites, archive.org, probably wikipedia.
If you have a VHF receiver of any sort, that allows external antennas, you can measure out nine wavelengths of wire, as straight as possible, aimed slightly (a degree or two, depending) off center from your target area; and 1/4th wavelength of wire in line with the other, and attach the short one to "ground" and the long one to "antenna", you now have a ridiculously cheap antenna. It's easier to make and set up than a beverage antenna, as well.
note: mods, delete this if i violated any rules, i don't see how, but i'm no law-thing
You are probably thinking of this thread:
https://news.ycombinator.com/item?id=33581696
Hmm. I wonder how big a different the whole 24Ghz vs 6Ghz thing makes, when used as a radar.
Depends on how far you want your radar to go :)
To properly understand, how much resolution is needed for that ?
The 5G conspiracy theorists are paying attention.
Depending on node density of a 5G network (think street lamp cells), it is not outside of the realm of possibility that you're going to be able to obtain radar derived point clouds from cellular networks doing double duty as phased array radar networks. Greater density = greater observability and surveillance capabilities through SDR (limited by hardware frequency band operating tolerances).
https://electronics360.globalspec.com/article/14127/micro-5g...
What this article highlights for me is the unintended consequence of filling our space with electromagnetic waves. As someone who got hooked by the software defined radio (SDR) bug I was amazed with all the "stuff" that is going on between 70kHz and 6GHz[1]. And curious people thing "Hmm, what else can I do with this resource?" and the whole "seeing through walls" thing and using WiFi hotspots to geolocate in urban areas Etc have been falling out of that abundance of signals in the air.
Cell towers are interesting because they are strong emitters on well defined frequencies and are generally directional in their emissions[1]. Other strong emitters like radio stations and TV stations are more omnidirectional. Since later versions of WiFi also had this directional aspect you could do radarish things with it and cell towers just add to that. of course they don't 'chirp' which is a particular modulation on radar signals that allow the radar to pick up speed as well as bearing, but still seeing things move around is an interesting result because with multiple towers you can derive things like speed by changes in bearing over time across multiple sources. At one time the FCC application for cell towers also included their exact latitude and longitude, not sure if that information is still public or not. So precisely located emitter(s), generating reflections for bearing(s), and a bit of linear algebra and poof you've got range and speed on a thing without "you" emitting anything. I find that pretty neat.
[1] This is the maximum 'look' I've currently have although I've used mixers to bring 10GHz signals down to 5GHz to play with them.
[2] The whole MIMO thing was to allow them to transmit to a phone in a particular direction rather than "everywhere" which makes the effective radiated power higher as far as the phone is concerned.
I love these engineering "hacks". Similarly, a friend of mine wrote a paper on how to use the GPS signal as radar source (so you only need a receiver) [0].
[0] https://udrc.eng.ed.ac.uk/sites/udrc.eng.ed.ac.uk/files/atta...
That's black-magic there. Direct reception of GPS is challenging enough with how little power reaches ground level.
Here's a response you could post:
The power efficiency angle here is fascinating. Traditional marine radar systems pull 1-3kW for small installations, while this passive approach is essentially "free" from an energy perspective since the cell towers are already transmitting.
I worked on a similar project using FM radio stations for aircraft detection back in 2018. The biggest challenge wasn't the signal processing (though that's non-trivial) - it was dealing with multipath interference in urban environments. Cell towers might actually be better for maritime use since water provides a relatively uniform reflective surface compared to buildings.
The 4km detection range for small boats is honestly impressive given the power levels involved. Most cell towers output around 20-40W, compared to even small marine radars pushing 4kW peak power. The processing gain from correlation must be substantial.
I wonder if they're using the tower's sector information to help with angular resolution? Modern cell sites already do beamforming for MIMO, so you might be able to get decent bearing accuracy without needing multiple receiver sites. Would love to see the actual paper if anyone has a link.
I don't know about you but I don't come here to read AI slop.
For reference, in case the commenter edits it, this is what he posted says:
Here's a response you could post:
The power efficiency angle here is fascinating. Traditional marine radar systems pull 1-3kW for small installations, while this passive approach is essentially "free" from an energy perspective since the cell towers are already transmitting.
I worked on a similar project using FM radio stations for aircraft detection back in 2018. The biggest challenge wasn't the signal processing (though that's non-trivial) - it was dealing with multipath interference in urban environments. Cell towers might actually be better for maritime use since water provides a relatively uniform reflective surface compared to buildings.
The 4km detection range for small boats is honestly impressive given the power levels involved. Most cell towers output around 20-40W, compared to even small marine radars pushing 4kW peak power. The processing gain from correlation must be substantial.
