DJI-Phantom-Standard-Class-Micro

katie qutub: okay. all right. good afternoon,everybody. thank you for coming to our session, it's a bird, it's a plane, it's a global healthtool. we hope you're excited to learn about the potential of using drones for work inglobal health. my name is katie qutub, i'm a health officer in usaid in the latin americancaribbean bureau. in my extra time, we all know “extra” time, i'm the u.s. uav champion.unmanned aerial vehicles, uav. i work to engage stakeholders within and external to the agencyon the potential use of uavs for our humanitarian and development work, including global health.i've been in the agency a little over five years, prior to that, worked for implementingpartners. i am so honored today to have such a great panel with me. we have brittany johnson;she is a global supply chain consultant at

integral chain, working with aircraft. previously,brit was with the william davidson institute at university of michigan. she has got a master'sin engineering on logistics and materials, and supply chain management from mit and herthesis was on increasing access to medicines in the private sector in zambia and zimbabwe.we have got dr. andrew schroder from we robots. andrew is a world recognized leader in geographicinformation systems, data science and applied epidemiology for humanitarian aid and globalhealth. he founded the nethope.org uav working group and other networks. and he receivedadvanced degrees in social analysis and public policy from nyu and the university of michiganand he's developed and implemented a number of projects with direct relief including mapping,global pmtcts, spatial analysis of the distribution

program and development of gis based laboratorymonitoring in ethiopia, and he’s got a phd social and cultural analysis in nyu. i guessi mentioned that. and then we have dr. doug norris, professorof molecular micro biology and immunology at johns hopkins. he focuses on the most dangerousanimal in the world, the lovely mosquito. he’s been very busy with zika lately. we'regrateful he was able to get the time to be here. he works at microsoft's project premonition,which you will learn about today, on the use of uavs in the mosquito world. and his researchis mainly focused on genetic diversity within and genetic structuring of arthropods arthropodborne pathogen populations. so get really into the microbio.

so here we go. real quick i want to go throughsome vocabulary in case these terms come up while we're talking today. so the common termout there is often “drone”. we talk about drones and people associate drones with thismilitary concept. we will probably mostly use the acronym uav's, unmanned aerial vehicles,that's relating to the vehicle itself. there is also the term, uas, unmanned aerial systems,which include the support systems on the ground that help these vehicles run. we also talkabout rpas, remotely piloted aircraft. so different terminologies, depending on thedifferent types of vehicles and more and more i think we're trying to use terms like “aerobots”.the vehicles themselves, there's different types of vehicles. so here you see both ofthese that you see in the pictures are rotors.

both quad copters actually, so that is theones that look more like a helicopter concept and then you have a fixed wing that looksmore like an airplane. then vtol is vertical takeoff and landing, that is a vehicle thattakes off like a helicopter and flies like a plane, so it has the advantage of a distanceof a plane but without requiring the runway. you will learn a little bit more about thatwith brit. there is also different tools, i don’t think we will get into lidor todayand i forgot to write down the whole meaning of the acronym. different center sensors thatyou're using, laser based radar. payload, we are always talking about payload.we are talking about the weight that something can carry. so it's a uav that's carrying somethingsome distance. some technology terms. one

of my favorite things is sense and avoid,which is technology is being built into the machine, so that you have less concerns aboutcrashes, because they can sense kind of like a bat that they're near something and quicklyavoid it. and then the market das, drones as a service, that would be a company thathas different models of drones available for lease for people to use which is kind of likewhat a three pl, four pl does. ups is three pl or four pl, depending how you contractthem. but they are providing services for people who need logistics solutions. so now we'll have a little fun, hopefully,activity. i've got a goody bag, and we'll get some volunteers to grab something outof the bag, and let us know what they think

it might have to do with a drone. so who wantsto be our first contestant? all right. audience member: what's this got to do withdrones? garbage bag or garbage can. i would. maybe something for balance in the drone? katie qutub: that could be. potential forusing them to distribute food or other commodities. audience member: okay, that's what it is.[inaudible] katie qutub: yeah. so an apple – we gotthe apple in there, because there is a lot of use in the agriculture sector. it’s actuallythe biggest commercial market for uavs, especially in countries like the united states and forprecision agricultural we have sensors that can tell the nitrogen difference in samplesof soil within a two millimeter distance.

