Hacker News Reader

Comment by ggreer | Hacker News Reader

Drone technology is advancing at an impressive pace. I think we're going to see some interesting (and terrifying) applications.

For example: Add some face detection software, then change the payload to a laser dazzler and pepper spray cartridge. Now you have a drone that can temporarily incapacitate people. When these get cheaper (say $100 per unit), a police department could buy a swarm of 1,000 for $100k. Whenever a riot breaks out, just unleash the drones (geofenced to the desired area). Instant crowd control.

Like I said, interesting and terrifying.

View on HN

It goes the other way as well: using a drone to commit arson or even an assassination (mounting a weapon is well within the weight profiles of many current devices), is something that is inevitable. Drones have range deniability for things like ransom pickups as well.

As an aside, I expect anti-drone systems to become a significant debate soon enough (e.g. if there's a drone peeping in my windows, what rights do I have to disable it?), and then to become a significant commercial concern.

> what rights do I have to disable it?

Your rights probably are limited to calling the police, like pretty much any other non-immediate danger legal situation that involve applying the use of force.

Interesting point. Quick question(s):

Is it "use of force" when targeting a piece of equipment? If you park your car in my driveway, I can move it.

I may have liability, but I think I can destroy it if it is "invading" my property.

Do I not have some "height rights" at home?

I don't know, but I hope you have enough money to pay a lawyer. Personally, I would just call the police.

Who will do what? Commiserate with you over the phone. I'd get a butterfly net or a paintball gun, take the thing out, and dump it in the trash compactor. Who's going to come demanding the thing? What can they say? "When I was flying my gadget illegally in your airspace, you damaged it!" to which you reply "Lets go down to the station and work this out".

I'm pretty sure the police will come by if you call them, that's their job. There may be the chance that they're all busy dealing with more important things, but they will come, if only to come and take a police report.

If you do destroy their drone while it is on your property what will they do, call the police?

Probably yes.

Ok maybe yes, not probably, but if they do call the police, are you going to admit to destroying their thing? Are you sure you're going to be in the clear there? Why take that risk?

I think there should be a centralized tracking system for drones, and each drone should have a unique cryptographic identifier, so you always know who made the drone and who owned it.

Then, while you still won't be able to stop assassinations and such, just like you can't stop them now anyway, you can at least know who did it, and punish them afterwards.

How would you make sure this cryptographic identifier cannot be tampered with?

Huh? What makes you think criminals will be sure to use a drone with such an ID system? There are no unique parts to a drone. Other than the motor controller and motors, it's mostly cellphone parts. Motors and controller have several other applications.

Drones that fail to ID themselves will be shot down on sight.

You could always spoof the ID so that kinda defeats the purpose.

Maybe have each drone regularly broadcast their GPS's most recent NMEA sentence (latitude, longitude, timestamp, altitude, etc.) signed with their public key?

That's what came up with after thinking about it for a bit. There's probably a better solution out there.

Pretty sure we have solved this problem, since we use HTTPS

Thus guaranteeing that unregistered drones are dangerous.

Who cares? Are you suggesting setting up an agency that monitors 'airspace' (I wouldn't call area 20 meters above ground an airspace)?

I don't want my drone to be seen at the moment of crime anyway.

Don't like drone in front of you? Accidentally throw a piece of string on it.

That's like asking for every gun or knife to have a unique cryptographic identifier.

Guns and knives aren't (usually) electronics.

And guns have serial numbers and records to trace their purchase, right?

> Then, while you still won't be able to stop assassinations and such, just like you can't stop them now anyway, you can at least know who did it

Even if you get criminals to comply as far as only operating drones with proper IDs, so that you can trace the legal purchase history of a drone, that doesn't mean you know who used the drone. Drones and their supporting control equipment can be stolen (and the owners prevented by force from reporting the theft) and people willing to commit assassination by drone certainly wouldn't be unwilling to commit theft, kidnapping, and/or murder of drone owners to acheive that.

The FAA has just this week said that drones flying under a COA will be required to have an N-number: http://www.uasvision.com/2014/09/25/faa-issues-requirement-f...

Do you expect to be able to regulate the electronics industry that tightly? You'd need some universal DRM system implemented in every accelerometer, GPS module, electric motor, and microcontroller.

There's little to prevent construction of unlicensed drones. Like they say, make operating drones illegal, and only criminals will have drones. And they can be made simply with mail-order parts.

Probably we're just going to have to get used to drones being around. Like we got used to cities and crowds on public streets and horseless carriages. This paranoia about drones will be seen by future historians as quaint.

