Hacker News Reader

Comment by geoah | Hacker News Reader

I've seen a lot of suggestions floating around on jamming the signal. But that assumes that there is an active operator and not a drone that has been told to go between a pre set list of coordinates. Most high end should support gps navigation so I assuming that you have enough money and don't need to ever recover a drone after it's out of battery you can probably increase your chances of not getting caught.

View on HN

There is a lot of discussion about how a drone might avoid jamming, which is worth considering with regard to going forward, but the fact is that jamming apparently did bring this incident to an end, so the delay suggests that the relevant authorities were caught woefully unprepared.

From media - police started with a commercial DJI anti-drone system. That did not work. Then "drone dome" was brought by the UK army. That apparently did work and the drone was grounded. This is in all likelihood the first time ever that system was used in a non-military context, so it's reasonable for the operation to take some time.

In the Gatwick case it appears the drone was remotely operated rather than autonomously visiting a series of GPS way points.

I figured the authorities could pretty quickly triangulate the signal to locate the operator. But Gatwick is a bucket of suck so maybe they're just unprepared?

Also, I understood that drones are geofenced out of restricted airspace such that they disobey commands to fly where they're not allowed.

Geofencing drones: https://www.heliguy.com/blog/2017/02/16/heliguys-guide-to-ge...

DJI drones are very geofenced. To the point where if you try to fly them inside an enclosure near an airport at an officially-sanctioned event like EAA Airventure, you won't be able to.

But many others are not. Drones without GPS for example cannot be geofenced.

It’s not that hard to build a drone as a hobbyist. Controls on commercially available drones are only going to deter the most casual operators.

That assumes it's running factory standard software.

If my memory of this talk is correct, he demonstrated how easy it was to un-geofence a drone.

https://www.youtube.com/watch?v=5CzURm7OpAA

Radio direction finding is reasonably straight-forward in theory, but practice is a challenging art.

First, what frequency is the control transmitter on? (or frequency-band/freq-hopping in). If it is one of the usual R/C bands, then that is a limited search space. Or if the perp is not so concerned with observing frequency allocation treaties and regulations, well then now you have a big hunt.

Second, what other signals are in that band? Suppose it is in the 2.4GHz ISM band. Given a spectrum analyzer plot of 300 spread-spectrum signals hopping in the ISM band, which 299 are WiFi and Bluetooth and other, and which is the perp? The microwave ovens are easy to distinguish, but still add clutter.

So if you can pick on a single signal, then you can use Doppler techniques to get a rough bearing, and directional antennas to zero in.

I suspect they had some crews working the problem, but it is a challenge.

Anyone interested can look up Amateur Radio “fox hunting” in the US, and in Europe “radio orienteering” is popular. Since signal hunters are not transmitting, no license is required to participate in a hunt.

I know public reporting says US is behind Russia in EW. I don't know how good UK EW is, but if the UK can't locate commercial drones signals, than we are in trouble.

(Fun fact, UK was very good in EW during WWII

https://en.wikipedia.org/wiki/Battle_of_the_Beams )

Military perspective on this:

http://www.thedrive.com/the-war-zone/25608/rogue-drones-have...

>pick on a single signal

You dont have to do it manually. Nowadays you can grab whole 2.4-2.5 range(most common drone remote frequency), process it in real time, at 3 locations simultaneously, and produce signal position map with every transmitter inside the triangle or interest. Even amateur fox hunts can be automated to the point of trivializing whole thing with couple of those https://www.indiegogo.com/projects/kerberossdr-4x-coherent-r... . More professional hardware does exist, quick search gave me http://spectrafold.com/quadrus/product/dru-244a-1-4-pci/ so Im sure MIL already uses this instead of primitive manual techniques.

Then you jam GPS

At an airport.

Which is fine because it’s not necessary for aviation.

GPS disruption is NOTAM'd or scheduled with Ofcom[1] from time-to-time, although not usually at the airport itself. Given that Gatwick was shut at the time, it wouldn't have affected the aircraft since they weren't able to go anywhere.

[1] https://www.ofcom.org.uk/spectrum/information/gps-jamming-ex...

I wonder how much of a problem that would actually cause? Surely they could announce the lack of GPS on ATC and the pilots could revert to ground based navaids. Of course it might be difficult to keep the effect localised considering how weak the competing signal from space is.

What about all of the other legitimate users of GPS?

They suffer a minor inconvenience as precision is lost. The US military frequently does selective jamming near bases and other ‘special’ areas.

The tricky bit here is that this is a civilian airport. Jamming GPS could reasonably be expected to cause significant difficulties for civilian planes and pilots. And thus passengers.