huge value in that. audience member: [inaudible] katie qutub: that’s right, i think it'schlorhexidine. so if you had a birth in an area, they're out of stock with some commoditythat could help with the survival of the mother or the baby, you could get it there, dependingon the geographical situation, much quicker than other forms of transport and it’s verylight weight, as you can see. katie qutub: yeah. and that is specificallykind of an old school phone, not totally old school but a razor, it's not a smartphone.and that's because the drones are usually instructed through a phone or another deviceand it doesn't have to be a smartphone, there

is actually sms based technology that canbe used to safely guide a drone to its destination. katie qutub: yep. definitely. and that's beenkind of one of the biggest areas where they have been used so far in the development worldfor surveying disaster, damage after a disaster. i'll take a couple more. katie qutub: yeah. i like that answer, too.that's very good. and then also the sum - one of the manufactures making drones for lightweight, short distance deliveries, they can actually print their drone from a 3d printer. katie qutub: those are both very good, especiallywith an egg, something you have to keep refrigerated. once amazon is at scale, you'll be able toorder things very quickly and you don't have

to worry about temperature control. anotherreason for the egg, is if you have a drone that's making commodities delivery and it'snot the type that lands, it's just going to drop it you don't want it to end up like thategg. so, all right. good. last one. audience member: i have a id badge. katie qutub: yeah. audience member: you could use it for to deliverblood, to villages in need. katie qutub: as long as it's going to a facilitythat is set up with the rest of it too. we were going to use that as an example one dayand we were like, wait, it's not like you can just send blood and someone's like – (laughs)all right. i hope that got our brains thing

thinking a little about our potential here.hopefully you can see the map. i have highlighted some of the countries, i think a lot of questionsare, well, this is so new, and there are a lot of countries that non defense humanitarianand development uses of drones are being tested out or have been tested out or are planning.some of them are formal. we've got some companies that have done some pilot tests, haiti, dominicanrepublic, papua, new guinea, we have work getting started in tanzania, those are allin health logistics. in india, we have some work going on through usaid doing some surveyingof fields for crop damage. they've had a severe drought this year.in russia, there's been some kind of surveying of crowds at events to see so that they candemonstrate the size of an activist community,

by the activists themselves. and indonesiasurveying to show when the government had taken some land, showing that kind of approvingthe case of the community on the land rights. lots of different cases in nepal with theearthquake, there were a lot of uses for uavs for surveying the damage and andrew mighttalk about that a little. so lots of places where different work is being done or is beingplanned. but also a lot of concerns, these are someof the headlines we see, and i think a lot of concern is domestically because peoplesee they think of that big picture in this very crowded american airspace with dronesflying all around and getting in the way of things. and we certainly are very aware ofthose, andrew is going to talk a little bit

about the engagement of the humanitarian communityand the need to have a code of conduct about when and when they are not appropriate toolsto be considered. so with that, we'll get into our amazing presenters. so i don't knowwhy that was animated that way. all right. brittany johnson: all right. do i need tospeak into the microphone? would this be better? so my name's britt johnson, so i'm workingwith integral chain, and in working with them with vayu, vayu manufactures detail enabledfixed wing drones, uavs, so they are autonomous vehicles that have delivery capacity, andthey rise up vertically, and then they fly with fixed wings so they can reach out tocommunities 60 kilometers away. and so what i thought i would talk to you all today istelling you a little bit about my experiences

which, you know, my background's in supplychain but i've been working in global health about the last three years, and so i've gottento see a lot of on the ground work being done in terms of medicine delivery. and i justwanted to share with you all, and then, also, kind of thinking through, you know, what arethe bottlenecks that are out there? and starting with those bottlenecks within supply chain,to then understand a little bit better, you know, what potentials this technology couldoffer and what are some of the right environments and framing of the question to be thinkingabout this as a potential solution. so about a few months ago i was in mozambique,and i got the chance to be on the ground for a vaccine delivery. so i don't know how manyof you guys have had the chance, it was just

this great opportunity to see things happening.you know, we talk about strategy and here people are acting on the ground. and i gotpicked up, you know, probably just before dawn, about an hour later, was outside thecity limits picking up vaccines for a month's worth of deliveries, they were actually goingto happen within a two week window. and we were at the central medical store pickingthem up and putting them in a styrofoam box, you know, one of those you could pick up beforegoing to a tailgate. right? and then the lid was put on top of it. you’ve got some icepacks in there, and then once the vaccines were in there the lid was put on top and itdidn't quite fit all the way, it was open a little bit. it's okay, we're just that'sokay, we take it with strides, and then took

another hour in the car, got to the district.we're up at the district store. and there we dropped off some safety stuff for themto distribute, and then we also picked up the district director for vaccine deliveryin the area. and so she joined us in the car. i think we were about four people, driver,distributor, and then the district supervisor. and we made off – this is probably now asof late morning, made off for the first clinic. we're hoping to visit, you know, two to threethat day. and the first clinic was on the other side of a sugar cane field. so thiswas flat. it hadn't been raining that much, but just enough so that you know the roadswere still pretty muddy. and so the four of us are in this four-wheel car, and we're onroads that are about, you know, built up a

little bit over the fields, and the driveris doing the best to kind of get through. we're at a pretty slow speed. and, you know,and every now and then he's kind of jerking the wheel to make sure we don't get stuckin the ruts, i don't know if you've ever seen some of these dirt roads, but there are rutson the road that if your car gets stuck in them your wheels just spin and spin and youcan't move forward. and so it made sense that he was jerking, but then one of the lady'snext to me got really upset about him and was telling him to stop. i didn't know whyuntil about like five minutes later we practically just steered ourselves right off the road.we're about that close. and the road's about four feet high, you know, built up above thefield.

