November 19, 2020
The future of warfare is heading towards a battle between software engineers rather than soldiers and pilots. World superpowers are pursuing an arms race to develop super swarm drones, which some have identified as weapons of mass destruction (WMD). In this podcast technical consultant, journalist, and author David Hambling talks about the state of the art of military drones and distributed approaches in warfare.
Samuele Lilliu (SL). David, thank you very much for joining me in this podcast. I want to introduce you a little bit. David Hambling is a journalist and author based in South London specializing in science, technology and strange phenomena. You're also a physicist, right?
David Hambling (DH). Yeah, that's my background...my degree, which was a very long time ago now was in physics. Yes.
SL. The first time we spoke was this summer. That was when I came across a series of articles written by you, while I was researching on topics reported by Leslie Kean on the New York Times. There were some articles on the Pentagon UFO videos, which I then discussed in my podcast. And then with the chat this summer, where you told me about an article you wrote on "USA Navy laser creates plasma UFOs". What was that patent about?
DH. Yeah, indeed, this is part of a long term effort to protect aircraft against heat seeking missiles. There's been several technologies which have been used in the past, particularly flares, the idea is you eject a hot flare and that then distracts a missile away from the aircraft. There's also towed decoys, which are laid out on long cable, but obviously, those both have their disadvantages. So the idea was to use a laser to create an object midair that would attract a heat seeking missile. They use what's known as a laser induced plasma filament. The idea is that if you have a high enough intensity laser, it can actually cause the air to glow. By focusing the laser correctly, you can have that occurred at a distance of 10s or hundreds of meters away. By using a series of laser pulses, you could create a shape midair. In the instance of the Navy patent, the idea was that you could create essentially a ghost airplane that would look more like an aircraft in the real target and would decoy incoming missiles away from it. But of course, this technology would have lots of other applications in that, if you happen to want to create a glowing ball of plasma that looked like UFO.
SL. So basically, you can do very weird maneuvers with this kind of technology. You could have something that moves like zigzag...
DH. Yeah, it doesn't need to obey the same laws as a physical object, it's more like a shadow. If you shine the beam of a torch or something, you can have that circle of light moving around wherever you want. It's exactly the same kind of thing. But that was only the latest patent. The American military had been working on related technology for about 40 years and the Russians possibly even slightly longer than that, though they were using microwaves rather than lasers.
SL. Because you're ionizing the air, then you should also be able to pick up maybe an infrared signal from this kind of ghost image. So the thing would appear in the detection systems of aircraft.
DH. Yeah, well, another interesting thing is that plasma is radio opaque, it reflects radio waves. So you could have something that would show up on both radar and infrared imaging, and which was also visible as a ball of light. So you could create something that looked very much like a real solid object, except there's really nothing there and it can just appear or vanish, as UFOs tend to.
SL. So this would be something of this world, rather than out of this world.
DH. Very much so. I mean, there's an even more bizarre version of this they worked on, which is known as… sometimes called the “Voice of God”. It's not a continuous ball, it's actually a series of flickering plasma that comes and goes, and every time it appears, you get a slight sound. The idea was that by modulating the frequency to disappears, you could actually carry a modulated signal. So you could have a human voice being carried by it. So what you have is effectively a talking fireball. In the very original idea of this, and this is going back about 30 years, the idea was to overall people with this, who would think that it was something spooky or supernatural. There is now a version of that which has been developed by the US Marine Corps as a means of conveying voice commands long range that they demonstrated. It's not very good, but it shows what the technology will ultimately be capable of.
SL. That's amazing if we want to be a cult leader.
DH. Yeah [sarcastic], I think that's the kind of thing they might have had in mind.
SL. So you also spoke about other instances of unidentified objects. There have been reports of UFOs flying around restricted areas, such as nuclear power plants. You wrote an article about a case of five or six drones flying over the Palo Verde nuclear power plant in the US. That was in September 2019. Now, the strange thing is that these drones were flying longer than 30 minutes, which is weird for commercial drones. So what do you think was happening there?
DH. Well, from the size of them, they were reported as being more than three feet across. So these are obviously much bigger than the normal consumer drones, but well within the size range of commercial drones, that sorts that are used for carrying cameras. But rather than just being a few hundred dollars, drones like that are several thousand dollars. Clearly someone had the money to be flying a fleet of those over Palo Verde. They hung around and they apparently surveyed the place for something like half an hour, and then they went away. The security forces were unable to ascertain where they'd come from. So we have no idea who is behind that. But the equally strange thing was the next night, they came back and did the same thing again. That's not the only nuclear power plant. The Freedom of Information Act requests have shown that there have literally been dozens of incursions like that, sometimes with one drone, occasionally with three or more at the same time, over the nuclear power plants all across the US. The vast majority of these are unsolved, and nobody knows who's behind them, or what they're doing.
SL. So maybe it could be some other adversary power trying to test the security of these facilities, or...
DH. A lot of people assume it's some arm of the US government carrying out its own test, just see what happens. But it's clearly someone with money. So it's not just a few hobbyists messing around, because it's a lot of drones. It's a lot of expensive drones. And it's all the way across the US. So there's clearly something rather bigger going on. Though, inevitably, I did actually have someone email me and say "No, they're not drones, they're UFOs".
SL. So it would be a sort of penetration test. If you're talking about network security.
DH. Yes, exactly. If it could well be something like that. One of the curious things is that the nuclear authorities have said that there is no threat from the drones. They are doing very little about taking additional security measures, they don't seem bothered at all, because their theory is that nuclear power plants are protected by several feet of concrete. So they don't see that there's any real danger, which may turn out to be a very naive view.
SL. We are talking about drones, but there are different types of drones, you can have flying drones, you can have drones on sea and land. How would you define a drone? What's a drone?
DH. A lot of people don't like the term at all. But you could use it to apply to any unmanned or remotely controlled system.