I didn't want to put too much speculation in my comment, but you're right that there are a ton of things that individuals could do with drones.

Here's a movie plot attack: A week/month/etc beforehand, plant 4-8 drones a half-mile downrange of the airport a VIP is scheduled to fly out of. When their plane takes off, trigger the drones to fly into the the plane's turbines. A fully-loaded plane with no working engines would quickly stall and turn into a fireball on the ground. You might not even need explosive payloads for this to work.

Hobby planes have been around for a long time. Wouldn't this have happened by now if it were such a risk?

Good point, but I think this hasn't happened because almost nobody is a psychopath, and most psychopaths are incompetent. If I believed otherwise, I wouldn't speculate about these things publicly.

Also, it would be hard to pilot multiple RC planes or helicopters simultaneously. Any attack that requires people cooperating is much less likely.

"most psychopaths are incompetent"

I wish you were right.

Psychopaths have a wide range of traits, but I believe those who would consider doing something like this, also tend to have poor impulse control that may interfere with their plans ahead. What makes it a terrifying possibility is that terrorists, who I don't think are all psychopathic, may have a better chance of pulling it off.

Not to mention that the vast majority of people who would be willing and capable of pulling off such an attack are probably people in government, so government regulation wouldn't make sense as a recommendation.

The same can be said about almost any terrorist action. They're not generally limited by what is possible to do, but mostly by what people are willing to do. That something like this hasn't happened just shows that there just aren't that many crazy people out there.

Drones are too slow to catch-up with a plane that's taking off. Not to mention the fact that those things are so fragile, the moment they get near a jet engine they will disintegrate.

The plane's future position can be accurately estimated as soon as it spools-up on the runway. So the drones wouldn't need to catch up, just fly up to intercept. To ensure turbine destruction, the drones could have a payload of ceramic ball bearings, explosives, or both.

A passenger plane taking off is going to have a hard time seeing (let alone avoiding) such small, maneuverable craft.

Of course, this is a movie plot threat. Really, nobody is going to spend the time and resources to program drones like that.

The drones you buy at RC shops, yes. Do a YouTube search for jet powered RC planes and you'll find several that exceed the takeoff speed of a commercial airliner.

People associate drones with multicopters. But really they're UAVs.. or.. you know, what you call drones, and can also be planes. small/cheap (800USD) electric powered planes can go about the same speed as an airliner at take off, that is, 400km/h. any faster and it needs to be turbine powered.

Also you'd go the opposite way (toward the plane) so the speed dont matter much...

Now, good luck to aim for the turbine - for all turbines, at once.

Targeting something stationary at a location known well in advance would be even easier; like say, a podium at an outdoor speaking event. I wonder what speed an inverted drone, accelerating toward the ground, can get up to while controlling its impact location to within a foot. Might be that adding an explosive is largely superfluous in that situation too.

I don't think kamikaze drones are even remotely practical as weapons for assassination. You'll have big issues with accuracy, lethality, and the element of surprise. It could work, but I don't see any circumstances where a hunting rifle wouldn't be a far better tool in every aspect.

Nothing you could not do 10 years ago without any kind of remote video or automatic piloting.

Jet engines have to survive the Chicken Test, which is to throw a frozen chicken into the engine running at full speed. A drone would have to be made pretty chewy to stop a jet engine.

And then the plane just lands. They can fly with an engine out.

http://www.straightdope.com/columns/read/1134/is-a-chicken-g...

"The most pyrotechnic test of all requires that dynamite charges be strapped to the compressor blades and detonated while the engine is going full blast."

RC Planes are far cheaper than drones and can fly FPV and have been around much longer than drones. They can also carry very large payloads as well. These whole what-if scenarios are good to hypothesize, but assassinations happen if they want to happen. They are not going to happen more because of drones.

RC planes ARE drones. I think the word you meant to use was 'multirotors'.

Looking forward to clouds of warring aerial nanobots, a la Diamond Age.

It’s only a matter of time before state and municipal governments start moving en masse to regulate civil drones – which is why there probably needs to be an association of some sort (is there one?) to get out in front and start the conversation in the right place.

I could imagine license tags being one outcome of this process: If a drone goes spying outside your window or damages your property, you now have a fairly simple recourse of reporting that tag number. The existence of this recourse would also go a long way toward making civil drone use less scary and more socially acceptable.

There is AUVSI http://www.auvsi.org/auvsi/home/ (Association for Unmanned Vehicle Systems International). I don't exactly know how effective they are what they do or even what they do, for that matter. They do seem to come up a lot in these drone issues, but their Wikipedia page appears to be self-written.