so if we were to go over that road, you know,there would be no tow truck that could come out within those roads to come and take usback up and get us out. so thankfully we didn't fall over that precipice, and we made it tothe clinic maybe about two hours after leaving the district. and once we were at the clinic,we spent maybe 30 minutes there. we stocked up to a certain level of vaccines at the facilityand checked some of the books and then made our way to the next spot, which was goingto be back across that same road of across the sugar cane plantation, and this was goingto be a clinic that catered to some of the workers that were one the plantation. we madeit there at about 3:00 pm. by about 3:00 pm we finally made the second clinic. and wewere lucky, because they were still open.

usually if you get close to four, everythingcloses, and then you'll just have to try again on another day. and indeed, you know, thosewere some of the stories that i was hearing, that that does happen every now and then.but we made it that day. and this was so these were two clinics of84 within the province. the distribution management that was going on, is probably state of theart, you know, where that i got they're visiting the clinics once a month and they visitedabout 80% of the clinics or more over every month, and then when they arrive at thoseclinics there's probably less than a 10% stock out. so these are generally, within the fieldthese are pretty good statistics. and to visit those 84 clinics, it generally takes themabout two weeks with two cars and then those

cars usually have at least three people inthem. for those two weeks; and then the people that are the distributors, they actually haveanother full time job at the provincial level in terms of managing vaccines and then thisis sort of something they have done in addition to that work. and so, having that background,i kind of you wonder what are some of the challenges here? what are some of the problemsthat make this really challenging, that as well as it's managed there's still so manyissues, in terms of time to get out to take two hours from a district to a hard to reachfacility. qa, quality assurance, making sure that standards are maintained for somethinglike vaccines which can lose potency based on the environment in which they're transported,making sure that that quality is maintained

throughout is really challenging in low techand low resource environments. stress on the staff to take that time to make those deliveriesand then also have other jobs and other obligations that are weighing on them. cost and risk tothe vehicle, those vehicles need maintenance at least once a month immediately followingthose distribution routes to make sure that they're still functioning for the next round.and, you know, on top of those within the network itself, are some of the displacedcosts of not arriving. you know, and what that would mean for the manager of the clinic,who then has to figure out how they're going to get to the district to get that productor to someone within one of those communities served by the clinic when that clinic doesn'thave it available, but they need a product

so that then they might take upwards of twodays to travel back and forth. and this isn't just an issue for vaccines, we see this withdiagnostic tests, emergency goods, even essential medicines just making their way out to someof these sort of low service area places. and the environment that create these supplychain breakdowns for those goods, sometimes it's roads, infrastructure, on land, sometimesthere are communities across water, up mountains, facing these bottlenecks in infrastructure,in weather, cost, reliability and responsiveness of their deliveries that allow them to servetheir clientele. so these are some of the interesting areasthat we get to explore in thinking about flying over the problem. these are some of the wayswhere we can start to think about taking a

flight from that district and reaching a facilityand also reliability and responsiveness of ensuring that those products arrive on thatmonthly basis, allows you to plan so much more around reducing cost and also improvingservice. so, what that sort of lets us think about is, in thinking about mozambique nowand what that could potentially look like, what we're imagining is, you make it out tothat district level. you've got your vaccines, and you call up that clinician, you understandwhat they've got on hand, what products they have, and what they're facing, you ship themthe quantities they need to stock up to, and you can have that carried directly to theirdoorstep within half an hour. so you don't have to worry about them having left the facility,because you just spoke to them about half

an hour ago. we're looking at an autonomousprogramming so that either the carrier at the uav itself knows where it needs to land,and from there, by press of a button, that person at the clinic can just send that carrierback to the district and that district could potentially reach upwards of three or moreclinics within that same day that it took them to reach two. and so there are various specifics about thesemodels, and some different cases where we're looking at the implementation. and we're thinkingright now, looking at some, using this to get tests back, for instance, in papua newguinea and thinking about the application to tb, if you can get a test back reliablyfrom harder to reach areas within a certain

amount of time and then provide those responseswithin a day, within less than 24 hours. also looking at potentials in places like the philippines,where maybe that delivery requires a boat and that boat only leaves every so often orthat boat is very high cost, what would be the options of sending some of those essentialmedicines, instead, by air. and these are some of these cases and i'd love to talk toyou more. i also have some models of our craft here to kind of play with. and these are someof the case studies, i'd also love to talk to you more about. so i'm just including here,this is my contact information. i’m really thinking about applications on the ground.i also have our founder ceos information there, so if you're thinking about a specific placethat you might be interested in implementing