SL. Yeah. You wrote a great book, I read your book called "Swarm Troopers, How Small Drones will Conquer the World". It's a great work. And you make some very good points about the military, drone technology and future developments. Can you briefly tell me what was this book about?
DH. At the moment the drones are beginning to have quite significant impact on warfare. At the moment, the chief users are the US Air Force and the CIA who've been using them for what are called counterinsurgency operations. So that's very low level conflicts, where you're dealing with insurgents and terrorists. That's well short of an actual hot war. But we've also seen very recently in Nagorno Karabakh, where there've been drones making a significant impact on the battlefield on a full scale war. And that's kind of likely to happen more in the future. Now, the big drones, like the Reapers that the US Air Force us go for $20-30 million apiece, whereas small drones, like the consumer drones made by DJI, the sort that you can buy over the internet, those are only a few hundred dollars. So you could literally buy 10s of thousands of these for the cost of one Reaper. And you can do a lot of damage with 10,000 small drones.
SL. Yeah. So you mentioned the Reaper and you said it's very expensive. I mean, a Reaper would cost, you said $20 millions, it can be even more than that. So it's about a third or a quarter of an F-35 fighter jet, right?
DH. I've had some interesting discussions with General Atomics who make it about what the exact cost is and it does get very complicated. But somewhere well over that. The UAE have just bought some. I would have to check how much they're paying a unit, but I bet you it's a lot more than 30 million a time.
SL. I guess you can customize them depending on what your needs are...
DH. Yes, and that doesn't make them any cheaper either.
SL. What's the state of the art of drones, big drones, at the moment?
DH. So the big drone, something like the MQ-9 Reaper is about the most advanced thing there is out there. In terms of military aircraft, it's very low performance, it only goes at about 300 miles an hour. It's not stealthy. It's really more like a World War II aircraft in terms of performance. But what it really has got is endurance, it can circle over a target area for a very long time with endurance of 40 hours or more.  Now, going ahead from this, they're looking at stealthy jet fighter-type drones. But that's still some way ahead.
SL. And in terms of cost you can keep improving and improving these drones until they get so complicated and so expensive that basically... there was a guy that wrote a set of rules...
DH. These are Augustine's laws, yes...
SL. ...by 2050 for the Air Force and Navy will only be able to afford a single aircraft between them.
DH. In many ways his he worked on both sides of the defense business, both as a contractor and for the military. And he saw that there is this inexorable rise of price of military systems that there is a tendency for them to just shoot up in price, partly because the government will always pay and there's no incentive to keep costs down. One of the interesting things is that that's in stark contrast to consumer electronics, where things do tend to get cheaper over time and you get increasingly more capability for the same price year on year. When you compare those two things side by side, one of them is going to tend to lose.
SL. There seems to be a solution to this large and expensive hardware. In fact, rather than focusing on the development of single offensive and highly sophisticated machines, there are people talking about distributed approaches. You wrote an article about a talk by Timothy Grayson from DARPA, where he spoke about Mosaic Warfare. He uses this mosaic as opposed to jigsaw. Can you tell me briefly what's this concept of mosaic warfare (and distributed approaches)?
DH. The idea is that you can have your warfighting capability. Particularly for him, what he's talking about is sensing. So that's the ability to see what's happening, cameras and radar and other sensors, and your communications and then your weapon systems. Rather than having it all in one big box, like, say, a single aircraft flying overhead, you can have lots of different devices at different levels. So you could have unmanned sensors on the ground, you could have small ground vehicles, you could have small drones, and you could have larger systems orbiting overhead, you can build up this whole collection of things. Then, depending on the mission, you can slot in more at whatever level of are needed. So you can have more radar sensors, or more thermal sensors, or more acoustic sensors. Like a mosaic, you can build it up to match whatever the particular mission requirement is.
SL. And instead the jigsaw is not flexible at all, because you can only place a piece of jigsaw...
DH. The problem is the pieces only slot together in one way. So you've got to have this piece, and then you've got to have this piece, whereas with mosaic everything is like open source, everything can plug into everything else and it all networks very tightly. That gives a lot of redundancy. So if you lose one bit of it, you can work around that. Whereas with a jigsaw, when you lose one piece, the whole thing's ruined.
SL. I think one of the main problems with this mosaic warfare is that when you have different pieces of hardware, it's very hard to make them communicate to each other because they have communication infrastructure and software that is very different, that is not done to communicate with other hardware pieces. For example, you can have the situation where you have two airplanes made by two different companies and there is no way they can talk to each other.
DH. Unfortunately a certain amount of that is because that's the way the aerospace industry works. As with computers, nobody there wants it to be open source because what they would much prefer is to lock customers into buying their equipment. So once you buy the aircraft, you have to buy all the ground communications equipment to go along with it. Then you're incentivized to buy the missiles from the same company because the missiles can talk to the aircraft and everything else goes from there. As soon as you go open source and have everything communicating to everything, it destroys the whole marketing strategy.
SL. The problem then if you have open source is also that you might be able to hack the systems, isn't it?
DH. It shouldn't. As long as you've got a decent firewall between you and the outside, it shouldn't really make any difference. But security is a huge issue and that does tend to lie at the center of a lot of drone communications activities.
SL. So basically, the concept with this distributed approach is that instead of having big devices, big airplanes, big whatever, you could have smaller devices that work together. And you can dispose, you can sacrifice one or two, or 10, or whatever, the system will still work. You spoke about that. Another similar concept is also the Ocean of Things.
DH. The Ocean of Things is a program by DARPA. That's a very distributed sensing approach. So again, rather than having one or two large sensing devices, the idea is that you have thousands of these floats, which are distributed across the ocean and each float has a number of sensors on it. Mainly it's like a floating mobile phone, because it's got the mobile phone type capabilities in terms of having GPS and communications and the camera and tilt meter and things like that. But it's also got an underwater microphone on it. So it's able to detect any vessels going by, it's able to see things under the water. It will not communicate with the other floats nearby. One of the things, one of the unusual things about the Ocean of Things is that all their data is going back to a satellite. They're not networking it because they found that was too power hungry. So all the sensors in the Ocean of Things bounce the signals back to a satellite. And then a cloud server puts together all the data from all of them and builds up a picture about what's happening on, above, and under the ocean.