The airport was already disrupted because of the drones, disrupted to the point of shut down. thus the article.

As has been mentioned earlier, there are 5 airports in London.

There aren't five airports in London. There are quite a few airports with "London" in the name, but they aren't particularly close to London. London Luton Airport, for instance, is ~45km north of London, and London Gatwick is ~50km south of London. There's ~100km distance between Luton and the airport that was affected by the drones.

How do you think those planes get to those airports? There are planes flying all over the south east going to if from numerous airports. Then there’s the impact on GPS in cars, trucks, phones, police vehicles, ambulances, all sorts of emergency services. We’d have to be extremely careful about potential impact before considering such a drastic step that has a very small chance of actually being effective.

I think you might have misread my comment. What I said was that there aren't five airports in London. You seem to have understood that to mean something very different.

All flights were grounded anyway

Not really, whatever you are jamming you can just make sure it doesn't make it 10 km up in the air where most of the planes are (especially given the GPS patch antenna pointing up from the plane), and I don't believe they need GPS signal on takeoff/landing, they use older tech for that.

10km up? What about all the planes landing at the other airports of London?

Most of which are geographically distant enough to not be affected by localised jamming.

Never mind that Aircraft will be under ATC in UK airspace, and ATC uses Radar + transponder to localise aircraft.

GPS is a navigational aid, not the primary navigation tool.

> (...) I don't believe they need GPS signal on takeoff/landing, they use older tech for that.

Much older. The basic "stack" uses radio beacons, not greatly different from early XX century tech. The more advanced Instrument Landing System is quite a bit more complex, but in very basic terms is a set of radio beacons that define a glide path. It is local, and requires no GPS.

I wonder, does that mean loss of precision or complete failure?

As always it depends. There was defcon talk[1] about messing with drones and one part of speekers research involved jamming gps signal.

[1] https://youtu.be/5CzURm7OpAA

It can still do dead reckoning using optical flow and the barometer.

That is not what dead reckoning is.

You’re idea is simply flying by other instrumentation.

Dead reckoning would be the accelerometer only, and estimating an assumed possible position based on known facts about the vehicle’s inertia and direction over time.

Using an accelerometer wouldn’t be dead reckoning either. Dead reckoning in an aircraft is measuring airspeed and heading and correcting for expected wind. Using an accelerometer is intertidal navigation.

> measuring airspeed

> accelerometer

What exactly would imagine an accelerometer meters? Something other than speed?

Accelerometers measure acceleration. This can be used to calculate the groundspeed of the aircraft. Groundspeed is different than airspeed. For example, if a drone was hovering in a 5mph wind, its groundspeed is 0 and its airspeed is 5mph.

I was a flight instructor for a few years, and have taught many students how to do dead reckoning. That doesn’t mean I’m right, but hopefully gives me enough credibility that you’ll consider my argument. Each of the Wikipedia articles for these topics is robust if you’d like more detail.

Anything without a global reference is dead reckoning. Optical flow sensors just give you a rotational velocity relative to ground. They don't globally locate you. If you combine it with an altitude sensor it gives the same as an odometer would in a car, or the number of steps you take while hiking.

Then the drones fly visually. ANY drone activity will shut down the airport. I think the crucial point to make here is the attacker is inherently in a better position then the defender.

There is a buffer zone around the airport anyway. I would aim to mercilessly shoot down detected drones in that buffer zone.

Visually? I have yet to hear of a drone that can fly by camera instead of GPS.

I assume most can fly without GPS. I can launch either of my Mavics without GPS and fly free, or it can try to use visuals to maintain location.

I've flown within a cave without GPS. The drone drifts and can't really hold position reliably, but it works:

https://www.instagram.com/p/BoayncJgAOO/

Sure, I’ve flown many times without GPS too, but if one is jamming 1575.42 MHz, it’s a simple matter to do the same for 2.4ghz and 5ghz that’s used for the drone to remote link.

Which is great, but the parent post was describing autonomous navigation and disposable drones as a way for a drone operator to evade capture.

But we were talking about not controlling it so they can't jam your controller signal. Sure you can fly manually...

I saw MIT opensourcing it 5 years ago. Flying through dense forests too.

https://github.com/andybarry/flight

It should totally be possible, I trust without clicking that they have an impressive proof of concept. But I've never seen it implemented in a commercial product. The only purpose would be when GPS is not available (indoors? Perhaps in a cave or on a space station? Not very common scenarios to not fly on auto pilot), so it makes sense that nobody bothered.

Yeah I'd be wary releasing that commercially, drones are easy enough already to crash, giving users an automated feature with the potential for monetary loss is asking for drama and bad press.