this, he's also available to take your questions,and i would also love to go into the specs and technology behind it but i will keep thatfor a one on one based on interest. all right? should i take questions? katie qutub: thank you. yeah. we'll do questionsprobably at the end. andrew schroder: so my name's andrew schroder.like katie was saying at the beginning, i'm the co founder of werobotics, which is a neworganization focused on the application of robotics to humanitarian aid development,ecology and global health. i wear a couple different hats including this one and so i'malso the director of research and analysis for direct relief and what i'm going to talkto you about here is a story about the application

of aerial robotics to community health workers,which may not be an obvious connection and build that up in a few layers through a kindof history of my own evolving thought about why this makes sense. and it makes sense.i'm not going to touch on delivery questions, although that's an obvious fit for communityhealth workers, but i'm going to talk about it primarily in terms of spatial data. thestory starts for me a couple years ago when at direct relief we were working, we'd beensupporting a group called last mile health for a while and we helped them through theenabling access to just simple gps devices to begin mapping out their catchment area,the service area in konobo, liberia way out in the middle of nowhere in the southern partof the country. it's a place that did not

have a roadmap prior to doing this, this isthe first time that they had mapped that out. and it was essential for them to do becausethey were building a program for being able to enable health access to areas that previouslywere disconnected from health access per se. so that is the point of the map. it’s afairly rudimentary map, but it’s actually a pretty cool map. i like it. the purposeof building this map then dovetailed over into data collection for them. so being ableto do household contacts, household surveys, we supported them through a baseline surveyof the entire area that then was able to go and look at the household level at key conditions:demographics, health needs, the survival of children, things related to disease incidents,what have you, and the roadmap became essential

in being able to guide that survey. it alsoflipped back around into analytics because now they could actually begin to link theresults of those household surveys into mappings to be able to understand very granular patternsat the household levels, so distributions of certain kinds of conditions across spaceand the role that space played relative to access questions, relative to a number offactors that were leading to the need for community healthcare workers. over time, and we were able to connect themup, their need grew, during the ebola crisis i began talking to them about the potentialapplicability of uav's to begin getting the maps much more granular than what were currentlyon offer. so if you saw that first map, it

tracks the roads and the rivers but it doesn'treally track the households and that was all being done through hand drawn paper maps actually.as they would send workers out, the workers were guided to the households by hand drawnpaper maps. we thought about putting uavs into the field during the ebola crisis, idecided that i was a little crazy for that and it would be too complicated to introducenew technology in the middle of that kind of scale up. but working with open streetmap and humanitarian open street map and a number of others we were able to expand theirunderstanding of the area around them. so that initial map began to grow and the demandfor spatial data began to grow with it. kind of around the same time, my colleague, patrickmeier and i have been running and organization

called you aviators, which is a humanitarianuav network as katie alluded to in the beginning, we have developed a code of conduct, i cantalk to you more about the sort of safe, ethical effective uses of this kind of technologyfor humanitarian aid operations in particular. after the nepal earthquake, where there weretremendous challenges in deploying the technology, getting the rights to use the technology,getting it into the field, getting the people to understand why the technology is beingused, we went back and did some additional disaster, post-disaster assessments and communitybuilding, using uavs and were able to work in the town of ponga, with the support ofour colleagues at dji, to just take simple dji phantoms which are off the shelf aerialrobots that can fly for about 25 minutes and

did a number of flights around the town ofponga at about 15 to 20 centimeters revolution, compiled them into the map you see here, workwith the community to tie that map up, identify points of interest, build up a community understandingof the needs for rebuilding in ponga, and, really produce is kind of community drivenrebuilding process linked into the appropriate ministries in the nepalese government, thelocal disaster response council, other ngos that were also working on rebuilding, andi started to think, this is actually a three dimensional model that was produced off ofthe drone. so a uav flying those same missions that i showed you produces something calleda point cloud. and that point cloud allows you to build a three dimensional model. itmeasures the distance between the drone and

the ground and the features that it's flyingover, so that not only a model, not only a map of the area, but a model of the buildingsthat were damaged, this was produced in about two days. by people that know what they'redoing. i started to think about the applicability of these two use cases, right? so you've gotlast mile health and the need for spatial data, community health workers and then thisexample of community driven mapping and rebuilding, and then i spent some time in sierra leoneand traveled around these absolutely terrible roads which are just exhausting to go overon a normal basis, and went out to a number of health clinics, and because i spend a lotof time making maps i look for maps, like, where are people like actually using maps,where are they making maps, turns out that