There may need to be something more than that; a "drone lobby" perhaps?

The AUVSI is known for generally representing more of the traditional military-industrial side of the drone business than the smaller, more entrepreneurial side. E.g. see their corporate membership list, which includes Boeing, General Atomics, Lockheed, Northrup Grumman, Rockwell & Saab (along with newer members like Google & Amazon) at the "Diamond Level": http://www.auvsi.org/membershipandchapters/corporatemembers

The Silicon Valley chapter of the AUVSI is more progressive with regard to small business drone advocacy, but it's hard to tell if that is having an impact nationally.

I support Bitcoins and disagree with most of the arguments that they promote criminality, but I have to admit that collecting a ransom would be much easier by accepting it as Bitcoin (or other alternative cryptocurrencies, some of them much more anonymous than Bitcoin).

Lots of "ransomware" malware out there is already using this tactic, and law enforcement have no money to follow.

Drones are less practical than existing tools for most everything you just listed.

For peeping in windows, a drone is extremely impractical. A telescope will yield much better results. The only thing a drone can better do is look down at your roof, or over your fence if your fence is higher than nearby buildings.

For assassination, traditional tools are clearly more practical in nearly every case. A hunting rifle at a few hundred meters is going to be a heck of a lot cheaper, more reliable, easier to use, and more difficult to trace than some drone with a mounted weapon.

I don't really know how arsonists generally do their thing, but I can't imagine that dropping something flammable is more practical than just driving by and throwing it.

I'm sure there are very specific cases where sophisticated drone usage could be useful for committing these crimes, but I don't think it's a major concern, and not even one of the biggest concerns of drones.

I think the biggest 'benefit' to having a drone perform these tasks is the fact a robot is carrying out the task.

For example, instructing a drone to assassinate is potentially closer to hiring a hitman, than murdering someone firsthand with a shotgun. Closer in the sense that the puportrator feels a sense of detachment.

You seem to be only considering the moment of crime and disregarding what happens before and after it. A tool that almost completely removes risk of getting caught and need for alibi will surely be seen as more practical by all except mad and ridiculously dumb people, regardless of how less effective it is.

I considered the likelihood of getting caught as well. I mentioned that explicitly in the rifle example. It's fairly obvious for the telescope example. For arson, again, I don't know much about how arson tends to be committed, but it seems that a drone is not inherently harder to track than a car, bicycle, or person on foot, given the range, speed, reliability, and conspicuousness of those modes of transportation.

How would you trace back a drone that shoots a bullet then wipes its data and dies?

The MAC addresses for the wireless interfaces are likely trackable to an individual purchase.

Often they can be wiped as well.

What? ifconfig or My Network Places in Windows. Adapter cards had to have their mac address programmed once; it can be done again. Its firmware.

In some cases yes, but that requires additional expertise.

Not much. The OS has a utility to do it.

Perhaps. But a wlan card taken from a discarded/stolen/recyled/used corporate machine, then transported half-around the globe and sold for cash, isn't necessarily easy to trace back to a perpetrator.

A drone with a gun is unrealistic, I think. The stability of a light, air-borne vehicle is just not good enough. A drone with fragmentation grenades, on the other hand, could work well and pretty much erase all evidence.

Do you watch the news? Flying machines with guns are decades old tech. Drone versions are actively used. Welcome to the present.

I'm mostly commenting because I think the "small drone with a gun" idea is interesting; no idea how it would play out in practice.

If you had a drone with a weapon, I would think you'd want to fire on someone using some variation of Tracking Point style hardware [1], except partially self aimed. Basically see the target, paint the target, drone uses internal sensors to detect a brief instant of random stability and fires the weapon

http://arstechnica.com/gadgets/2014/08/gun-linux-on-the-rang...

On the show Sons of guns, they already made a drone that can shoot a pistol. Not very hard to do. Drones can withstand very high winds and hold position very easily, lots of hand guns have very little recoil and can easily be shot from a drone with very good accuracy.

Fragmentation grenades are presumably fairly difficult to come by. I think it's laughable to worry about an autonomous or remotely piloted drone assassinating people with fragmentation grenades when a normal firearm operated by a human is extremely practical and seems to work well enough. The latter is designed for nearly that exact purpose, and refined over centuries. They're cheap, easy to use, easy to get (in some places), and reliable.

when a normal firearm operated by a human is extremely practical

Again, doing a crime and getting away with a crime are two enormously different things. The truth is that the majority of humanity can commit terrible crimes with utter ease. There is absolutely no question about that.