SL. And you also wrote about another type of underwater device that works more or less as an RFID, but using sound waves...
DH. ...using sound rather than radio waves. Yeah, that's a new backscatter device. The interesting thing about that is that it's basically harvesting the ambient sound from the ocean, the acoustic background noise, and it's then using that to charge its batteries, and also to transmit signals out. So it's a battery-less system, it doesn't require any external power source, it's just powered by the ambient sound. That makes it very rugged and very reliable. In theory, it means it can sit on the seabed or on a float indefinitely, and just keep on working. Again, that can be used for communications, and you can network a load of them together to communicate over long distance. You can use those as sonar sensors. Because they can be made cheaply, the idea is that you could sow the sea with many thousands of these, and they would all connect up and they would then combine to produce an image of whatever was traveling around, on, and under the ocean.
SL. And you can use them for multiple purposes. For example, monitoring the sea temperature, for climate change, you can use it to monitor the animals in the sea, not only for military purposes. So there are multiple applications for this.
DH. Absolutely. That's one of the big arguments for ocean of things as well that it will be collecting a lot of weather and oceanographic data. DARPA are making efforts to make sure that all gets shared with scientific organizations across the world. They're making that all very open source, which in one sense, is very nice of them. But in another sense, it gives perfect cover. So it allows them to put their sensors everywhere and say “No, this isn't really military spying on you”. This is just harmless scientific sensors, which, interestingly, is something that the military have been doing for a very long time.
SL. So there is a long history of military technology being transferred to civilian applications...
DH. There is that that's as well, but there's also a long history of the military, using... For example in the 1950s, they sent balloons with cameras floating over the Soviet Union, which they claimed were just for sensing cloud formations and weather monitoring, but which are actually trying to take pictures of Russian nuclear installations.[13,14]
SL. So the opposite in this case.
DH. Well, not totally dissimilar.
SL. What's the advantage of having small portable and cheaper drones for military applications?
DH. If you're a soldier, one of the things you always want to know is what's happening on the other side of the hill, because the last thing you want to do is actually stick your head up and go over and get shot out, while trying to look. Whereas drones give you the perfect way of seeing everything that's going on around you. You can send a drone up and it can then take video of everything in the surrounding area, you can send it over obstacles, you can send it to the other side of the hill, you can look over walls, you can look down streets, you can look on top of buildings. So small drones are extremely handy as a means of seeing where the opposition is without them being able to see where you are. That's extremely useful. That's why some military are producing their own drones and some, as the Israelis are, are simply buying commercial drones and issuing them to infantry units to give them their own organic air reconnaissance units. So that even at the lowest level soldiers can have their own aerial reconnaissance.
SL. So what do we have at the moment? What are the best, small and portable military drones?
DH. That depends what you mean by we. But one of the most commonly used by the military is the Raven drone used by the US Army. And that's a lot more expensive than a sort of small quadcopter. But it looks like a model aircraft, it's got a wingspan of a bit over a meter and that can fly around for about an hour and a half. And as well as a camera, it's got a thermal imager, so it can get images of what's happening even at the dark or through fog or smoke.
SL. The one you just mentioned, I guess it's a glider. It's a glider drone.
DH. Well, it's powered, right, Yes, effectively.
SL. Yeah. So there are different types of drones. There are quadcopters. And we're more familiar with the quadcopters, because those are the commercial ones, basically. But those ones can only fly for, let's say, 30 minutes or so.
DH. Yeah, that's about it. For a consumer quadcopter 20-30 minutes is usually about the maximum, whereas a fixed wing one can fly for several times longer, and also quite a lot faster. So it gives it considerably more range and endurance.
SL. If we want to improve the flight time of this quadcopters, what do you think we should do?
DH. The problem with any kind of helicopter-type device is that unlike a glider, it has to work very hard just to stay in the air. So it's got to constantly shift a lot of air. That makes it inherently quite inefficient. The ways around that are mainly just find ways of recharging it. One of the most common things used is drone in a box type solution. There are several of these out there at the moment. There's an Israeli company called Airobotics produces so a very effective one. That is effectively a base station on the ground that the drone can land on and recharge itself automatically. So rather than having a very long endurance, it can just keep coming back and recharging itself. There's a number of projects as well, for vehicles. So you will have a ground vehicle with several drones so that it will launch one drone off. When that one runs out of power, it'll send another one off to cover the same area and the first one comes back and recharges. So you have this constant stream of drones going out there. So it can maintain a drone high in the sky indefinitely, even though none of the individual drones have that long of an endurance.
SL. Or you could work with the type of power that drones use. At the moment we have batteries, standard lithium batteries. You wrote an article where you mentioned the CIA Project Aquiline, where, in the 70s, they were planning to use drones propelled by nuclear batteries.
DH. Yes, I mean, that's ultimately that's one of the best ways of powering aircraft. You've to remember, back at the time they were also looking at nuclear powered bombers, and various other things. Given that the military already have nuclear powered submarines and nuclear powered aircraft carriers. It didn't seem quite so outlandish. The nuclear powered drone would be a slightly different prospect though, because there's a much greater chance of it crashing and problems. However, there is actually a company that's working on a similar concept at the moment, who believe that by using betavoltaics, which is a sort of power system driven by radioactive decay, they can get enough power to have a compact power system to drive a drone. So the idea isn't completely dead.
SL. The problem with this nuclear batteries is that they are not very efficient. That's my understanding. I'm not an expert, but they're not very efficient. I think efficiency is something around this 6-5% or something like that.