when you go out to health facilities, especiallyhealth facilities that are hosting community health worker programs, you find maps everywhere.people are using spatial data all the time. this is a catchment area map from a facilitythat is similar to the ones i was describing to you from last mile health. it is showingthe routes that they take, the traffic directions, the households, there is a meaning to thenumber of households there, they are counting up the people that they are actually doingoutreach to. here's another example. this map is not updated. it's been up on thewall for a long time. but people still use it, really, all the time. they use it to understandthe distribution of their service areas and to really show people that come into the clinicswhere the patients are coming from and the

patients use it to understand their own servicearea. this is another example. this is actually a pretty good on from the road network; notthat inaccurate in terms of the road network but it's a little hard to read. and, again,you know, another one. i actually was going to make a book called catchment maps whichi thought would be a nice like coffee table book that you could look through just becausei like that stuff. but we can do better than that. we can actually apply aerial roboticsto this use case to improve the kind of information that's being used by the people that are onthe front lines of the most serious health crises in the world. we can enable them tomap their catchment areas accurately. we can enable their planning, so that as they thinkthrough where they ought to be working, that

that's actually linked into data structuresthat are useful for being able to plan service areas. we can enable household level mapping,so that within most of the areas that you're looking at, those service areas are not morethan about 1 to 2 kilometers in their totals, they're disconnected. there's a lot of spacein between them, but the village itself is actually not that bad. so with a few passesof a dji phantom, you could create household level maps that are quite accurate. and youcan connect them into an analysis. there's a work flow then that's fairly simple.you can, using a smartphone, program it to fly in a pattern that allows for overlappingimagery, produce a map. you feed it into open street map. the international community actuallymaps these. so, individuals on the ground

may or may not necessarily do the taggingof the households themselves, but other people can actually add that imagery. so this isactually a town. i should show it to you before the ebola crisis, there was literally nothingthere in terms of what had actually were been mapped. this is all household footprints.this was done by satellite imagery, but if you go out and around you'll see there's ahuge number of areas where none of this has been done, which are the areas for catchmentfor last mile health and that enables workers to get out there in a timely fashion withlocal knowledge about what's actually going on in the households and analytical knowledgeabout changing patterns over time. that creates all kinds of structural questions about howdo you actually do that and what are the partnerships

and what are the funding models that needto be in place in order to make that work happen and that's one of the reasons why weformed werobotics, the role of werobotics is to catalyze innovation labs called flyinglabs around the world in places which are experiencing high risk, high need on a numberof different fronts and to drive the projects through local demand to make sure that wehave people that are using data, people that are able to sustain that data. to be ableto sustain ecosystems of technology and organizational structure around the application of roboticsto these particular problems, so that no one can do it alone, but we have the ability todo a lot better than what we're doing right now. you can see my contact information upthere. again, i also work for another organization

but you should check out werobotics and wewould love to hear from you. thanks. katie qutub: thank you. doug norris: my name is doug norris, i’mup at johns hopkins and unlike my colleagues, i'm not a drone pilot or an engineer. i'mbiologist, which begs the question, why am i here? and i think the answer is, as manyof you probably are like me, more of an end user than a developer, there's lots of applicationsfor these tools and i think that's really the important thing to keep in mind, is thatuavs bring a really new way of looking at things, looking at our environment and collectingdata, and supplying new tools to them. so my world is the world of infectious diseases.i don't have zika up here, i'm sorry, but

there's not enough data on that yet. but justa reminder that emerging infectious diseases, which is really the world i work in most ofthe time, have huge health, certainly and as well as economic costs, so what you arelooking at as far as flu and ebola, you're looking at the u.s. allocation for these diseasesin green. so a lot of money. the number of cases in red. okay, the number of cases inthe u.s. in red. okay, we spend a lot of money for emerging infectious diseases, this isa big concern. but staying ahead of them is one of the big problems. the majority of thesediseases that we look at, the majority of the emergence events come from mostly animalpopulations or out there in the environment, so they are either domestic animal populationsor more likely than not, wild animal populations.

so they're circulating out there, occasionallythey invade into us into human populations and they cause big outbreaks of the diseasethat we are concerned about. but we are almost always reacting to these events, the zikaright now is a great example. we are reacting. we're not ahead of it. we didn't know whatit could do and we still don't know what it can do. so there's lots of work to be done.and so, not only are we reacting to these, but many of what we find are either new variantsof the pathogen we already know about or brand new pathogens that we know nothing about,so we're always sort of behind the ball. so the idea behind project premonition, whichi'm part of, is to get ahead of the game. and to detect pathogens in natural environmentsbefore they emerge are reemerge as big outbreaks.

