I'm surprised this isn't already happening in Ukraine. They could fly small surveillance drones deep in enemy territory, perch on a power line, and send back lots of data. Not just video, but also sound and triangulating signals. This would also be useful in fog by monitoring major roads where high altitude drones and satellites would be obstructed.
I'm not like a drone-i-ologist or nothin, but from what I gather, both sides have gotten really good at detecting and jamming drone communication in the Ukraine/Russia conflict, which would probably make that a tough use case. I've read that the newer attack drones are controlled by a reaaallly long, reaaalllyy thin fiber optic line!
New attack models are using shielded electronics that don't need GPS and are immune to traditional jamming. Relying on computer vision and old school navigation math.
Go ~X speed for Y distance(+/-) on Z heading until you reach A landmark and then start a new set of instructions.
Yeah, but not in ukraine. They brute forced fiber, fly by wire.
Dead reckoning via inertial sensors, cameras, etc are way to complex for the flight controllers without heavier hardware since theyre hugely inefficient.
AI at the sophistication to do this stuff is essentially bloatware. Like running electron instead of a bare metal gui.
Everything I’ve found about the drones in that conflict indicates they use automated navigation for pursuit once a target is locked, but human pilots before then.
you described two different steps: human pilots get to desired area and target locked. For the human part, if the drone isn't using fiber optic and is getting jammed (many types of jamming), the human pilot might not be able to communicate with the drone. If that's the case, how will the drone get to the desired area? that's where the two things i posted come into play.
Surveillance gathered by an completely autonomous drone with no outside data, stationed far enough away to require refueling, close enough to enemy operations to be useful, that then needs to make its way back to origin, intact, through hostile territory, quickly enough for the gathered information to be useful, seems like a preeeetty big lift. Something a startup would promise to tackle with a star team of technologists over the course of like 10 years? Sure. Something they’d have designed within the past, like, year while getting shot at? I’d have to see that believe it.
Those fiber optic lines only work 50-60% of the time. Often the drones are carried 20km on foot into position which sucks as you know half the equipment on your back won't work.
A jammed drone that's perched on a power line wouldn't fall out of the sky, and doesn't need to transmit 24x7, only when it detects some activity. The lack of a signal from it would itself be a signal of where the next attack is coming from. Anti-jamming weapons (missiles and autonomous drones) would also be useful, that lock on to any signal jamming sources and deliver the munitions directly to the target that's advertising itself.
There's also powerline communication that these drones could use, relaying a signal to a second drone perched back in friendly territory. And if the military is going around blowing up all of their power transmission lines, that's also going to hurt them.
Use the transmission lines to link up with an RF node to hand off. Friendly territory might just be some place without good Russian internal security coverage (perhaps deeper in Russia rather than towards the front lines if the node uses satcom).
A friend of mine worked on a covert comms system for the Rangers to use in the Battle For Berlin that thankfully never occurred. The idea was to clamp onto plumbing, fences, and similar infrastructure where possible. Nodes with radios handed off to other comms systems. It worked reasonably well in tests but I don't know that it went anywhere, point is that the theory is sound.
The prototype in the linked video was first tested back in 2023, and since then a few startups have set their sights on the technology (e.g. Nomadic Drones and Voltair).
When working on the linked prototype we ran into some fundamental issues. Firstly, the recharging will only work with AC lines, and there's currently a lot of hype around UHVDC lines which are not compatible. Secondly, the AC lines must carry substantial current (ideally thousands of amps) for the recharging to not take forever. You can of course carry a much larger transformer on the drone to compensate, but this will in turn severely limit your flight time (ours was 1kg on a drone of 4.5kg and we could charge with 50W from 300A line current). You also have to account for significant daily fluctuations in the line current. It'll be interesting to see how the tech evolves, and I'll definitely be following these startups closely.
I'd assume otherwise you have to have a way for the drones to meter their usage and pay the power company. It will likely make power theft easier, but it seems entirely viable to have an account with the power company where you report "I drew X joules from line Y" and for them to bill appropriately.