Most of us don't, however, because of morality and our nature. The rest don't because they don't want to spend the rest of their life in jail, and for major crimes the probability of getting caught is enormously high. Things that reduce that probability (and a device that means you don't have to be on or even near the site, and can extract 100% of evidence with a very high degree of success, does that) open up opportunities for people to do such crimes.

As a parallel, someone mentioned bitcoins, and bitcoins absolutely opened up the world of casual extortion, because suddenly you don't have to go to a mailbox or meet in a park.

Fingerprints, hair or clothing samples, asking nearby drone hobbyists if they might know anyone into drones who might want to murder the victim, sales records of retailers of drones or drone supplies, and of course the normal stuff, like who would want the victim killed in the first place.

I'm not a detective, but I suspect having the large, complex piece of electronic equipment that was used as the weapon would be a godsend compared to the evidence available in a normal murder or assassination investigation.

Obviously the plan wouldn't be "drop the drone at the scene of the crime". Why would anyone ever do that? I mean, of course they would minimize the reveal were that to happen (though note that drones are becoming absolutely pedestrian, so this notion that you talk to the niche club and find the drone enthusiast is absurd), but the rational plan would be that the drone would escape as well. Because its odds of escape, currently, barring mechanical defect, would be close to 100%. No one is going to catch a drone.

A drone leaves all that bomb trace leaves and more - its construction, hair skin and dna traces of the builder, even its radio control details. So its not perfect; its more like leaving the rifle at the scene.

And the idea that shooting a gun brings down some hailstorm of official response is myth. Fire a gun in most places and ... nothing will happen. Folks will rush to the scene of any murder, but will only find the sniper's location after hours or days of investigation. It won't be like the NCIS show where the investigator glances around and says 'must have come from that rooftop; its the only secure sniper post in a half-mile!' You can shoot from nearly anywhere with line-of-sight, from the partly-open door of a parked delivery van, without much risk of discovery.

"Shotgun-Toting New Jersey Man Shoots Down Man's Drone", charged with criminal mischief and weapons charges: http://www.nbcphiladelphia.com/news/weird/Drone-Shot-Down-Lo...

As hamstergene mentioned, drones are the before and after risk abatement.

A hunting rifle at a few hundred meters is going to be a heck of a lot cheaper, more reliable, easier to use, and more difficult to trace than some drone with a mounted weapon.

Aside from the difficulty of a few hundred meter shot in urban settings, shooting a weapon from a location announces where you are, exactly when you are there. Aside from the immediate and overwhelming response likely, the human traces would be enormous and unavailable. A drone, in contrast, comes from some unknown place off in some unknown direction, can with ease get anywhere near the target (on a nearby roof for instance), and then when the task is complete it can go low and fast and get back to a drop-off. For arson and peeping, the same premise holds -- a drone can go to exactly where is needed, do what is necessary, and get away with, currently, essentially zero risk of capture.

Will we eventually have drone registrations and tracking systems, comprehensive urban radio tracking and announcement, police overrides, etc? Yeah, probably, which was what I was getting at -- events will inevitably happen, and regulations will adjust.

These types of assassinations don't seem to happen very often, so there probably isn't much actual data. But I would expect that a single shot of a rifle is quite difficult to pinpoint, especially in an urban setting. There are audio gunfire locators that some cities have, but even then, the assassin can be long gone by the time anyone could respond. Just drive a car into a nearby parking garage, do the thing, and leave. For a drone operation, the setup and "mission" time is almost certainly much longer. I don't think we're anywhere near fully autonomous missions that could assassinate with any appreciable reliability, so you're going to need to fly line of sight (very impractical because of the range) or FPV (still impractical because of the required skill, equipment and mission time).

Could a professional band of assassins pull off an assassination with a drone? Certainly. Would it be far easier to just use a rifle? I'd say absolutely.

The us military finds drones more practical for bombing.

> A hunting rifle at a few hundred meters is going to be a heck of a lot cheaper, more reliable, easier to use, and more difficult to trace than some drone with a mounted weapon.

Perhaps, but you've still got to pay the guy holding the rifle. A drone could self-terminate. Most parts can be 3D-printed & assembled in a clean tent, untraceable. The most expensive part is possibly the infrared camera ($1~2k).

There's Seek Thermal Camera that has a decent resolution (200x156) and a low price ($200): http://obtain.thermal.com/category-s/1818.htm

"For peeping in windows, a drone is extremely impractical. A telescope will yield much better results."

I'm not sure I agree with this at all. Drones allow you to access viewpoints that are otherwise impossible using telescopes at a fraction of the overall cost.