Military hardware (the Tomohawk cruise missile, in particular) has had optical-only navigation since the early 1980's.

https://en.wikipedia.org/wiki/TERCOM#DSMAC

Huh? This is very common. For example, look at these capabilities from DJI, one of the most important drone manufacturers: https://www.dji.com/products/mavic?site=brandsite&from=nav

What knowledge/experience is your speculation that flying by camera doesn't exist based on?

There are two classes of drones. The ones which need C&C link from their pilot and the ones which fly autonomously. With the ones requiring C&C link the best bet of the authorities is to triangulate the transmitter and/or disrupt the link.

The ones which fly autonomously are more complicated. And the thread you were responding to was talking about this case. There are no commercially available drones which can fly autonomously by on-board visual navigation. The ones you linked can send a camera image back to the operator, but can't fly themselves based on that image.

what if they C&C via internet? (4g). Lantencies are low enough nowadays. Controller could be anywhere in the world.

Same issue as with a controller: jamming. Heck, they could detect the signal and just tell the carrier to shut the customer off. Would probably be unprecedented and therefore take some hours, but if the person persists, surely an option.

Oh right, the following mode! Hadn't thought of that, that surely counts as flying by camera. But it doesn't apply to the situation where you give the drone a preconfigured, uh, visual coordinates (a path, I guess that must be?) and it will try to follow it. So it still cannot wreak havoc at an airport without either you being there (follow mode), GPS being available (autopilot to coordinates), or flying manually using a remote control.

Ever watch drone racing? They're not using GPS. Here's an MIT video of a drone flying using computer vision through trees at 30mph. They've improved this quite a bit since the video was made. This stuff is getting pretty easy to do. https://www.csail.mit.edu/news/self-flying-drone-dips-darts-...

I have about a 6 mile range for video/control with my drone.

What sort of drone to you have?

It's a heavily modified 3D Robotics Solo that I picked up for $300.

The Register has an article discussing why shooting down wasn't a great idea either:

https://www.theregister.co.uk/2018/12/20/gatwick_drone_non_s...

This is pure bs.

In the UK, they ram people of scooters. The biggest danger here is 'a drone may fall on someones home'. Clearly there is either a HUGE force imbalance, or a hidden situation. Shotguns, as posted ad nasuem here, are perfectly safe.

Not sure I agree entirely. A guy with a shotgun in a helicopter could of done it, both of which were available at Gatwick. The shot looses velocity quite rapidly so wouldn't do much damage on the ground. I guess you'd have to hope the drone didn't land on anything much.

Londoners mostly got though the Blitz ok with a million houses destroyed or damaged by bombs. We can handle the risk of a crappy drone crashing in someones garden. https://en.wikipedia.org/wiki/The_Blitz#No_collapse_of_moral...

I've been duck shooting and felt shot from shooters on the opposite side of the lake pattering down onto me and around me. I can't imagine a shotgun itself would present any real risk to those below, compared to the enormous distortion these drones caused.

Jamming GPS at an around an airport is not that great idea given that the airliners rely on it for navigation.

Granted, they have other means of navigating when the GPS isn't available but in the extremely busy airspace around London this would be just asking for a big problem.

How big is the area of a GPS jammer? Gatwick is 10km further out than London's biggest ring road and 45km from the centre.

> How big is the area of a GPS jammer?

However big you want to make it. All a “GPS jammer” is is a radio transmitting on 1575.42 MHz. Modeling the signal propagation of a transmitting device (Aka radio / antenna combo in this case) is a very very well known field.

But the airport was closed anyway. And there's some work by yours truly into targeted GPS spoofing, allowing not only to jam but also to take control of the device using GPS.

There are like 5 airports in London. You want to be pretty confident your jamming will be very local.

There's actually only two "international" civilian airports in London.

London City, and London Heathrow.

Luton, Stansted, and Gatwick are "London" in name only. All about 30mins drive from the M25.

I can reach Gatwick in 30 mins from West London using an Oyster. That's faster than my commute into the City. So it's London, for all intents and purposes.

Isn't London City the only airport actually in London? See e.g. https://www.telegraph.co.uk/travel/news/heathrow-and-gatwick...

Heathrow and Gatwick are about the same distance from London City Airport.

Yes, that sort of local. I'm not talking about kilometers, or even a whole airport. Something like 100m³ should be achievable without too much trouble.

I don't really want to link it to avoid linking this username to things, but I guess this should be okay: http:// rp dot delaat dot net/2017-2018/p95/report.pdf

I don’t know much about GPS jamming, but I know a fair bit about how airliners navigate. GPS is usually not used for final approach, which is typically either visual or ILS (a radio signal transmitted from the ground). This in contrast to the earlier stages of approach which may use GPS depending on the airport and the specific approach in use. If you can keep the jamming signal contained to the immediate airport vicinity and a few thousand feet above, there should be little or no impact on (non-drone) aircraft navigation.