The simplest might be for the drone company to act as an intermediary. They'd bill drone users for charging and have contracts with utilities. The drone company could do some authentication / DRM / etc. so that you'd basically have to jailbreak your drone to charge without paying.
Yes, I'm sure the markup would be large as a percentage, but for most customers the convenience would be worth it. Most of the customers are probably commercial and don't want to risk getting banned or sued.
Feels like this is likely to be targeting government and major corporate clients, in which case they're probably in a strong place to negotiate agreements based on charge reported by the drone's on board software. Not to mention the utility companies themselves, who are mentioned as the initial market.
Unmetered electric service based on "trust me bro" is actually the default (at least in the US) for a huge variety of devices, like streetlights, cell towers mounted to electric poles, public irrigation systems, etc etc.
Almost every US utility has a "UM" process to self-assess an unmetered load's consumption and be billed. So, yes, it's not only viable but widespread.
I mean, if I have to pay them by how much power I draw, I'm pretty glad they have a way to measure that, because I don't.
What's there alternative in this case? If I can land a drone on the power line and suck up some power, they can either charge me when I tell them I did it, or they can not charge me.
B2B transactions like this are handled fine with contracts and lawyers all the time, I doubt it would be an issue. In the worst case, the utility could own the recharging module on the drone, just like they own your power meter.
Presumably they'd be doing inspections for the power company, who probably don't care if some minuscule amounts of power are consumed directly during operations.
Liability is probably the biggest issue, rather than using the energy. If it causes damage because it fails to connect properly, or if it has a trailing wire to pick up other phases (not actually connect to it but to pick up induction)
There is a not so subtle hint in the description that they were mainly inspired by military applications (Air Force, DARPA). Legality doesn’t matter when you’re in enemy territory.
If this works via induction could you even eliminate the need for the drones to land?
Assuming flight conditions are good, there would be a region around the wire (line of charge) with an electric/magnetic field that the drones could use, any shielding notwithstanding.
Coupling falls off rapidly with distance. It's why a thick enough phone case will interfere with magnetic charging, and that's only on the scale of a few millimeters.
Drones consume a lot of power while in flight. You can get a little bit of power out of powerlines standing underneath them with some tricks, but it's not enough to keep a drone in flight. At least not without something prohibitively large to couple to the line.
It's also risky to hover near a stationary object because the longer you hover, the more you're exposed to the risk of a wind gust knocking you into the nearby obstacle.
iirc efficiency loss in wireless energy transmission is exponential? someone correct me. But basically after just a few mm the losses are so great that the amount of electricity needed becomes ridiculously wasteful.
to power a running drone at more than a few inches would be just...a lot.
It's actually an inverse first power law, assuming that the distance of the drone from the wire is much less than the length of the wire. The inverse square law applies only to a point source of electrical energy, to a sphere, or to an object whose largest dimension is much less than the distance to the drone.
Yes, feels like perching via some insulated "feet" and only using energy for stabilisation (as opposed to flight) would allow the drone to get very cloe (and suck much more power) from the line.
It'll definitely charge faster, if only because it's drawing less power to stay up and getting closer. The only question is, is it like 10% or 100% faster?
Maybe this is silly to complain about, but how would power companies charge these drone companies for electricity consumption? Seems… just so easy to steal?
When I was a kid, we had a neighbor who had somehow wrapped some wire around the power lines enough times to siphon off an interesting amount of power. I never saw the wire myself, so it might not have actually happened. But people certainly talked about doing it way back when.
We had a neighbor that put loops of wire under a 600kV line and got enough power to run a water pump for cows. The square law makes that kind of thing pretty low-return.
> Removing battery swaps is the last step to deploy UAVs autonomously at scale.
I can't say, as a citizen, that I'm particularly excited about this.
> Autonomous drones can deliver over 20x the inspection coverage for the same cost.
And we have 20x the manpower to review this footage? I wonder if you're just generating a bunch of data that cannot be practically used.