DH. That's certainly one of the problems. Nuclear batteries, there's only a tiny, tiny handful of applications that they're actually used for. They're used for satellites and space probes. The Russians bizarrely used them for lighthouses out in very distant locations out in Siberia, which didn't end well. One of them got attacked by some people who were scavenging for scrap metal. So the Russians then had to track down some highly radioactive material that had been stolen when they stripped this place.
SL. Drones, different drones have different capabilities. I mean, if you think about the fixed wing drone, it can fly for a long time, but the amount of maneuvers that it can do is very limited. But if you think about quadcopters, quadcopters can go in between trees, they can go inside the forest, they can go in, even inside houses and buildings. But now they're trying to expand this capabilities and build things like perch and stare drones, like the Spiderman drone.
DH. With quadcopters, yes, you're quite right. The ability to go inside buildings and to maneuver around indoor spaces, is a key advantage. Not just for the military. I recently wrote a piece about drones that are being used in mines. They're being used to travel through old mine workings and spaces that are unsafe for human exploration. But you can have a drone that can wander around and map and see there and test the gases. One of the things is, because there's no radio communications on the ground, the drone has to be smart enough to be able to find its way and explore autonomously. Then as you say, perch and stare, that's a good way of coping with the limited endurance of a quadcopter drone. By simply finding a flat surface and sitting there, it can remain as basically like a fixed CCTV, and that way, it doesn't actually use any power.
SL. And you could also have sensors to see maybe what's happening inside the building, what's happening inside the room...
DH. There are indeed some... there's through the wall radar, which allows them to basically pick up movement on the other side of a wall. The so called Life-Detection Radar, which is actually triggered by heartbeat. So they can see whether there are living people inside. One of the uses of this would be in an emergency situation. So you could tell whether there are people who needed rescuing inside a burning building. Obviously, the military are also very interested in this as a means of determining whether there are potential insurgents or other people inside a building that they're planning to clear.
SL. And there are also the types of drones combined with missiles, the so-called Kamikaze drones. There is a Switchblade 600, if I remember well...
DH. Yeah, the Switchblade has been around for about 10 years, which is used by American Special Forces. It's a small Tube Launched drone. You fire it from a little thing, like a bazooka, but it's very small, it only weighs a couple of kilos. And then after you launch it, it pops out wings, and that can then fly around again, beaming back video, and you use it to locate a target. When you do find the target, you line it up in the crosshairs and press the button. The drone then locks in on that and crashes into him and destroys it with a target with a warhead, which is basically like a hand grenade.
SL. Okay, so I'd like to talk about a little bit about China. So China is a big topic. It has been a big topic since the beginning of the trade war with the US. There are tensions in the Pacific Ocean, and also, more recently, with the pandemic... So the predominance of China, in terms of commercial drones, and also military drones, do you think it represents a risk for the West?
DH. I think it represents a problem in the sense that because these drones are very cheap and convenient, a lot of militaries have been using them. And also a lot of insurgents have been using them. I think the biggest surprise was when the ISIS insurgents started using Chinese drones to drop grenades on Iraq in American forces in Mosul. Obviously, anyone can use them. It's very advanced technology. It's very convenient. But the problem is that DJI, the company who makes these things, is essentially answerable to the Chinese government. So any data that gets picked up by those drones, DJI may be able to hack, DJI may be able to tell where drones are, there is potentially a huge, huge security risk there. So the American government, among others, has told everyone to stop buying DJI drones while they tried to find an alternative.
SL. There were similar problems with Chinese mobile phones about data breach...
DH. Exactly. And there's the same issue about whether it's safe for Huawei to be part of our national communications infrastructure, because, in theory, the Chinese government could demand that Huawei cooperate with them.
SL. Parrot was recently selected by the USA Defense Innovation Unit as a major drone supplier.
DH. Well, they've selected a whole load of them, they've now opened up a new standard called Blue sUAS, which is Small Unmanned Aerial Systems. The idea is, rather than being relying on a single supplier, they're going to have just a, basically an open source system, which all different suppliers will produce hardware that will fit into. So it's a sincere effort to get themselves out of this historic problem of being locked into large military contractors. And one of the big problems is that unfortunately, none of the companies involved has anything like the industrial cloud or the R&D base, or the experience of a company like DJI.
SL. I think when in the West, in, let's say, USA, or even UK, Europe, we have capabilities to build highly sophisticated small drones. I think the main problem is the cost of labor, and maybe procurement. I mean, Chinese can leverage the very cheap labor cost and they can procure materials and electronics so easily because everything is made there. I mean, if you think about Shenzhen, and everything is there.
DH. Yeah, one of the big advantages of building in Shenzhen is that they are close to lots of other Chinese electronics companies. DJI, from the start, realized that the best way to do it was rather than just building a drone, and then sticking a camera onto it, and sticking communications onto it, was to have everything totally integrated. They were able to go to manufacturers, and they actually have chips that are specially built to their specification for their drones, by chip makers. That's partly because they're doing it in such huge volumes, because they're ordering a million at a time, and partly because they are closely integrated under the Chinese system. So they they've got companies who they can work with them who they know and can trust and it's very difficult to get that outside of China. And equally right from the start they threw a huge amount of research on it. DJI has about 1500 people in their research and development department, which is just massively greater than any small startup is ever going to be able to match.
SL. Yeah. Do you fly drones yourself? Are you a drone pilot?
DH. I'm not a real drone pilot. No, I just fly occasionally toy drone.
SL. Yeah, I've got an Inspire 1 here.
DH. Okay, have you found it?
SL. The thing is that, when you get DJI drones, if you if you are interested in filming and all these things, that's the best stuff you can get out there. Now you can get an Inspire 2. With the Inspire 1 that I have, you can film in 4K, RAW 12-bit. With the Inspire 2, you can even film with the better cameras. I think they even have full frame sensor. And you can use this stuff for films. If you want to go for something better, you can go for a Matrice 600 and you can mount a RED camera command a big Canon camera there. But if you go for Parrot, there is nothing like that. That's the problem.