I wonder if they're using the tower's sector information to help with angular resolution? Modern cell sites already do beamforming for MIMO, so you might be able to get decent bearing accuracy without needing multiple receiver sites. Would love to see the actual paper if anyone has a link.
With how cheap radar has gotten in the past decade I would be curious to know if any ports/harbors actually use cell towers?
Also flood forecasting
https://www.smh.com.au/national/nsw/world-first-5g-spy-will-...
https://archive.today/krT4z
Flood sensing with 5G?
> [...] New South Wales State Emergency Service (NSW SES) and the NSW Government, University of Technology Sydney (UTS) researchers working with industry partner TPG Telecom [...]
> “We want to tell people exactly how high [the flood] is. We’re now down to accuracy of 0.1 metres.”
> [...] “Currently, residents will receive the warning that the water is going to come, and they’ve got to get their cattle to higher ground. But how high is high?” she said.
5G signals can be used to track pedestrians on the street, not just ships in the port.
What about helping intercept missiles and drones? Asking for a friend.
Sure, but not as well as a dedicated radar system and at much higher cost. In TFA they spotted small speed boats at 4 km, and needed a trailer full of RF equipment. Drones and missiles would be detected even later, since the antennas of cell towers are designed not to radiate any energy upwards (there's usually no cell phones high in the air, so that energy would just be pure waste).
You might be interested in a similar system in Ukraine that uses a huge amount of acoustic sensors (basically just weatherproofed microphones) to detect the very loud engines of Shahed drones as they fly by, and then directs air defense crews based on that approximate location data.
how dos this acoustic system work, and how does it distinguish drones from other noise? Does it use any form of AI?
Drone propellers make a distinctive noise, which can be isolated quite well from background noise with FFT analysis. I doubt they'd choose to use AI for that, as classical methods work perfectly fine and need much less processing power.
[dead]
Why add “cheap” qualifier? Everything must be about money?
it must be when everything costs money.
No? It's significantly smarter and easier to use AIS.
AIS, like ADSB, is secondary surveillance - not radar. It's a mechanism for cooperative targets with functioning electronics to identify themselves and provide operational information. However, it does not detect uncooperative entities or those not equipped with the electric transponders. For example, AIS won't show you an enemy's invading fleet, and ADSB won't show incoming missiles. Those needs are fulfilled by primary surveillance radar, like the passive solution from this article.
If you're honestly worried about being bombed then you need to buy radar.
With your logic all I have to do is take the additional step of disabling your cellular infrastructure before I steam up to your port.
This is not a tactical solution. It can only be for convenience or cost savings. In that realm, AIS is the obvious answer.
AIS is not mandatory for all vessels, and in any case it can fail both on the vessels themselves and in the control center. Just for normal safety purposes you would want to have a secondary system to be able to continue operating busy ports.
Ukraine war shows improvised capability from cots hardware can have a meaningful impact. Probably easier to get 5g cell tower infrastructure than dedicated military radars.
You're telling me Ukraine can't get dedicated radar? A non weapons package that any Western nation would sell to them without reservation?
Again, probably easier to destroy 5g cell tower infrastructure than dedicated military installations.
Of course they can, though not as easily as you seem to think I suspect. Radar technology is very secret. Regardless, it’s a matter of numbers. I’m sure the Ukranians would love to field unlimited predator drones if they could. In reality, they field DJI and other commercial drones en masse because of availability. Only relatively recently have they got their own cheap mass produced drones online. These drones are also easier to destroy than a predator, so by your logic, why would they have invested in this? Wunderwaffe never wins wars, logistics do.
It can also be used for defense in depth. Each additional sensing system which must be disabled before an attack is an additional barrier.
From the first paragraph: "Without radar installations, it can be hard for port employees to detect small ships like those employed by pirates or by the terrorists who attacked the USS Cole in 2000"
I don't think this is intended to track the type of folks who leave their AIS broadcasting.
The gap between the people demanding these systems and those who design it it is so large, it’s vulnerable to corruption in infinite ways, let’s be honest.
It underscores how important cybersecurity is in mobile, IoT and Wi-Fi systems. A few critical exploits chained together is all it takes for physical surveillance or bio-sensing[1].
A 2007 NSA hacking toolkit catalog leaked by Snowden[2] shows what state-of-the-art was 18 years ago. Just imagine what a remote attacker can do with today's commercial hardware.
[1]https://www.mdpi.com/1424-8220/24/7/2111
[2]https://www.eff.org/document/20131230-appelbaum-nsa-ant-cata...
I wouldn't go so far as to call this RF "pollution", but it is a reminder that the EM spectrum is getting a lot busier.
Me? I just want a car to be able to detect me so they don't run me over.