"More coverage" isn't always the best answer. "Better informed coverage" is probably the problem to solve here. Aside from that what is the maintenance interval on those drones? How does that incorporate into this system?
I think this is solving the problem in the wrong direction.
We do semi-automated analysis of imagery of things like utility right-of-ways and it's pretty scalable. We triage the vast majority of images automatically and then surface a small subset to human experts to review, but it's much, much faster than having the experts be in the field, while having high coverage. (Most images are really boring.)
And in most cases of inspection, you need to look at the stuff anyway, so any cost reduction is very welcome. And if you do increase the total coverage volume while reducing human time, you get a double benefit of being able to provide more granular information, either in time or in space, which can often be useful.
(And as a commenter below notes - this all works pretty well with several year old CNNs. We use a limited amount of image-LLM stuff to surface things zero-shot, but a lot of what we end up doing is a very conventional classifier with a lot of engineering work to make it very fast for the experts to see only the important things.)
> And we have 20x the manpower to review this footage?
I was involved with a startup that did inspections of power generation windmills. The computer vision anomaly detection was really good and that was about five years ago. The goal was to have the automated visual inspections route images with suspected anomalies to humans for review and it was working well the last time I heard.
Compared to having a human who needs to rappel down to the blades for a manual inspection, this is a huge productivity and safety boost.
I've worked on similar systems for oil & gas that combined hyperspectral imaging and LIDAR. The analysis of data collected by drones was fully automated. It was at least as effective as humans at detecting anomalies (something which was thoroughly verified prior to adoption).
The more thorough coverage, potential issues being detected much earlier, and increased automation greatly reduced the total manpower required. Humans only came into the picture when the drones found a problem that needed mitigation. Humans have long been the bottleneck for finding operational risks and issues before they turn into a headline. The more humans you can remove from that loop the bigger the win.
I live in the PNW of the US where many fires have been started by transformers exploding or whatnot.
Basically every community that has a fire as a result of transmission lines rebuilds them above ground/on poles. Just last month I was going through Detroit, Oregon and their 2-3 year old power lines were all down because of the wind storm. Detroit had a transformer explode a few years ago and it took out much of the community. They immediately rebuilt above ground.
So was rural electrification to start with, but it still more than paid for itself. It has also never been easier to bury lines with horizontal boring machines.
As long as they can cover the liability when inevitably one of this sets off a wildfire in California that costs Billions. (I think it’s a cool idea if done right, they are after all trying to fix the problem of wildfires anyway which makes me hopeful)
if you don't have a contract with the utility you almost certainly violate the law, at least in Europe, but then again I don't know what the US regulation is
Could robo taxis steal the idea and get recharged w/o going back to base station? They can eject a rod similar to an E-train or how a military plane get refilled.
This seems like a solution looking for a problem. If the power companies were satisfied about the safety issues of this idea, which are obviously many and severe, then they would also be satisfied with simply mounting a charging platform for the drone atop the tower or pole where the HV line runs, which cuts out a lot of the uncertainty regarding the drone's autonomous guidance.
Also, the whole idea triggers my reflexive skepticism about any technology that seeks to remove the last human from the system. Usually there are exponentially increasing costs to removing the next human, and at some point it's not worth it. People want (wanted) to make sealed, autonomous data centers maintained by robots and it just isn't worth it. Even in manufacturing where robotic automation is ubiquitous and advanced there are still tons of humans.
Hopefully all of the perfect engineering candidates who previously worked in this specific space (autonomous line inspection at an industrial scale) live in San Francisco……………
Seems odd that this would be all in-person roles. Not the most apparent path to relevant talent.
I would guess that its just a coil thats near the cable. however I can't imagine its actually that much power, as the field isn't that strong.
_but_ at 33kv, you start to get corona discharge, which gives you a potential difference, so that might be the mechanism they get power.
(thats a wild guess. )
Edit I did some more digging, induction appears to be magnetic induction, long cables do produce a magnetic field, but not that much (unless you put it in a coil)
Capacitive coupling is much more effective at high voltage.
