"Shaping the Future Rise of Drones"

CAS Author

10/11/2019 Added

29 Plays

Description

CAS Inquire lecture by Carrick Detweiler. The second in the series around the theme "Rise of the Machines." Visit https://cas.unl.edu/cas-inquire for more information.

Searchable Transcript

Search:

[00:00:04.050]Thank you all so much for coming to the second installment
[00:00:07.470]of this year's Inquire lecture series, Rise of the Machines.
[00:00:12.260]The Inquire program is structured around these lectures
[00:00:15.300]allowing students, faculty, and staff the opportunity
[00:00:19.210]to investigate how we as a society engage with technology,
[00:00:24.000]not just its benefits or negative aspects
[00:00:27.680]but the nuances and effects of our cyborg entanglements.
[00:00:33.317]It engenders the opportunity to investigate learning
[00:00:37.390]from the fascinating research of our faculty members
[00:00:41.030]in the College of Arts and Sciences
[00:00:43.240]and it allows students to see the various disciplinary
[00:00:46.540]and interdisciplinary approaches to the study of a topic.
[00:00:50.880]For a distinctive set of intellectually curious
[00:00:53.860]and inquiring students, the Inquire scholars,
[00:00:56.630]if y'all could please stand,
[00:01:02.790]it allows them the chance to foster
[00:01:04.920]a meaningful academic engagement with the topic
[00:01:07.730]through discussions with the lecture faculty
[00:01:11.080]and through year-long research projects.
[00:01:14.290]These students are edited along the way by the peer leaders,
[00:01:17.310]if y'all could stand,
[00:01:22.098]and if you could please join me in thanking them
[00:01:24.890]for their participation in this program.
[00:01:26.371](audience applauding)
[00:01:31.110]Tonight I have the pleasure
[00:01:32.500]of introducing Dr. Carrick Detweiler,
[00:01:34.960]the Susan J. Rosowski associate professor
[00:01:37.780]in the Department of Computer Science and Engineering.
[00:01:41.040]Dr. Detweiler co-directs
[00:01:42.510]the Nebraska Intelligent MoBile Unmanned Systems Lab
[00:01:46.720]and he is also the co-founder of Drone Amplified
[00:01:50.750]which sells drone-based fire ignition systems
[00:01:53.770]to assist in controlled burns.
[00:01:56.600]His talk tonight, "Shaping the Future Rise of Drones",
[00:02:00.000]explores how drones, which gained notoriety
[00:02:02.880]from their use in bombings associated with conflicts
[00:02:05.950]in the Middle East, can be used for good,
[00:02:08.770]aiding not just farmers and scientists but society at large.
[00:02:12.750]Please welcome Dr. Detweiler.
[00:02:14.736](audience applauding)
[00:02:19.500]Thank you, so apparently if you're in this class,
[00:02:22.340]you have to be in the front unless you're the leaders
[00:02:24.420]or the future leaders, is that why you two are back there?
[00:02:28.245](audience laughing) (Carrick laughs)
[00:02:30.200]Okay, so we can see what happens there.
[00:02:32.651](audience laughing)
[00:02:33.540]Yeah, so it's great to be here.
[00:02:35.450]Thanks for the kind introduction.
[00:02:37.080]My name's Carrick Detweiler and I'm an associate professor
[00:02:40.330]in the Computer Science and Engineering Department
[00:02:43.120]and today I'm gonna talk to you a little bit
[00:02:45.470]about the history of drones,
[00:02:47.970]also some of the work that we do in the NIMBUS Lab
[00:02:52.220]related to developing new types of applications
[00:02:55.580]of drone technology, and I'll at least give a little bit
[00:03:00.590]of thoughts on where, maybe, they may be going,
[00:03:03.320]although I'd love to get your input
[00:03:04.740]and at any time in the talk if you wanna interrupt me
[00:03:07.360]with questions, feel free.
[00:03:08.510]I'm happy to try to answer them.
[00:03:12.930]So I've been here since 2010.
[00:03:15.970]Before this, I did my undergrad at Middlebury College
[00:03:18.090]which is best known for its actual language program.
[00:03:22.520]I was one of the few students who ended up
[00:03:24.870]actually learning computer languages
[00:03:27.780]as opposed to foreign languages,
[00:03:30.420]and then I did my PhD work at MIT and came here.
[00:03:35.730]And really my research has always been centered around
[00:03:38.960]how do you get robots that might work really well
[00:03:42.370]in kind of structured lab environments
[00:03:44.470]to work out in the real world
[00:03:46.620]where there's a whole lot of uncertainty?
[00:03:49.240]And really one of the main things that motivates me
[00:03:52.400]is actually solving real-world problems
[00:03:55.560]that are impossible to do now with robots as we know them.
[00:04:02.100]Just a little bit more about the NIMBUS Lab,
[00:04:04.910]so there are actually three faculty co-directors.
[00:04:10.100]So Dr. Brittany Duncan and Dr. Justin Bradley and myself,
[00:04:14.150]and we work with a really good group of students
[00:04:17.620]both at the undergraduate and graduate level.
[00:04:19.780]So if you're interested in learning more about robots
[00:04:23.650]and getting involved in the robotics research,
[00:04:25.810]definitely feel free to send me an email and we can chat.
[00:04:31.575]We also do a lot of different types of research projects
[00:04:36.320]from National Science Foundation to USDA
[00:04:38.970]to DoD-type projects, and when I started the lab
[00:04:46.170]about 10 years ago, when I said,
[00:04:48.567]"Okay, we're working with drones,"
[00:04:50.113]I actually tried to avoid the word drone
[00:04:52.560]because when I said drone, most people actually thought
