Hacker News Reader

Comment by alacombe | Hacker News Reader

On a purely technical level, it's nothing very difficult, a video feed and a bunch of I/O to control the truck (either with linear actuator for steering / braking, throttle control is trivial to do electronically). Same for the gun control, a two axis mount for aiming, fire control can be done with motor and actuating a cam mechanism, and add another video feed for aiming.

I'm only surprised it's not done more...

View on HN

I'm surprised we haven't seen this with $200 drones and a single shot gun attached, either 3D printed or otherwise DIY. Regardless of the level of security surrounding anyone other than maybe the president it seems anyone would be pretty susceptible to this and the odds of it getting traced back to someone who bought the device with cash and was careful to not touch it or leave DNA would be pretty small

Because these things aren’t reliable enough.

Operations like this take years to plan and you often get only one shot.

This also means that this story is quite likely fiction.

Have you seen people doing drone racing? I think this would be pretty easy for them, and even if it's not reliable you're out $200 and some DIY time. Even if it takes 5 attempts, $1000 is cheap as hell to assassinate someone

That’s not the point. The cost of assassinations on this level is in the prep work.

It takes very large amounts of money, skill and time to develop and run an intelligence network capable of pulling something like this off.

As counter intuitive as it seems these organizations are extremely risk averse because to them risks means years of work down the drain, opportunities lost and often for good and quite likely dead assets and or agents.

A $200 or even a $2000 drone doesn’t minimizes these risks it introduces additional ones.

Killing someone in a hail of bullets is easy and reliable, using a toy with a small explosive attached to it isn’t.

Use a remote control drone airship and drop 25-50 guidable mini bombs to hit the target in a dispersion pattern with a concentrated center. Small radar footprint, quiet operation, large destruction area.

Have backup gunman to mop up if needed, likely same pointing illuminator at target.

Not what happened but easy enough.

I don’t think you realize just how ridiculous what you proposing is...

Might as well call a B-52 to carpet bomb the place.

You aren’t getting an airship and a few dozens of drones into a country like Iran.

Nonsense, that plan worked perfectly in Call of Duty 4.

I'm not talking about coordinated organizational assassinations. I am talking about lone wolf extremists that want to murder someone who maybe isn't even that high profile in the national stage. I could rig up something like this in a single day and could probably practice flying casually in the evenings to become proficient enough. It's a really low barrier of entry for anyone technically savvy and violent enough

Disclaimer: all hypotheticals, I don't advocate violence, I would never do this even for hypothetical purposes

A gun has even a lower barrier of entry if you want to kill someone, and kitchen knives even lower.

Small drones can’t cary a lot of payload getting your hands on high explosives that would actually do enough damage isn’t easy, rigging this up in a safe manner isn’t as easy as you think either.

Like I’m not saying you should try to do this to prove a point but you have much higher chances of blowing yourself up or even more likely be arrested prior than you think.

That would only cover the delivery method, no the payload, but this is very much of a covert way to assassinate someone. Killing people is easy, the biggest shortcoming is that terrorist are trying to be as visible as possible. This is the same shortcoming seen is so many Hollywood movie... when you have to shoot, shoot, don't talk.

I suspect the chances of actually hitting a target would be very low, if the weapon was rigidly mounted to the drone?

Although if a grenade was used rather than a firearm, it could be worryingly effective.

Then again, RC aircraft have been around for many decades, and haven't become the terrorist or criminal's tool of choice. Even flying with a first-person view isn't particularly new, here's a video about a hobbyist doing it in the 80s: https://www.youtube.com/watch?v=e5_gPhT61ek

Terrorists want to cause terror and are unlikely to carry "surgical" assassination.

Or a small swarm of like 25 high speed drones that just dive bomb the target and explode on contact. It’s like poor man’s cruise missile.

How are you going to get your video feed out of the country remotely, and get the control signals into the country?

It's not trivial at all. It's trivial-ish if you're operating over a custom data link from a nearby location.

The more remote the operator, the bigger the lag and the bigger the threat of signal loss.

And you can't rely on the Internet because of lag and potential packet loss or congestion.

It's hard enough to get everyone on the same screen in a Zoom meeting. Driving a car remotely and shooting at people over a patchy mobile data link is a whole other level of complexity.

Bounce it off of a satellite, like on-location news broadcasts do? Those all operate out of a civilian style van with an antenna in it.

Use a multi-cellular modem. Starsky Robotics remote-drove a truck (in the US) that way.

The Nissan was not driving remotely it was fixed on the side of the road. So signals needed to remotely fire gun and blow up truck. I wonder how they were able to get the scientists car to stop in the right area.

Zoom meetings carry HD video, generally both upstream and downstream. You hardly need 1080p 30pfs at a high bitrate to aim through crosshairs. You can totally split the feed to a low-resolution wide field of view (FOV) feed, then crop your feed to a low FOV once the target is acquired to enhanced apparent resolution. Common thermal scope are generally 320×240[0], and that's plenty to hunt at night. Reduce the frame rate to 15fps, and that's 54x less raw bandwidth, without accounting to whatever gains you can gain with H.264. I'd probably implement some sort of variable bitrate, aim to the general target direction with the high FOV low-resolution low-bitrate feed, then zoom to the crosshair section (low FOV, high resolution, even though the "effective" resolution is the same) and higher bitrate.

[0: btw, I'm not not implying a thermal scope was used here, but simply drawing a parallel for comparison with off the shelf commonly used hunting equipment. If 320x240 is enough to kill a coyote at night at 100 yards, it's plenty to target a human. Not to mention full-auto fire will cause recoil and whack your aim, so you're not looking to repeat a 5-shot group within dime either, just shoot in the rough vicinity and spray the target with projectiles.]

Also, the US army (not implying US involvement here) is piloting drone from the other side of the planet, so getting a signal out of a country is not very difficult.

Imho this will only become more feasible as the costs/ease for such implementations keeps getting lower with further advances towards autonomous driving.

Very comparable to what happened with drones: Used to be that only very sophisticated state actors could field such weapons.

But with the rise of consumer-grade drones we started to see DIY implementations where consumer-grade drones were repurposed to drop explosive packages [0].

It's even become a thing for boats, back in 2018 three remote-controlled boats attacked a Saudi-flagged oil tanker in the Red Sea [1], the most recent attack like that just happened this year in July [2].

[0] https://www.defenseone.com/technology/2017/01/drones-isis/13...

[1] https://www.marineinsight.com/shipping-news/swarm-three-remo...

[2] https://uk.reuters.com/article/uk-yemen-security-saudi-boats...

Exactly. I guess the only real challenge is in ensuring that the wireless link between the controller and the vehicle is both uninterrupted and low latency.

Reply on news.ycombinator.com