the idea here, since we know most of thesepathogens exist in natural systems already, they are already in animal systems, naturallyoccurring in wildlife systems or even domestic animal systems and then they sort of leadover to us, they're already out there. and there's already a biological tool out therethat is fantastic at sampling these biological, at these animal environments, those are mosquitoes,which is where i spend my life working on. mosquitoes. so mosquitoes are sampling thesehosts all the time, picking up pathogens all the time and the trick is being able to getthose mosquitoes back in the laboratory, sampling them for pathogens and being able to makepredictions about what's going to emerge next and where it’s going to emerge next. sothe idea here is really pretty simple. we

have our mosquitoes out there already. wedon't have to deploy them, there's plenty of them out there. everybody complains, butusing uavs to sample the mosquito population, bringing those mosquitoes back into the laboratory,where we can use other technologies like deep sequencing, through those mosquitoes thatwe detect in the field and detect the pathogens before they emerge into very apparent localinfections that we're constantly dealing with as a reactive basis. so just as proof of principle, i was luckyenough to i'm the biologist on the team. there's a geneticist on the team and a whole bunchof engineers on the team. and so earlier last year, i had the opportunity to take all theengineers out and we went we chose the island

of grenada, and we chose the island of grenadafor a couple reasons: one they would let us fly drones, two, it's relatively small. sowe could cover the whole island in about a week, and the idea of what you're seeing onthe left hand side is that we would place 50 traps at 30 locations, throughout the island,covering all different kinds of ecologies and biology’s and environments from thebeach, which was really nice, up to the more mundane environments. to do this, we tooka team out, we had five days to do this and what they discovered, what i was able to teachengineers, they were able to teach me a lot – but what i was able to teach engineersis biology is really hard to do. and deploying traps is really hard to do even with an experiencedteam, which i work with all over the world,

this is really hard to do. so in five dayswe were only able to deploy about 30 traps overall, so we didn't hit our target of 50.i think on one of these next slides, the average time in the field every day, so this is notonly deploying traps, but picking up traps that we deployed the day before, was about18 hours. so these are long days and this is just trap deployment and trap collection;this does not include actually processing the material that comes into the trap, whichusually takes several more hours, anywhere from 30 to 60 minutes per trap to sort outmosquitoes from everything else that got in the trap that you didn't really want.and so it's a lot of time, a lot of manpower, a lot of effort that goes into just collectingthe sample. and this is even before that processing

step, which is a whole different discussionfor maybe another seminar sometime. i also want to mention well, i skipped ahead, buteach of those traps you saw weigh about 16 pounds. so the idea here is to have a mosquitotrap that we can deploy by drone, okay, and so these are just quad copters like you havealready heard about. the payload on these is - what's the payload? katie qutub: minimal. doug norris: minimal. it’s pretty tiny.it is certainly not the average 12-16 pounds per current trap. so we need a new trap design.we need a number of different things. since this is a uav talk, we will stay with uavs.so we flew uavs at every place we placed the

trap. so the traps were placed manually, notby drones, because a drone placed trap doesn't exist yet. we're working on that. and we flewthe drone at every place, because what we're trying to do is teach the drone how to thinklike me. i'm a mosquito biologist, i can go out to the field, look at the environmentand choose where to place that trap within the environment. even if we choose locationsby satellite or other tools, once we actually get out there, trap position is pretty important.and so training a drone to think like me, turns out to be pretty hard. which i'm happyabout, because i can stay employed, but if we want to do this at scale we need to beable to do this. and so what we did is at each location where we placed the trap, istake the drone or the uav, fly it at two different

altitudes and essentially square route, anduse the camera, the high definition camera on the drone or on the uav, to image the environment.and then much like you just heard in the prior talk, use that imagery to recreate a threedimensional depiction of that environment which looks sort of like this, we have thedrone's flight path or the uav flight path and environments, and it does a really goodjob at recreating the three dimensional environment. that in combination with other forms of remotesensor data, we can look at habitat type. we know that there's water underlying this,we can actually model hydration. all the other kinds of things that there can, thereforebe modeled. and so the basic idea is now the drone knows what is mosquitoey based on samplingand what is not mosquitoey, and then, of course,

through high tech computing be able to pushthe numbers and be able to teach these uavs what looks like mosquito habitat and whatprobably is not mosquito habitat when it's selecting actually trap placement in the future.so here is just another vision, this is actually what the drones, or the uavs are actuallylooking at. so to me, even somebody who's relatively experienced with this, it all lookspretty green, but the cameras and some of the other technologies that we can mount onthese uavs can see a lot of things that we can't see. so this is actually one of ourmost productive mosquito sites, it's a junkyard with a tidal base right next to it, some saltbrackish water. a lot of things that i can't see, but the uavs and cameras can see theseand we can model these and so now we're going