DH. No, and I think Parrot will have trouble competing with them in the open market. I mean, it's interesting that their nearest American competitor 3D Robotics, went out of business or went out of the drone business several years ago. It's just very difficult to compete with people like DJI.
SL. The only thing I found annoying was the customer service, which is totally rubbish. That's totally rubbish.
DH. But they are a huge success. Because they give you a capability, which people just didn't have before. Before, the only way you would have to get that kind of shot would be to rent a helicopter or, in some cases, to rent a crane. And those are just fantastic amounts of money. Whereas these days, the world is full of amateurs who are able to shoot these incredible movies with overhead imagery and drones flying over buildings and through forests and other things which would simply be impossible without this type of technology.
SL. Have you heard about those people doing experiments with insects to control how insects fly?
DH. Yeah, the American military have been pursuing this for quite a few years. It all looks fairly gruesome, but part of the idea is that you can insert a microchip controller into a chrysalis and then, as the adult insect forms, the chip will have connections inside its brain. They have been able, by using electrical impulses, to steer an insect to the left, to the right. It's not exactly remote control, but it gives some idea of the kind of thing that they're after. It's all at a fairly basic stage, but it's an indication of an interest more than anything else. I don't think there's any actual capability out there. What's more interesting, in some ways, is what's known as biomimetics, which is copying biological systems for drones. So, all the way far as back as the 1970s, they the CIA developed a drone, which was essentially a copy of a dragon fly. That was able to fly and carry a camera. They didn't know how the aerodynamics worked or anything, they literally just copied a dragon fly. Similar approaches have been used more recently, and they're getting a much greater understanding of how the flight dynamics work. But in some cases, you're seeing quite convergence. So you're now looking at drones that look a lot more like living things. In China, allegedly, the government does use spy drones that look just like birds. But given how much the Chinese government spies on its population anyway, I'm not sure whether that's really necessary.
SL. Okay. In terms of unconventional drones, there are also other types of unconventional drones, like boat drones or USVs. And I think you wrote an article about the USVs operating in Australia... but then there were also USVs operating in Iran, I guess.
DH. Yeah, I mean, lots of people are. Similarly, with aerial drones, the technology is now there so it's now possible to exploit that. So you can have something that can do a mission, which would previously have taken a manned boat with several crew, which is an expensive piece of kit with all the running costs. And you can just have a small unmanned boat, doing the same task and sitting out there in the ocean for days or weeks or months to do the same thing, and then communicating back generally via satellite. So that's a capability you're seeing, not just with America, and Britain and China, but also with all sorts of countries like Australia and Turkey and Israel. And in the very near future, everyone is going to have their own robot boats.
SL. And so before you mentioned, in terms of non-commercial drones, drones used by terrorists, you mentioned the ISIS in Iraq, trying to retrofit, I think it was Phantom drones, with the grenade drop systems. But there is also the case of the Houthis from Yemen, which successfully attacked Saudi Arabia with the homemade drones.
DH. Yeah, the Abqaiq oil processing facility in Saudi Arabia was effectively put out of action, by an attack by 20 or so of these essentially garage built drones, which came from the Houthi rebels in Yemen. They clearly had a lot of help from the Iranians. There was an engineering and support from that side. But these are basically very crude, very cheap garage built devices that are capable of carrying out a strike over hundreds of kilometers and doing tremendous industrial damage. That was a bit of a wakeup call as to just how much of a threat there were, particularly because the facility was defended by some very modern air defenses, which completely failed to deal with the attack.
SL. Yeah, then systems like, I mean, in Israel, they have the Iron Dome. That wouldn't be effective against these small drones. I don't think so because those are too small probably, isn't it?
DH. Iron Dome is actually one of the better ones, Iron Dome is quite well geared up to deal with a range of different threats. But something like the American Patriot System, which was one of the ones that had at Abqaiq, it's only really geared to deal with fast, high flying threats and threats like cruise missiles. Something that's very small and slow, and looks like a bird on radar, it wouldn't be so good at. And that's one of the big problems they're having.
SL. So what do you think... what would be an effective strategy to disarm small drones trying to cross a border? What would you use the?
DH. Well there is a, you might say, a mad scramble among the defense establishments of the world at the moment to come up with effective counter drone defenses, particularly in what they call class one, class two, these very small quadcopters and fixed wing drones. The problem is that they're actually very difficult to shoot down, because they're very difficult to see with just with a machine gun or a rifle or something like that. It's possible to shoot them down with missiles, but missiles are very expensive. On one occasion, we believe the Israelis did take out a small quadcopter with a $3 million Patriot missile, which is clearly not a viable long term approach. Jamming is very effective, because all you need to do is break the link between the operator and the drone. With the DJI ones, they will normally automatically returned to base. But if you will also jam the GPS, so it doesn't know where it is, it will just land softly on the ground. So jamming is actually quite cheap and quite effective and there's a lot of these drone gun things being marketed for defenses for this kind of low level threat for something that's simply a commercial drone using standard radio communications. But that type of jamming tends not to work against military grade drones, which have got much more robust communications, which can't be blocked that easily. Then moving up from there, you're then moving to the really advanced stuff. Two of the ones in the pipeline are high energy lasers and high energy microwaves. There are a lot of laser systems out there. In a sense, this is a godsend to them, because we've had lasers for 60 years now. But there, they haven't actually managed to reach power levels that can do any damage to a tank or a ship or an airplane. But they are powerful enough to shoot down small drones. So there are lots of 10-20-30 kW systems out there, which are being marketed as anti-drone systems. They do a very good job of that. And they have very high precision, deliver effects at the speed of light, and they are quite deadly against small drones. But they are extremely expensive and there's a real question as to how many drones you're going to be able to shoot down before they overrun you.
