It’s a chilly morning in late May on the coast of Belgium and I’m standing on a beach beneath a brilliant blue strip of restricted airspace courtesy of the Lombardsijde military base. I’m here at the first ever training course in humanitarian unmanned aviation, along with a motley collection of representatives from various UN agencies, NGOs and EU institutions. We’ve been gathered together by my friend and colleague Patrick Meier, the founder and director of the UAViators network, as well as by faculty members like Rob de Roo from the nearby VIVES University in Ostend, which maintains an excellent program in civil aviation and unmanned aerial vehicles.
Laurent, our jaunty flight instructor, is cranking hard on the bright red handle of a long metal catapult, at the base of which sits a sleek piece of black foam and carbon fiber packed with electronics and optics, called the Trimble UX5. As he does this a thick bungee cable inside the catapult grows alarmingly taut.
“Ok we’re almost ready, stand back everyone!” Laurent shouts to us over the westerly wind.
We snap to attention and scurry to a semicircular perimeter around the launch area. Safety is at a premium out here at Lombardsijde. No one wants to take a speeding UX5 to the face, that’s for sure. Flight safety is a theme that will be drilled into us again and again over the next three days.
The UX5 is a fixed-wing drone designed to capture high-resolution aerial imagery, the kind we use to make both two-dimensional and three-dimensional digital maps. And it flies fast. With a twang and a whooosh the catapult releases and the drone leaps into the sky at a sharp angle, then makes a wide swooping arc out over the water before settling into the familiar zigzag pattern that lets the camera record row upon row of overlapping high-res imagery.
It might not seem obvious from our breezy northern European environs, but this training group is present at the creation of a potentially enormous change in data and analytics for humanitarian relief and development. Over the past 20 months since Typhoon Haiyan hit the Philippines unmanned aerial vehicles have seeped into the workflows of a number of different humanitarian agencies responding to disasters, as well as a smattering of disaster recovery and long-term assistance programs. New means of collecting and using aerial imagery to understand the dynamics of disaster situations and the needs of affected communities are coming into focus. And the field of humanitarian UAVs is finally starting to come into its own, with attendant needs for better knowledge, skills and collaboration structures.
Drones of various shapes and sizes, from the ubiquitous white plastic DJI Phantoms to more sophisticated fixed-wings like the eBee from Sensefly and the Trimble UX5 are now appearing as part of the core assessment equipment being hauled to the front lines of major global catastrophes. Yet we are still figuring out the rules of the air for these vehicles, the use cases which are going to make the most difference, the possible means of collaboration which they may enable, and the subtle but maybe far-reaching shifts in our assumptions about feasible emergency response practices.
A little while later we get our hands on a DJI Phantom 2 to work through a series of exercises in aerial navigation. Discussions while we’re flying dig into the relative merits for relief efforts of something that many people may view as a kind of toy. And it’s true that the Phantom looks surprisingly insubstantial up close. But looks can be deceiving in terms of performance and field utility.
Patrick points out a couple of crucial factors which make this type of inconspicuous quadrotor almost something of a mandatory element in humanitarian UAV missions, even if it’s not the most powerful by far. For one, it just works. With only about an hour’s worth of background, and minimal maintenance and balancing, most people can safely fly the Phantom. This can be a crucial factor when you need to make sure that no matter what, some imagery is definitely captured. For another, the Phantom’s toy-like appearance helps to sooth concerns among local communities, police and customs officials. Slap a few friendly stickers on it and you’re sending a message that this technology is not a threat. It’s these types of social nuances that crop up again and again in thinking through how to apply drones to relief and development work.
Back at the base Laurent puts us through the paces of flight planning and image analysis. We learn how to structure unmanned flight patterns to avoid prohibited ground zones, how to measure the required takeoff and landing zones for fixed-wings, how to ensure proper image overlap for three-dimensional data capture, and by the afternoon of the first day how to begin processing and interpreting the wealth of data which returns to us even through our brief beachfront overflight.
3D point clouds float above our image mosaics once they’ve been loaded into the Global Mapper software, revealing hidden terrain dimensions and even the basis of an exterior building model. Laurent of course is a professional UAV pilot and image analyst, but the simple fact of his being able to produce a complete aerial map in 5cm resolution from start to finish in less than one day is one among several signs that we may well be on the cusp of a remarkable change in rapid mobile image data collection.
A big chunk of the remaining three days is spent alternately exploring lessons learned from recent humanitarian UAV missions in Vanuatu and Nepal, from the flight authorization process to equipment checklists for disasters, and reviewing the core details of unmanned aviation from a standpoint which aims to place humanitarian UAV pilots on roughly equal footing in their discussions with civil aviation professionals.
The two levels of discussion intersect in useful ways. As the hours roll by it becomes clearer and clearer that pretty much anyone who wants to deploy drone technologies for humanitarian missions needs to come up to speed on more than a few areas of knowledge which tend to be unfamiliar to humanitarians. In order to talk effectively with civil aviation authorities in a language they can hear as professionally appropriate rather than risky and amateurish more specialized technical and policy background is required. Whereas the wave of activity over the past 20 months was driven largely by newcomers to unmanned aviation who may not even see what they’re doing with drones as “aviation” per se, often in countries which may not have much of a regulatory framework governing drone activity, upcoming missions will need to take up the discourse of civil aviation proper in order to ensure that humanitarian UAVs are taken seriously and teams are allowed to operate expeditiously.
Of course, saying that one should learn the ins and outs of civil aviation and actually doing so are two quite different matters. By 6pm on Thursday my head is swimming with charts and checklists, aerodynamics algorithms and systems of meteorological coding. The field is complex, and it will require long-term engaged communities of practice to move us towards at least quasi-professionalism in the humanitarian space. But after what Patrick reports about the ups and downs of recent relief missions, particularly in Nepal, it’s crystal clear that some degree of professionalization is the field’s next major iteration.
Just as important as engaging civil aviation authorities as professional peers is the necessity of engaging local communities in understanding what these technologies are and how they can be a benefit rather than a threat. The ability to understand and explain the physical properties of drones as flight and data tools can help to demystify the devices. Combined with the Code of Conduct which UAViators has developed in part to ensure appropriate community engagements, we may be able to avoid unnecessary fears, mistakes and misunderstandings.
On the other side of these discursive divides lies what I would describe as an imminent humanitarian future. It’s not here yet — but the shape of that reality is getting clearer by the day.
None of us present at this initial training agreed that drones in relief and development actually do right now solve the specific problems confronting NGOs and other humanitarian agencies, at least not at a rate that suggests they’re about to replace other means of gathering and processing data. But the promise is enormous, of a day sometime soon when our field assessment survey teams are linked practically by default with rich, customized and hyper-localized 2D and 3D imagery, based on specific questions about terrain distributions, hydro conditions, damage reports and frequently changing ground circumstances and weather patterns. There is, just over the horizon of the present, a possible future where networks of payload drones are relieving at least some of the burden of routine physical transportation of things like essential medicines in remote areas. And there is an exciting future coming into focus where local communities and organizations are working in tandem as partners on these technologies with the organizations tasked with emergency assistance.
There is much that will need to go right for these futures to materialize — and again, we are still in very early days with humanitarian drone applications — but after three solid days of training I came away more convinced than ever that these technologies can and will, if we understand and apply them correctly, change in significant ways our basic assumptions about what is knowable and achievable in humanitarian operations. And that possibility is something I think we should all be rooting for.