Not going to argue the finer points. Don't know nearly enough to sound even vaguely knowledgeable. Hacking a Parrot is a fine proof of concept, but it certainly doesn't link 1:1 with hacking a GPS-powered delivery drone. However, it does illustrate a problem I hadn't though about before, namely that novel delivery methods must necessarily engender novel methods of hijacking and theft. Amazon is tied to a warehouse, no matter what. Their launch spot is gonna be fixed. Their drones can either be stone cold locked out of the airwaves - an unlikely proposition - or they're going to have to transmit and receive. I can sniff their traffic pretty easily if they're broadcasting. If they make it tricky, I can fly a sniffer drone to follow them around logging commands - there's nothing Amazon can do about it. Once I have their traffic patterns, it's gonna be a lot easier for me to interfere with a drone shipment than a UPS truck for practical, legal and philosophical reasons. I don't even need to steal Amazon's shit - maybe I'm a college nerd who enjoys harrying the traffic to another dorm. Maybe I'm a disgruntled UPS worker. Maybe I intercept Amazon's shipments and force them to land near my van. Frankly, at the build cost an autonomous hexacopter capable of cargo, it makes more sense for me to steal the drones than their cargo… and if I can sidle up next to one, blast it with RF to make it forget its life and then throw it in an RFI box, I own it. I can chop it up and sell it off. Educated guess, Amazon's drones are going to come in between $5k and $10k. That's "motorcycle" prices, and bikes get stolen all the time. Would it work exactly like this? Certainly not. But if I can program something to fly next to an Amazon drone and spoof its comms, a whole new world of mischief opens up.
My first thought when I saw Amazon's video was a teenager shooting down the drones with an airsoft rifle and pawning off them off. I might be undervaluing how tough they are, but even if it takes real gun, for <$10 of ammo you could net yourself a $5k+ robot after a bit of repairs and robo-brain surgery. A heavy duty net could definitely hold one of these things down to the ground if you're more trap-minded. They're totally ripe for the picking. With regards to a few other miscellaneous things, I doubt they'd use the same frequency that iPhones, wunderplastic consumer routers, and every damn internet-thing-under-the-sun uses. 2.4 GHz is crowded enough already and switch to 5 GHz isn't going to to help for long in crowded cities as people switch to 802.11n / 802.11ac. So my lappy might not be able to sniff their traffic any more, but that's not going to stop a determined adversary from buying / building a transmitter on whatever new frequency they pick. It's not legal to use an unlicensed device on alternate spectrums, but neither is hacking devices you don't own. But while switching wavelengths may change the transmission protocol, it's not going to significantly affect the encryption technology being used. WEP and WPA were doomed by RC4#Security), but WPA2 uses AES, which is pretty rock-solid, and any newer protocols aren't going to use ciphers that have been broken. If you break the encryption being used on the drones, you've got tech that's worth a lot more than amazon packages, or even a fleet of drones. You're better off selling that tech to the highest bidding government agency or sitting a receiver outside your nearest megabank. It's definitely easier to break the physical security of these things than taking on the behemoth that is cryptography. But hey, who knows how they'll respond if you gag them with frequency jammer!I hadn't though about before, namely that novel delivery methods must necessarily engender novel methods of hijacking and theft.
Yeah, there isn't a lot of spectrum to go around. Probably safe to presume they'll be somewhere easy like 3G. I'd run them from a cell tower; doing anything else would require more infrastructure. There's also the fact that once you have one, you have the methodology for the rest of them.