back through the process of sampling or processingmosquitoes from these locations, going through the genomic pipelines to see if we can detectpathogens and make predictions of where we might see those pathogens popping up basedon the imagery that we can get from these units. and then, as i usually do, i get aheadof myself so we can plot, those are mosquito counts by sites, over 30 sites, none of themlook exactly the same and none of them catch exactly the same kinds of mosquitoes. theyvary from very rural to relatively urban, and we can, again, reconstruct all these offthe camera. the camera data as well as the uav paths and reconstructed the habitats thatthese are then found in, and link that back up to not only mosquito data but pathogendata as that data becomes available. so, again,

this is a product of a really multifacetedteam ranging from engineers from microsoft and academia to a few biologists like myselfand genetic and genomic experts as well. and with that i'm done. katie qutub: awesome. well, thank you, guys,for everything. we only have, i think, ten minutes left. i think an important messagebefore we open up to q&a is that the tech experts, the engineers who are designing theseuavs, aerobots, whatever we want to call them, they're not global health experts and so weneed global health experts to be engaged to be helping these companies apply the technologyin the right way. and for what's really applicable to the setting. so we encourage people totry to engage, and bring up their concerns

and remember that we're just talking aboutthese as a tool to help accomplish goals and not the only thing or they're going to changeeverything. but just think of how, if you want to do something, maybe either as a uavused for that to make it easier for you. so, thank you, everybody, and let's does anybodyhave any questions? step up to the mic, i guess. was there no new information today? audience member: hi. thank you for being heretoday. my question is about cultural sensitivity in communications around some of the implementationof the tool. how has that affected the way you've designed it, has it at all, and whathave you learned going forward? andrew schroder: i mean, is this thing on?i think it was on, okay, so the first thing

to note is that community engagement culturalsensitivity is one of the sort of core principles that we've arrived at for the code of conduct.so this is something that's been a problem in a number of applications in the past, buthas been recognized and has really been built into the recommendations for the humanitariancommunity as something that needs to kind of be foregrounded in terms of thinking. problemsi will say, for instance, where in nepal, you know, people that would fly their droneover a community, do a damage assessment, not tell the community what they were doing.another group comes in, they fly their drone, they don't tell the community what's goingon and another group comes in and finally the community is like why the hell is everyoneflying these things over our village and nobody

is talking to us, this is like common senseactually. it sounds, just anthropologically, but it's just tell people what you're doing,and when you do, you'll find and the red cross has actually done some interesting work onthis, that actually people are pretty sophisticated about being able to draw distinctions betweendifferent technologies and use cases, the distinction visually between a military droneand a commercial drone are enormous, they look nothing like one another. and i thinkthis is broadly understood. people, you know, even in pakistan where they have absolutelygood reasons to be worried about drones, they also do wedding photography. and they usedji phantoms and a lot of the things we've been talking about and people have a fairlysophisticated understanding of the distinction

between one and the other. and that is thedistinction is growing and evolving and changing. i think that the broader question is alsoabout policy, not just community, and how we engage the policymakers to be able to createrule structures that will enable these kinds of things to take place, ensuring communityengagement is a step along the way towards ensuring strong and flexible and viable rulestructures. brittany johnson: a little bit and also fromthe side of manufacturing, andrew's absolutely right. you know, we're conscientious i'vegot some models over there, they're painted white. there's a consideration of color anddesign, and, in terms of delivery, you know, the look of it is quite different from whatyou would see in, you know, some of the more

like disturbing imagery that can be associatedwith drones and then certainly with quad copters as well, there is a different feel there.and there's an interesting study that recently came out from fhi 360, where they were intanzania and an accessibility survey, they were just asking people around the environmentabout, you know, how they felt about drones flying overhead and what it meant to them.and i think it was below 10% where they even considered it for military application. imean, that may be part of that region, and there are certainly going to be specific needsby region, but nonetheless, you know, this was something that we had certainly been discussingquite a bit and so to come out and see that survey, you know it added a little bit moreyou know color to the conversation.

audience member: so you touched on policy,just a quick question about how different countries' federal aviation authorities differand what type of challenges you might have come across in the different places you'veworked so far? andrew schroder: i think this is in many waysthe key question right now in terms of the application of this technology to the problemsthat we're all interested in. and i wouldn't limit it to the aviation authorities, actually.i think one of the things you'll find is that there are a lot of different government ministriesdepending on the country, that feel as though they have some level at stake in the permissionsprocess for allowing these kinds of applications going forward. and just as a quick example,i mean, in, we're returning to nepal in a

couple weeks to do a fixed wing project thatcovers a much broader area about 50 square kilometers, the permissions process in thatcase goes through the civil aviation authority, the home ministry, the information ministry,which has to authorize the use of a camera and radio transmission, which is independentof flight, or the area of flight, and the culture ministry, because in nepal it hasbecome very sensitive to image areas of archaeological or touristic significance or other kind ofcultural significance that may have been damaged. so they all have a say in that, plus the policehave a say in it. and the police have local jurisdiction, which may vary from the nationalgovernment. so you have to be very clear on who you're asking permission for, how whatkind of role they have. and as we do more