SL. And the advantage of directed energy weapons is that you don't need to reload them, you just need the power.
DH. Yeah, you've got an infinite number of shots, it's extremely accurate. So it's a lot easier to shoot down a target two kilometers away than it is with a machine gun. The ammunition is extremely cheap, cost nothing, and it can be linked to a radar or other system to give you very high precision. So at the end, unlike bullets, there's very little risk of collateral damage. One of the systems the Americans uses is a thing called Phalanx, which is a 20 millimeter rapid fire cannon, which is very effective against things like missiles. But the problem is it fires a spray of these 20 millimeter bullets into the distance and those are going to ruin someone's day if they land anywhere near them. Whereas with a laser, it only hits what you aim it at. And it shouldn't do any damage beyond that.
SL. So let's talk about the swarm drones and what are the current capabilities of the West in terms of swarm drones.
DH. A swarm is simply a large number of relatively simple units that are able to carry out complicated coordinated actions just by following a few rules. Swarming technology is a lot of it draws from nature. So this is from the way that flocks of birds fly or that shoals of fish swim, so they can all move together without crashing into each other. And the whole thing can change directions smoothly and seamlessly. So it's that sort of approach. The advantage of a swarm is that you can have 100 or 1000, or even more small drones, all carrying out a task and they don't require hundreds or thousands of operators, you just need one operator to say go over there and the entire swarm can then have a commands. So it's got a huge potential for allowing the deployment of what the military called mass, it means you can have large numbers. So it means you can have large numbers of drones, but with very simple command. And so that's quite appealing.
SL. So what are the technical challenges in the development of swarm drones?
DH. It's getting smart enough swarming algorithms and the way the command and control works is quite tricky, particularly if you want a sub swarm to break off and go off and do something else, or if you want to have a second swarm join in become part of it. So the swarm interface is a big challenge there as to how humans can effectively deal with it. You see a lot of these drone light shows at the moment using 1000, 2000, 3000 drones. That's not actually a real swarm. Because that's all centrally controlled. All the drones there are getting their orders from a central controller. Real swarming is where each of the drones is independent, and is making up its own mind, just based on what its neighbors are doing, that's a bit trickier. But the technology is moving fast. An Israeli company launched their first commercial swarming software. That's for operating swarms of drones to do things like agricultural or industrial surveys or search and rescue. But there's a lot of work in different parts of the American military and the Chinese military, in terms of putting together swarms of drones that will be useful, not just for scouting and for finding the opponent, but also carrying out tanks on them autonomously.
SL. So the main advances that we need are in terms of software, basically, and probably computing power within the drone itself. So the drone itself needs to carry out all the computation required to behave within a swarm. You don't want to use an external communication systems, you don't want to use GPS, you don't want to use a central control station, you want to do everything within the drone.
DH. Exactly, you want to have as much of the processing carried out on the drone as possible. One of the things that's helping with that a lot at the moment is the so-called neuromorphic computing, which again, is copying from biological systems. Buzzwords like Deep Learning get thrown around a lot. That's just a very efficient algorithmic method of doing things like recognizing objects, or learning how to fly, for example, could be done by a deep learning system. Because there's a lot more neuromorphic chips coming onto the market, that's greatly increasing the capability for this type of drone.
SL. We even have them and in some mobile phones, right?
DH. Yeah, exactly. The last two or three Apple phones have had these bionic chips in them, which have a dedicated deep learning processor as part of them. That makes them supposedly thousands or millions of times more efficient than a normal digital computer for doing things like facial recognition, or converting text to speech.
SL. So what are the current capabilities of adversary countries if we can call them adversary countries? I'm talking about Russia and China.
DH. In terms of their swarming drones?
DH. Well, the Chinese military have flown, what they say is the largest ever swarm of small drones. They're certainly looking at using that type of swarm as a way of countering America's advantage in particular items. So they're looking at drone swarms to neutralize American aircraft carriers and air bases, for example. That's highly significant. On the battlefield, the Chinese recently showed video of a barrage drone launcher, which is a vehicle that can fire off up to 48 drones at a time. They say that you can combine several of those so you can launch a swarm of 200 drones all together at an enemy target. With existing air defenses, 200 drones would completely overwhelm them. There's nothing currently that's able to deal with that number of targets at once.
SL. But what happens when the number becomes very high? I'm talking about 10,000 drones or something like that. There was an article that spoke about drones as...
DH. Weapons of mass destruction. Yeah. That's when drones get really scary when you start looking at thousands, 10s of thousands or even hundreds of thousands of drones. In the simplest case, if you just have a drone with an explosive warhead, each drone can kill one person. So a swarm of 10,000 drones can kill 10,000 people. That's vastly more destructive than anything else in the inventory at the moment, simply because it's such an efficient device, a killing machine. There are arguments that drone swarms of that type should be legally counted as weapons of mass destruction, and that there should be controls over them as there are with nuclear, chemical and biological weapons simply because they are so dangerous, and because of the risk of proliferation. Unlike nuclear weapons, which require a very advanced scientific establishment to develop, anyone can develop a drone swarm and simply acquire drones and the software and put explosives on them and build bigger and bigger and more destructive swarms. So the potential is there, even for relatively minor nations to have huge disruptive power capable of killing very large numbers of civilians.
SL. Yeah. And there was a Netflix show about that. I think it was Black Mirror.
DH. Yeah, Black Mirror did and it did an episode about that. There's also a video called Slaughter Bots on YouTube, which was put out by the campaign against autonomous weapons, showing very much this idea and that, if the technology of having small drones that can autonomously track down and kill people gets out there, it will have almost unlimited potential for destruction. Though in fact, simply drones attacking people may not be the worst aspect of it. What would be worse, if you had smart drones, which are able to do things like start fires and attack industrial infrastructure, you could get much more damage done.
SL. Yeah. What are the ideas at the moment in terms of stopping these large swarm drones?