Example: https://www.youtube.com/watch?v=kLB66lge6oA&list=UUZfBMqj1pN...

https://www.youtube.com/watch?v=2PKG8q94D98

> For peeping in windows, a drone is extremely impractical.

Well, say I set up a drone for remote control over the internet, then rent it out to people that are located half around the globe. I act as a cut out, and whomever is spying will (may[1]) be hard to track.

[1] If I'm getting paid, there's the issue of anonymous transactions, and as for control, latency might be too high to use Tor -- but the main point is that drones already facilitate something new, if not revolutionary new.

Agreed. Murders/assassinations via drone, with no hope of determining who did it, are coming sooner than later. Drug cartels, street gangs, etc. can already easily afford a weaponized drone - they just may not (yet) know how to obtain/build one.

The ability to anonymously open fire on virtually any target one chooses is a scary reality, but it's almost here. With no risk of prosecution, murders will happen that never would have happened before. Be nice to your neighbors :) .

to be honest all you see there are features of the APM, which is also sold by 3DR and developed by open source enthusiasts + 3DR.

The features have been there for a couple of years now albeit they've been improved and made more and more reliable over time.

That includes stuff like follow me (follows phone gps coordinates) auto land, auto return to home point, etc.

I've had an APM board (well, many APM boards really) along other boards for flight control since many years.

Now they sell is as a board called PX4 which is more powerful but uses the same software (but should evolve into something better than the APM).

In reality tho stuff like follow me doesnt work as well as on the video. The short 3s extracts are best case scenarios.

The main issue is that it cannot track the subject (no image recognition) so it really just follow some gps location and hopefully the camera is aligned (it auto aligns it depending on the trajectory, else you need a copilot to care for the camera).

Funny trivia, a large amount of kick starter projects with their "awesome" return to home, follow me, etc are using the APM board and the APM is the only reason they even get support... except the APM is free/opensource hard and software.

DJI being an exception, but their board isnt as full featured as the APM/PX4...

Finally, remember that these helicopters fly for about 15min and are still relatively prone to failure and interference. we're far from reliable large scale reliable delivery, mass governmental use, etc.

That follow me implementation is an OpenCV computer vision application, not GPS. https://www.youtube.com/watch?v=gr0jIAT68eI&list=UUcp-_sF8s4...

Yeah its coming eventually!

It’s very easy to take down a drone. And I don’t mean shoot it down with an assault rifle. You could even take it down with a sophisticated slingshot once it’s in a low altitude. And in order to tase or spray someone it has to get really low. These things are very slow and not that agile.

  >  in order to tase or spray someone it has to get really low. 
  > These things are very slow and not that agile.

For now.

And it will become hard to even see/notice them as they become smaller and quieter. What then? Shoot down a fly?

You’re talking science fiction now. If it becomes so small as a fly how on earth would it be energy sufficient? With a battery? And that battery will hold enough energy not only to keep it flying but also to provide for a taser that could incapacitate an individual? And at that size it should also carry some communication infrastructure to receive instructions from a command center, a gps, and a cpu capable of running face recognition or even detect movement patterns because otherwise you can’t have efficient crowd control. Sure, in a couple of generations we’ll get there, but not in the immediate future.

And 640K should be enough RAM for anybody. Betting against technology is a losing game.

My nephew has a palm-sized copter with remote that flies for 10 minutes, right now, got it for XMas.

There have already been dragonfly-sized gadgets made by graduate students. Check out youtube.

RFID tags are all the computing power you mention, already on your high-end product you bought from WalMart.

The future is clear. Maybe a one-use metal-air battery on a chip-bot? If I can say it, its conceivable.

> It’s very easy to take down a drone.

For sure. Laserpointer to blind it, jam gps to confuse it...

> And in order to tase or spray someone it has to get really low.

For tasing, sure. And for spraying an individual in the face. But if we're talking swarm-based crowd control, there might be other options. Anything bad that drops down (say 1000 minature cs-cannisters -- effective even if a few are "thrown back", other gas that is heavier than air, perhaps created by blending two components "on board" the drone, perhaps forms of acid or irritants liquids that are arezol'ed with pressure... audio-attacks...

In fact, swarm based drones might even be hard to take out, if they could rely on low-bandwidth, near-field comms, such as IR, lasers, (directional) ultrasound or whatnot. Not even sure if there's much current research in combining directional ultrasound with the type of advanced coding that is used for high-bandwidth radio short-wave/microwave comms.

At that point people will pull out their guns and play a real-world version of Half Life 2.

Provided rioter don't have their own drones for a drone war.

Reply on news.ycombinator.com