Drone Series: Sensor technology to pinpoint the location of drones

In my first article for International Airport Review, I looked at the probable legislation changes around drone use and whether or not these changes might make any difference. In this second instalment of my contribution to the drone series I will explore in more detail how airports in particular might benefit from the new legislation, considering specifically how they may use some current and emerging technologies to detect and identify drones that are in breach.

But first, back to that legislation…

The Royal Aeronautical Society recently hosted a conference entitled “Drone Regulation and enforcement in Europe: Harmony or Fragmentation?”. The attendees and speakers represented a good cross section of the relevant industries (airports, drone operators, counter drone companies, UAS trade bodies and so on) and the various legislative / advisory bodies (CAA, EASA, DfT, NATS). A full analysis of the content of the day is not appropriate or necessary here; however, what was clear is that the future legislation is far from determined and the implications for law enforcement, drone users, airport operators and others are far reaching. In addition to new EASA regulations, there is also a pending amendment to the Air Navigation Order, the UK Government’s own Drone Bill, a private members bill (sponsored by Peter Bone, MP) and, of course, the issue of what if any existing legislation may be relevant. Even assuming all of these bodies can put legislation that is coherent in place, between them there is then the question of enforcement.

Do we really expect a police officer to be able to judge the height of a drone accurately enough to be able to issue a fixed penalty notice? Will new drone owners/users have to take some kind of simple test and, if so, how will that be administered? Will commercial drone operators (who have built a business around the current rules) still be able to do things they can legitimately do today? Will there be ‘grandfather’ rights for existing operators? (A not uncommon mechanism in aviation licensing). These, and many more questions, are still to be answered (along with how the European angle will play out after Brexit). Current and emerging technologies can certainly help but, ultimately, these laws will be tested by the courts. Such tests, coupled with the pace of technological change in the drone sector, strongly suggests that the new legislation will not be a one-off change but more likely the start of a series of legislative changes over many years. I was going to joke that before long we will see specialist ‘drone lawyers’ emerging – but checking the attendance list for the RAeS I see there were a couple at the conference so they’re here already!

Detect, Track and Identify Technology

We’ll leave mitigation / defeat technology aside for now – partly because that gives us something exciting and punchy to talk about in future instalments but also because they are currently pretty much all illegal in the UK (a broad-brush statement but it’ll do for now). As the Gatwick and Heathrow incidents showed, the ability to know where a drone is (and perhaps an idea of its intent) is fundamental to informing the decisions an airport operator needs to take in response. This can be as simple as when to close a runway but, of course, follow-on decisions, not least when it is safe to re-open, also require real-time and accurate supporting data. Absolutely essential is to drive the false alarm rate as low as possible without going so far as to risk missing a drone that is a genuine threat. One of the first issues airports face when a drone detection system is put in place is that the number of drone incidents will almost certainly go up as they become aware of drones of which they were previously ignorant. Clearly the potential for more runway closures (and therefore passenger disruption) is the opposite of what the airport operator wants to achieve. If some of that disruption stems from false alarms the negative effects will be even worse as the justification of ‘better safe than sorry’ won’t hold water. Therefore, as well as investing in appropriate technology, airports will need to consider their concept of operations in light of the new information flows they are going to be presented with by that technology.

In addition to this need for situational awareness to inform decision making, we also need reliable detect, track and identify capabilities to support law enforcement, evidence gathering and, of course, to maximise the use of our airspace whilst reducing risk to an acceptable level. Remember, this is not about stopping drones from using the airspace, it is about facilitating the best use of the airspace by a multitude of users. It was very refreshing at the RAeS conference that this is the approach all of the vested parties are taking – there are, after all, commercial and societal benefits to be reaped and, if we get it right, the UK is well-placed to be a world leader in unmanned technology.

At the highest level there are two types of detect, track and identify technologies and these will be familiar to anyone from the manned aviation sector – namely cooperative and non-cooperative. Of course, ‘non-cooperative’ doesn’t necessarily imply deliberate rule breaking, in this case it simply means not participating in a cooperative system.

In manned aviation cooperative systems include IFF and ADS-B (and many others) and indeed procedural methods (for example confirming to a known flight plan). There are corresponding systems in the drone world. These both leverage existing technology (mini ADS-B solutions for drones are on the market) but also use new ones (4G based systems). Additionally, many drones report their position to their ground control station – often a hand operated controller linked to a smart phone or tablet. This, in turn, can then interface with other systems to report a drone’s position to other airspace users.