DH. They're very difficult stop one at a time. So individually, shooting them down is challenging. One approach is using high power microwave weapons, which effectively burn out the electronics. Raytheon have a system on that, the US Army have a system for that. These are reasonably effective at destroying electronics, and they've done very well in tests. What we don't know is how easily drones can be shielded against this sort of effect, and the potential for collateral damage. This technology has existed actually, for many years. But one of the reasons why it's never been deployed, is because of the risk that it could harm other electronics. The military is highly dependent on this sort of thing. One reason why they won't ever put a microwave weapon on an aircraft is because of the risk of damaging some of the aircraft's own systems. So there may very well be issues with that. The other way is fighting the swarm with another swarm and having interceptor... there are already some interceptor drones out there. If you had a swarm of them, they would be able to take on an incoming swarm, particularly if they are smarter than the drones coming in and are able to anticipate what they're going to do. So you could have a situation where your defenders were able to very efficiently take out the attacking drones. But equally, you could have a situation, where if the attackers are smarter, they might be able to evade the interceptors. So I think it's going to be very much a software race between attack and defense on there.
SL. Yeah, so it seems that, that's the kind of perception that some people have, the ultimate battleground seems to be between rival software engineers, not between soldiers and pilots basically.
DH. Yeah, it's, it's in many ways it is heading towards that situation. Yes.
SL. Did you see the Alpha Dog Fight Trials?
DH. Yeah. So this was some seeing how well, a piece of software could compete against a human pilot. This is about the third year they run it. And so before, it's just been the different AI systems competing against each other. But this time around, they had the winner of that in a simulation flying an F-16, against a highly trained human pilot in a simulated F-16. The AI won five out of five, rather easily. There's a few questions about that, and whether it's an accurate simulation, and whether that tells us anything about what would happen in the real world. But it certainly looks like the AI has got to a point where it's highly competent at dogfighting.
SL. Everybody talks about AI, I would call it a machine learning because I don't think there is much of intelligence in those things, because those things do what software engineers write them to do. They learn things, but they don't take initiatives. For initiatives, we still have humans.
DH. The argument about whether machine learning counts as AI or in fact, what AI is, is going to go on for a while. I think so. Yeah, I think that's a discussion that will certainly go on. But yes, there's an argument for saying it's not showing any real form of intelligence. It's just very good at what it does.
SL. And there was there was also another thing about this dog fight that they were using F-16. They couldn't use like F-35, because of security. Those F-16 are old fighter jet.
DH. Yeah, it's an aircraft from the 1970s. So they could share all the information about its performance very easily. Whereas with the F-35 or some of the more modern ones, a lot of that is still secret. Another interesting thing about the F-16, is that they've actually converted a number of the old F-16 for unmanned operation. Those are mainly there as aerial targets for air to air combat training. But if you've got a an unmanned computer controlled F-16, you could quite easily put your AI, if you want to call it that, dogfighting pilot in there, and turn that into a robot dog fighter.
SL. You end up in a situation like Terminator or something like that.
DH. Yeah. But it's interesting that there's actually a follow up to the Alpha Dog Fight. Because people were saying that part of the problem was that the human pilot was doing everything he'd been trained to do. Whereas the AI pilots were behaving like video gamers, and they were just using the system to its advantage. So there's actually going to be a rematch between the AI and a human who's the winner of a gaming competition to see whether the game is better than human pilots.
SL. So David, I want to ask you one thing, do you mainly cover stories from USA? Do you cover the UK at all? Or do you do anything about...
DH. Yeah, I cover the world. US, UK, Russia, China. Yeah. But the Americans tend to be the best communicators. Whereas particularly for military stories, it's very difficult getting anything out for the establishment over here.
SL. Okay, David, thank you very much for doing this. I hope next time we won't have any pandemic so we can meet here in Cambridge.
SL. That would be great if we can do that. Thanks very much for having me.
DH. Thank you very much.
 Hambling, D. U.S. Navy Laser Creates Plasma ‘UFOs’. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/05/11/us-navy-laser-creates-plasma-ufos/?sh=642b81111074
 Hambling, D. Court to Defendant: Stop Blasting That Man's Mind! Wired (2009) https://www.wired.com/2009/07/court-to-defendant-stop-blasting-that-mans-mind/
 Hambling, D. ‘Drone Swarm’ Invaded Palo Verde Nuclear Power Plant Last September — Twice. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/07/30/drone-swarm-invaded-palo-verde-nuclear-power-plant/?sh=6941bd4743de
 Hambling, D. Dozens More Mystery Drone Incursions Over U.S. Nuclear Power Plants Revealed. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/09/07/dozens-more-drone-incursions-over-us-nuclear-power-plants-revealed/?sh=109c12f66296
 Hambling, D. Swarm Troopers: How small drones will conquer the world. (20156).