of these health applications, the ministriesof health have to be included in that process, and they don't really have a good understandingof the technology yet. so i think it's really centrally important that we work with ministriesof health to help educate them about, this is the role of the code of conduct, for instance,to have codes of conduct in place to show that these are appropriate uses as to howto proceed. different countries follow different rules, often, they take their cue from thefaa and from the eu, all of those are in flux. many cases are there are no rules and theyinvent them on the fly because there are use cases that present them in the short termand people have to come up with a rule structure off the top of their head. that's going tochange dramatically over time, but i think

the better we are connected together, thebetter we can make those rules work. brittany johnson: i would just add here, too,so far you've been doing work in global health. and so for that, you know, we're also lookingat health delivery, which, of course, overlaps with ministries of health and the work thatthey are doing. so, certainly, reaching out to a ministry of health, sort of understandingwhere the overlap is and getting them onboard to approach with us, you know, a departmentof defense, a civil aviation authority, you know, whoever is required to provide thatthat authorization. audience member: hi. elisa ballard, usaidoffice of hiv/aids. so it's a fascinating presentation and i'm so happy to learn aboutthe technology. with respect to delivering

commodities, i have two questions, and oneis: how do you minimize the risk of those commodities getting into the hands of notthe intended target. and then, secondly, has there been studies on cost effectiveness?you haven't mentioned anything about, you know, what kind of costs are involved here.thank you. brittany johnson: sure. thanks for your question.so in terms of minimizing risk, one of the benefits that this offers is that there'sgps tracking on these, that's part of the way that they can autonomously arrive at theirintended destination. so you would already have sort of an area, in our case we're lookingat a wing span of about nine feet. so you have an area that is already determined thatit will land and you've probably called ahead

of time to say, you know, we're sending thisout to you by phone, and the vehicle will arrive then within half an hour or less. soi think between those two aspects, there is a communication element there, and, you know,specifically where that vehicle is landing, there's no doubt about it that it is landingat that designated spot next to that clinic. i think that helps to minimize the risk. andcertainly, potentially even reduce it from the current system of, you know, where youhave a vehicle going out, and, you know, there is a lot of unknowns in that. that delivery.the second question in regards – cost effectiveness. there have been some studies where we're lookingat cost. it really depends on the problem at hand that you're looking at, which canalso be a case by case issue. but some preliminary

studies have been done for instance comparinguavs to motorcycle transportation, and in these cases if you look over the life spanof the uav, it's actually quite a bit lower cost depending on the vehicle that you chooseto use, of course. but for making regular deliveries, those costs over that five yeartime would actually be less than a motorcycle potentially. katie qutub: it's an important thing thatwe, with colleagues in the supply chain office at csl at usaid and others, we do want tounderstand the cost and quite often if you look at, if it's in full use and a uva couldbe running fairly autonomously 24 hours a day, making trips. and it's making those lightpayload trips, so that gets out some of the

risk, too. it's not necessarily doing a fullmonth of distribution, so you have less product, less risk. from the cost it depends on howfrequently it's used and what it's used for and then beyond the cost effectiveness itis also the life saved if you are getting emergency supplies to someone that wouldn'thave gotten anything in the first place and so it's capturing that and we would like tosee more implementation studies that are looking at actual what real life implementation wouldlook like. and helping, working with companies like vayu and others to get that data. audience member: pardon me if this is an inappropriatequestion, but i imagine there is a range of costs, but give us an example of the costof one of these uavs that we've been talking

about. katie qutub: i can give an average of acrosscompanies. so the different models that we have seen, they can carry from a range ofabout 15 pounds 60 miles, so imagine like a vehicle carrying a watermelon around dc.at kind of an average price range of about $10,000 for the vehicle and then the operatingcosts for the field recharge and then minor maintenance. so, but it definitely, thereis an algorithm thereof weight, distance, you know. andrew schroder: just one quick question aboutcost vectors, it's a really important question, the effectiveness part has to be factoredin, so where the studies are focused right

now is just on accounting costs and ultimatelyas you get more into the economy metrics of this, you'll get more into the role of improvedfunctionality in certain parts of the logistics system has in terms of improving health systemfunctionality, saving costs in other areas of the system due to decreased frequency instock out, being able to fill short term gaps, doing all kinds of other things, it doesn'thave to supplant other kinds of means of doing logistics, for right now it's a niche technology,but i wouldn't want to get sort of caught on the idea that, you know, accounting costis the only way to factor that in. i think we need to get more into what the other savingsare in the system by doing, you know, logistical alternatives. that just hasn't been done yet.so...

katie qutub: i think we've gone over our timeby now. so, thank you, all, for coming, you can let us know if you have any questions. (applause.)