> The bulb is a prototype "fluorescent" light he invented consisting of a partially evacuated glass bulb with a single metal electrode. Nearby, probably behind the curtain, there is one of his Tesla coil high voltage oscillators which produces a radio frequency electric field. The electric field ionizes the gas in the bulb, causing it to glow similar to a neon light. This photo is a 2 second time exposure taken by the light of the bulb
Yes. That should be done along hiking and biking trails under power lines. There's one in Silicon Valley along the bay shore line. The fluorescent tubes don't wear out; the filaments at the end are not in use. Just slip them inside polycarbonate tubes.
Go ~X speed for Y distance(+/-) on Z heading until you reach A landmark and then start a new set of instructions.
Dead reckoning via inertial sensors, cameras, etc are way to complex for the flight controllers without heavier hardware since theyre hugely inefficient.
AI at the sophistication to do this stuff is essentially bloatware. Like running electron instead of a bare metal gui.
There has been lots of work to make fibre connected drones, so that they can't be located as easily (also the pilot)
A friend of mine worked on a covert comms system for the Rangers to use in the Battle For Berlin that thankfully never occurred. The idea was to clamp onto plumbing, fences, and similar infrastructure where possible. Nodes with radios handed off to other comms systems. It worked reasonably well in tests but I don't know that it went anywhere, point is that the theory is sound.
So the limit isnt batteries, its fiber spools.
https://www.youtube.com/watch?v=C-uekD6VTIQ has a video of their drone on a power line.
"'Vampire drone' can leech electricity from power lines to live forever" https://news.ycombinator.com/item?id=40317484
"Autonomous Overhead Powerline Recharging for Uninterrupted Drone Operations" https://news.ycombinator.com/item?id=39945733
Yes, I'm sure the markup would be large as a percentage, but for most customers the convenience would be worth it. Most of the customers are probably commercial and don't want to risk getting banned or sued.
“Trust me, bro!” is something I wish my power company would do, but they installed a meter instead.
If it's the latter then hand editing is all it takes to create fraud.
Almost every US utility has a "UM" process to self-assess an unmetered load's consumption and be billed. So, yes, it's not only viable but widespread.
I wouldn't talk too loud about this or you will ruin it for all of us. If I discover the street lights on my street mine botcoins I will blame you.
What's there alternative in this case? If I can land a drone on the power line and suck up some power, they can either charge me when I tell them I did it, or they can not charge me.
Assuming flight conditions are good, there would be a region around the wire (line of charge) with an electric/magnetic field that the drones could use, any shielding notwithstanding.
Drones consume a lot of power while in flight. You can get a little bit of power out of powerlines standing underneath them with some tricks, but it's not enough to keep a drone in flight. At least not without something prohibitively large to couple to the line.
It's also risky to hover near a stationary object because the longer you hover, the more you're exposed to the risk of a wind gust knocking you into the nearby obstacle.
to power a running drone at more than a few inches would be just...a lot.
You get a million times more power if you can put a coil 1mm away from the wire than if you hover 1 meter away.
[0] https://user.physics.unc.edu/~deardorf/phys25/rwp/exam1rwpso...
I can't say, as a citizen, that I'm particularly excited about this.
> Autonomous drones can deliver over 20x the inspection coverage for the same cost.
And we have 20x the manpower to review this footage? I wonder if you're just generating a bunch of data that cannot be practically used.
"More coverage" isn't always the best answer. "Better informed coverage" is probably the problem to solve here. Aside from that what is the maintenance interval on those drones? How does that incorporate into this system?
I think this is solving the problem in the wrong direction.
And in most cases of inspection, you need to look at the stuff anyway, so any cost reduction is very welcome. And if you do increase the total coverage volume while reducing human time, you get a double benefit of being able to provide more granular information, either in time or in space, which can often be useful.
(And as a commenter below notes - this all works pretty well with several year old CNNs. We use a limited amount of image-LLM stuff to surface things zero-shot, but a lot of what we end up doing is a very conventional classifier with a lot of engineering work to make it very fast for the experts to see only the important things.)