Truck driver uses a GPS jammer so his boss can't track him, then drives by Newark airport, disrupting their "Smartpath" precision navigation system.

Source: https://www.cnet.com/news/truck-driver-has-gps-jammer-accide...

Smartpath: https://aerospace.honeywell.com/en/products/navigation-and-s...

There are many more airports that have an RNAV GPS/GNSS approach procedure, than ILS. ILS requires expensive ground equipment, and fairly low obstruction clearance for miles surrounding the airport. I expect RNAV GNSS is used with some frequency at even big airports like Gatwick, on days when conditions don't require ILS in particular when the ILS is out of service for planned maintenance.

Example RNAV GNSS for Gatwick: https://opennav.com/pdf/EGKK/EG_AD_2_EGKK_8-4_en_2010-07-29....

A drone capable of jamming GPS in any useful area should carry a huge load of batteries to develop enough jamming power, not the best requirement for something strapped to a light flying machine where every gram counts. Anyone can buy a GPS jammer at the usual online stores, but they draw watts to cover an area of tens of meters, that is, to jam a single airplane you must fly the drone very close to it, and lose all other planes in the process. A flock of pigeons is many times more dangerous than a drone, even if carries a functioning GPS jammer.

To me any concern about GPS jamming is plain bullshit added to the soup to justify further restrictions on drone flying. Drones with cameras are powerful weapons these days.

>A drone capable of jamming GPS in any useful area should carry a huge load of batteries to develop enough jamming power

GPS is extremely low power and relatively speaking easy to jam. The received signal strength is far below the noise floor to begin with and it's basically RF black magic that it works to begin with. Those cheap GPS jammers that only radiate white noise at like a watt cause huge problems in the surrounding area.

There was even one case that I know of that would pertain to this subject in particular. A truck driver wanted to prevent his company from snooping on his whereabouts while driving a truck with a tracking system on it so he bought and used one of those cheap chinese jammers. He delivered stuff to an airport and just sitting in the parking lot was enough to keep aircraft from maintaining a GPS fix. There was also the case at Newark airport where just a truck driving by regularly was enough to cause problems.

And really those GPS jammers are the bruteforce and pretty ineffective approach. White noise is essentially what GPS is designed to work in, it just raises the noise floor 20 - 30 dB. GPS uses a gold code to transfer information and it's essentially looking for statistical deviations to be able to get a lock on such a weak signal. If a jammer was designed to be biased and transmit with the intention of preventing a receiver from being able to lock onto a satellite instead of just overpowering the signal entirely it could jam more with less power.

https://en.wikipedia.org/wiki/GPS_signals#Coarse/acquisition...

https://gizmodo.com/jamming-gps-signals-is-illegal-dangerous...

https://www.cnet.com/news/truck-driver-has-gps-jammer-accide...

I've never seen noise generators used as GPS jammers, that would be hugely inefficient. All boards I've seen so far, although not emitting a single frequency, jammed the target using sweep generators operating on the desired GPS frequencies. Some even do that simultaneously, which is why they use multiple antennas; if they transmitted noise they could do that using a single noise generator and a wide band antenna.

Example: http://jammergroup.quality.chinacsw.com/pz66df87e-handheld-6...

And by the way, Boeing had jamming resistant GPS back in 1998.

https://boeing.mediaroom.com/1998-04-23-Boeing-Anti-Jam-GPS-...

n00b question: are the GPS signals used by drones and commercial airliners the same? I know the military have their own special class of GPS.

Yes, same signal. The military signal you mention is military-only.

It is very possible for a drone to fly in a GPS denied scenario.

All of them? My drone uses a uBlox M8N which uses 4 different GPS systems (US GPS/Galileo/Beidou/Glonass). It's nice because you can get a GPS lock in just a few seconds, and it's a lot more accurate if you're connected to 20 sats at the same time than just 5 or 6. You can pick them up for about $20 on Amazon. You can also use satellite imagery and machine learning/computer vision with no GPS at all. You just draw a boundary on the sat map that you want it to stay within and it uses its cameras to match. MIT developed some cool drones that can actually fly through a forest at 30MPH without hitting any trees. This stuff is becoming very easy to do. https://www.csail.mit.edu/news/self-flying-drone-dips-darts-...

If you jam GPS these days, you also get confused ubers, delivery riders and tourists milling around.

First priority in not getting caught is recovering the drone. Physical evidence is the bane of all crime. At least have it crash into a deep or fast-flowing river.

Reply on news.ycombinator.com