 Hambling, D. The ‘Magic Bullet’ Drones Behind Azerbaijan’s Victory Over Armenia. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/11/10/the-magic-bullet-drones-behind--azerbaijans-victory-over-armenia/?sh=2e258b7a5e57
 Hambling, D. Why The Air Force Needs A Cheaper Reaper. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/06/10/why-the-air-force-needs-a-cheaper-reaper/?sh=5d6ac08946ff
 Hambling, D. Why Selling Reaper Drones To UAE Is More Controversial Than F-35s. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/11/12/why-selling-reaper-drones-to-uae-is-more-controversial-than-f-35s/?sh=3b28615825f2
 Force, U. S. A. MQ-9 Reaper (2015) https://www.af.mil/About-Us/Fact-Sheets/Display/Article/104470/mq-9-reaper/
 Wikipedia. Augustine's laws (2020) https://en.wikipedia.org/wiki/Augustine%27s_laws
 Hambling, D. DARPA Progress With ‘Ocean Of Things’ All-Seeing Eye On The High Seas. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/08/13/darpas-ocean-of-things-is-an-all-seeing-eye-on-the-high-seas/?sh=2cbbec6df270
 Matheson, R. Wireless communication breaks through water-air barrier. MIT Media Lab (https://www.media.mit.edu/articles/wireless-communication-breaks-through-water-air-barrier/) https://www.media.mit.edu/articles/wireless-communication-breaks-through-water-air-barrier/
 Hambling, D. Stratospheric Balloons Will Rain Tiny Electronic Spies From The Sky. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/07/08/stratospheric-balloons-will-rain-tiny-electronic-spies-from-the-sky/?sh=2ee4153a1985
 Hambling, D. Not just pie in the sky. The Guardian (2002) https://www.theguardian.com/science/2002/jan/10/physicalsciences.highereducation
 Force, U. S. A. RQ-11B Raven (2007) https://www.af.mil/About-Us/Fact-Sheets/Display/Article/104533/rq-11b-raven/
 Airobotics. Airobotics (2020) https://www.airoboticsdrones.com/
 Hambling, D. CIA Reveals Details Of Bird-Like 1970s Stealth Drone — With Planned Nuclear Propulsion. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/07/31/cia-reveals-details-of-bird-like-1970s-stealth-drone---with-planned-nuclear-propulsion/?sh=51870f08428e
 Wikipedia. Betavoltaic device. Wikipedia (2020) https://en.wikipedia.org/wiki/Betavoltaic_device
 NTI. Stolen Lighthouse Radiothermal Generator Recovered from Gulf of Finland. NTI (2003) https://www.nti.org/analysis/articles/stolen-lighthouse-radiothermal-generator-recovered-gulf-finland/
 Hambling, D. Israeli Video Gives Glimpse Of ‘Spiderman’ Urban Assault Drone. Frobes (2020) https://www.forbes.com/sites/davidhambling/2020/07/28/israel-gives-glimpse-of-spiderman-urban-assault-drone/?sh=5caf32344948
 Briefs, T. Life-Detecting Radar (2014) https://www.techbriefs.com/component/content/article/tb/pub/techbriefs/physical-sciences/19384
 Hambling, D. AeroVironment Aims To Disrupt Industry With New Loitering Missile. Frobes (2020) https://www.forbes.com/sites/davidhambling/2020/10/01/new-loitering-missile-aims-to-destroy-targets-and-disrupt-industry/?sh=124c3260390f
 Hambling, D. How Islamic State is using consumer drones. BBC Future (2016) https://www.bbc.com/future/article/20161208-how-is-is-using-consumer-drones
 Novak, M. Pentagon Official Warns About Chinese Drones Without Explaining Specific Security Risks (Again). Gizmodo (2020) https://gizmodo.com/pentagon-official-warns-about-chinese-drones-without-ex-1845022250#:~:text=The%20U.S.%20Department%20of%20Interior,susceptible%20to%20cyberattacks%20nor%20that
 Sevastopulo, D. Huawei and ZTE classified as security threat to US. Financial Times (2020) https://www.ft.com/content/943f8c85-58f7-467b-b774-9963dfaa91b5
 Statt, N. French drone maker Parrot selected to develop spy aircraft for the US military. The Verge (2019) https://www.theverge.com/2019/5/28/18642728/parrot-drones-us-military-department-of-defense-contract-china-dji
 Defense, U. D. o. DEFENSE INNOVATION UNIT ANNOUNCES sUAS PRODUCT. US Dept. of Defense (2020) https://www.defense.gov/Newsroom/Releases/Release/Article/2318799/defense-innovation-unit-announces-suas-product-availability-to-provide-secure-c/
 Hambling, D. 7 Reasons DJI Won the Drone Wars. (2015) https://www.popularmechanics.com/flight/drones/a18486/6-reasons-dji-is-winning-the-drone-wars/
 Mac, R. Behind The Crash Of 3D Robotics, North America's Most Promising Drone Company. Forbes (2016) https://www.forbes.com/sites/ryanmac/2016/10/05/3d-robotics-solo-crash-chris-anderson/?sh=6a85ad1f3ff5
 Wod, M. Backpack Turns a Beetle Into a Remote-Controlled Cyborg. Wired (2015) https://www.wired.com/2015/03/watch-flying-remote-controlled-cyborg-bug/
 Hambling, D. In the 1970s, the CIA Created a Robot Dragonfly Spy. Now We Know How It Works. Popular Mechanics (2020) https://www.popularmechanics.com/flight/drones/a30795266/cia-robot-dragonfly/
 Liptak, A. A US ally shot down a $200 drone with a $3 million Patriot missile. The Verge (2017) https://www.theverge.com/2017/3/16/14944256/patriot-missile-shot-down-consumer-drone-us-military
 Hambling, D. China Releases Video Of New Barrage Swarm Drone Launcher. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/10/14/china-releases-video-of-new-barrage-swarm-drone-launcher/?sh=d5fb7322ad7f
 Hambling, D. U.S. Army’s New Drone Swarm May Be A Weapon Of Mass Destruction. Forbes (2020) https://www.forbes.com/sites/davidhambling/2020/06/01/why-new-us-armys-tank-killing-drone-swarm-may-be-a-weapon-of-mass-destruction/?sh=5edf1f32ece8
 Wikipedia. Metalhead (Black Mirror) (2017) https://en.wikipedia.org/wiki/Metalhead_(Black_Mirror)
 YouTube. Slaughterbots (2020) https://www.youtube.com/watch?v=9CO6M2HsoIA
 Hambling, D. AI thrashes human fighter pilot 5-0 in simulated F-16 dogfights. NewScientist (2020) https://www.newscientist.com/article/2252760-ai-thrashes-human-fighter-pilot-5-0-in-simulated-f-16-dogfights/
 Mizokami, K. How the U.S. Air Force Turns an F-16 Fighter Into a Drone. Popular Mechanics (2019) https://www.popularmechanics.com/military/aviation/a29847417/f-16-drone/