I was involved with a startup that did inspections of power generation windmills. The computer vision anomaly detection was really good and that was about five years ago. The goal was to have the automated visual inspections route images with suspected anomalies to humans for review and it was working well the last time I heard.
Compared to having a human who needs to rappel down to the blades for a manual inspection, this is a huge productivity and safety boost.
I've worked on similar systems for oil & gas that combined hyperspectral imaging and LIDAR. The analysis of data collected by drones was fully automated. It was at least as effective as humans at detecting anomalies (something which was thoroughly verified prior to adoption).
The more thorough coverage, potential issues being detected much earlier, and increased automation greatly reduced the total manpower required. Humans only came into the picture when the drones found a problem that needed mitigation. Humans have long been the bottleneck for finding operational risks and issues before they turn into a headline. The more humans you can remove from that loop the bigger the win.
This was years ago, the tech has only improved.
I live in the PNW of the US where many fires have been started by transformers exploding or whatnot.
Basically every community that has a fire as a result of transmission lines rebuilds them above ground/on poles. Just last month I was going through Detroit, Oregon and their 2-3 year old power lines were all down because of the wind storm. Detroit had a transformer explode a few years ago and it took out much of the community. They immediately rebuilt above ground.
They'll rebuild them on poles again.
So ubiquitous surveillance, literally overhead, without any need to have a nearby/local charging/physical-management station/crew?
> After power companies, we will service rail, road, telecom, real estate and other inspection markets.
Oh?
> After building drones for the Air Force and DARPA, ...
Oh
2. Find more fires
3. Profit
Like any good startup.
Sounds like having contracts with the power companies is their #1 goal.
Could robo taxis steal the idea and get recharged w/o going back to base station? They can eject a rod similar to an E-train or how a military plane get refilled.
https://youtu.be/3VEkzweBJPM?si=_5hCR8AE1Ii3sI-l
You mean like birds have been doing for decades?! OPEN YOUR EYES, SHEEPLE! BIRDS AREN'T REAL!
/s (I really shouldnt need a /s, but people these days believe anything...)
Also, the whole idea triggers my reflexive skepticism about any technology that seeks to remove the last human from the system. Usually there are exponentially increasing costs to removing the next human, and at some point it's not worth it. People want (wanted) to make sealed, autonomous data centers maintained by robots and it just isn't worth it. Even in manufacturing where robotic automation is ubiquitous and advanced there are still tons of humans.
Seems odd that this would be all in-person roles. Not the most apparent path to relevant talent.
Do the drones need transformers? Won't those be heavy?
Don't the drones need a way to circulate the electricity? IE have a path to ground?
Is it harvesting energy from the magnetic field (via induction), or does it extracts its energy from the electric field instead?
And does the drone just happen to land on the power line for saving energy while doing so, or is the contact necessary somehow?
Some lamps do use induction, like this one[2].
I guess it uses induction looking at the drone, it also works with an open coil.
[1]: https://en.wikipedia.org/wiki/Balisor
[2]: https://www.delta-box.com/en/marking/bht-marking-light-for-h...
_but_ at 33kv, you start to get corona discharge, which gives you a potential difference, so that might be the mechanism they get power.
(thats a wild guess. )
Edit I did some more digging, induction appears to be magnetic induction, long cables do produce a magnetic field, but not that much (unless you put it in a coil)
Capacitive coupling is much more effective at high voltage.
Remember that the primary and secondary side of transformers never touch.
https://commons.wikimedia.org/wiki/File:Nikola_Tesla_holding...
> The bulb is a prototype "fluorescent" light he invented consisting of a partially evacuated glass bulb with a single metal electrode. Nearby, probably behind the curtain, there is one of his Tesla coil high voltage oscillators which produces a radio frequency electric field. The electric field ionizes the gas in the bulb, causing it to glow similar to a neon light. This photo is a 2 second time exposure taken by the light of the bulb
https://www.bbc.co.uk/programmes/p00l1slw < although its not available at the moment