Non-cooperative systems rely on sensors to detect and report the position of a drone that is not known through cooperative means (or, rather more nefariously, where the cooperative position might be considered untrustworthy such as through GPS spoofing). Literally thousands of experts have been wrestling with which sensors to use for many years now and, sooner or later, all come to the same conclusion – namely a layered system of sensors operating in different spectrums is required. Radar, radar frequency interception, electro-optical/infra-red and acoustic sensors have all been used successfully to detect drones. Each has its advantages and disadvantages, in many cases dictated by the characteristics of the target drone. To take an obvious example, a drone that is not emitting any RF will not be detected by an RF sensor! The basic problem here is very easily stated, namely: Reliably detecting, tracking and identifying non-cooperative drones is hard, really hard.

After all, they’re small, often slow moving (or hovering) and have relatively little metal in them – so they are a challenge for radar. Size and even colour make them difficult to see with even high-end cameras. They are relatively quiet, and the world is relatively noisy (especially the bits of the world used by airliners for taking off and landing!), so acoustic sensors are easily swamped. Of course, the manufacturers of these various sensor types will all say they can overcome many of these issues to some degree – and in many cases they are right, but the fact remains there is no single sensor ‘silver bullet’. Therefore, multiple sensors working across multiple spectrums has to be the answer.

This then gives rise to two more problems – technical complexity and information management. Technical complexity in that the sensors needed to detect drones generally need to be optimised for that purpose (a radar that is good at seeing an A380 at 150 miles is pretty hopeless at seeing a DJI Inspire at two miles). The RF spectrum is ever more congested and many (probably the majority) of drones use 2.4GHz as their control and/or link frequency. Picking that signal out against the cacophony of WiFi, Bluetooth and other devices in that band is not easy and requires very sensitive equipment that is very well set up. Cameras capable of seeing drones at useful ranges need exceptional quality optics and, even then, will be operating at high levels of zoom (50 times and more is not uncommon) which in turn need high quality pan, tilt and zoom mechanism to point them the right way. At those sorts of zoom levels, it is rather like looking at the sky through a straw so maintaining tracking once you’ve got the drone in the field of view is also complex.

Information management is a key issue as however complex the outputs from the sensors, there is clearly a requirement to present this to the user in as unambiguous a way as possible. Indeed, in many cases the requirement is for existing security personnel to be able to use the system. Put simply, the issue is to present users who are not aviation experts with probably the most complex air picture we need to build. There are also issues around fusion and correlation between sensors – we need to be sure that what we present to the user is not an artificiality induced by the system. For example, by the very nature of how they work a radar and an RF sensor may ‘see’ the same drone in two slightly different places. This can be due to minute time differences in detection, misalignment of sensors, or even simple environmental factors such as RF sensors reacting to a signal that has reflected off another surface (noting, of course, at least two RF sensors are generally needed to fix position as a single RF sensor usually gives a bearing of the source signal). There is then the challenge of resolving if that really is one drone being detected by two sensors or if it is in fact two drones flying very close together.

A number of approaches are being taken with these issues. There are existing software solutions that can be re-purposed for this role, for example, military grade systems that already take multiple sensor feeds to build up an air picture over a battlefield/contested airspace. However, these are often extremely expensive and come with a degree of user complexity that drives a high training and expertise burden. A number of companies have therefore developed dedicated alternative command and control and sensor integration solutions that are very much optimised for the task and designed from the outset to be extremely easy to use, rapid to assimilate and inherently low false alarm rate. It is worth noting that if potentially lethal effectors are required (for example the automatic triggering of kinetic effectors) there may well be assurance requirements that make the military derived solutions more suitable, albeit at the significantly higher price they command.

Conclusion

This is just a brief insight into the sensors and other technologies we can use to detect, track and identify drones and then present that information to users. At all times the aim is to support effective decision making. In particular, the techniques of identification have only been briefly touched upon, these will be returned to in a subsequent article. It is very tempting to think of the drone problem from a point of view of “what can we do about a bad drone?” but, actually, until we know where it is and that it is actually bad, that is something of a moot question.