[00:04:57.030]of this type of thing.
[00:04:58.320]So 10 years ago when I said drone,
[00:05:00.950]people were like, "Oh, these military vehicles,"
[00:05:04.030]and actually for that reason,
[00:05:06.380]I strongly avoided using the word drone
[00:05:09.030]and instead used terms like UAS or UAV or aerial robot.
[00:05:14.980]But then actually, Lisa from the Public Policy Center and I,
[00:05:20.200]we did a study actually, and we've been doing this now
[00:05:23.700]for a few years, asking people what they think
[00:05:26.050]about various terms and it turns out
[00:05:28.280]that nobody knew what UAS what
[00:05:29.960]but a lot of people knew what a drone was.
[00:05:32.800]And as this has evolved, we now, when you think of drones,
[00:05:37.700]you probably think of things like this,
[00:05:40.240]these little things that hover around our heads
[00:05:42.470]and we wonder what they're doing
[00:05:43.668]and why people are flying them,
[00:05:46.430]or in Nebraska, there's a lot of applications
[00:05:48.900]of actually using drones to better understand
[00:05:53.260]how our crops are growing, to help aid precision agriculture
[00:05:57.380]where you're going to actually apply fertilizer or water,
[00:06:00.280]detect pests before you can visibly see them.
[00:06:03.760]So there have been a lot of changes.
[00:06:05.430]So how many people have flown a drone?
[00:06:11.240]Okay, so ignoring my students there,
[00:06:14.420]a few of you but not too many.
[00:06:17.060]And I wanted to start out actually
[00:06:18.310]by talking a little bit about,
[00:06:19.730]just a brief history of kind of how drones came about
[00:06:22.897]and in particular how they've transitioned
[00:06:26.290]from military applications to these other types
[00:06:30.530]of applications from just for fun
[00:06:33.190]versus our scientific applications.
[00:06:36.420]And if you go read on Wikipedia and other places,
[00:06:39.470]it actually turns out that probably this 1849 attack
[00:06:45.570]of Venice, this is kind of the first unmanned military use
[00:06:49.380]of flying things where they actually launched balloons
[00:06:53.560]and dropped explosives on the city.
[00:06:57.860]And so this is, again, these kind of developed
[00:07:02.300]through the military and when airplanes were first developed
[00:07:10.840]shortly thereafter in World War I,
[00:07:14.010]they actually had a project in World War I
[00:07:17.930]when first major use of airplanes
[00:07:20.660]and they were still not widely used,
[00:07:22.610]and any historians out there,
[00:07:24.430]please correct me on all these things.
[00:07:26.660]But they actually worked on trying
[00:07:28.110]to automate these aircraft
[00:07:30.390]and they were using a lot of new technology.
[00:07:32.690]So they were using aircraft which were new.
[00:07:36.140]They were using radio transmitters.
[00:07:38.250]So they're actually trying to send radio commands to these
[00:07:42.920]and another thing they were using was a gyroscope,
[00:07:46.930]which, a gyroscope basically tells you the rate
[00:07:49.160]at which you're turning which is important for an aircraft
[00:07:51.570]to know if you're level or if you're actually starting
[00:07:55.050]to turn into a bank.
[00:07:56.580]And so they had to use gyroscopes on here
[00:07:59.120]to actually get them to work.
[00:08:00.900]They were not widely successful
[00:08:03.390]but they did actually try to deploy some of these.
[00:08:05.200]I think they said these were able to go,
[00:08:08.010]let's see, I wrote this down,
[00:08:09.560]they were able to at least demonstrate
[00:08:12.290]that it could fly 30 miles and drop, I guess in this test,
[00:08:17.440]a bag of sand within two miles of the target.
[00:08:20.280]So, not great, but given other options at the time,
[00:08:24.807]that may have been good.
[00:08:30.718]World War II, there are further developments,
[00:08:33.730]but actually during the Cold War,
[00:08:36.070]they took these World War II aircraft
[00:08:38.360]which they had actually had automated to some extent
[00:08:41.790]where they could remotely pilot them during World War II.
[00:08:46.110]But they actually used these during the Cold War
[00:08:48.520]to fly through mushroom clouds
[00:08:49.990]and actually collect data on the atmospheric conditions,
[00:08:53.630]radiation, and all those things
[00:08:55.390]where you wouldn't wanna send too many people through.
[00:08:59.200]So a lot of development on the military front,
[00:09:01.730]kind of around the same time,
[00:09:04.110]RC aircraft became popular as a hobby.
[00:09:08.210]So you could now actually build small aircraft
[00:09:12.490]with remote controllers and be able to fly them.
[00:09:16.580]And this, I'd say, really,
[00:09:20.655]in the '50s and '60s, this became very popular
[00:09:24.410]and lots of clubs forming, competitions,
[00:09:27.820]doing this to try to both build
[00:09:29.810]your model aircraft and fly them.
[00:09:32.910]It was probably, it wasn't until actually kinda the '80s
[00:09:37.890]that they began to really pick up
[00:09:40.202]in their use in the military, again,
[00:09:43.840]in part because Israel effectively used them
[00:09:47.660]in the war in 1982 to actually confuse
[00:09:53.040]a lot of the anti-aircraft systems that Lebanon had.
[00:09:56.290]And I think this is the point
[00:09:57.920]that at least if you read a little bit online
[00:10:00.880]that you'll find that people say, "Okay,
[00:10:02.527]"this is when the the U.S. military said, 'Okay,
[00:10:04.683]"'we need to do this, we need to do it better.'"
[00:10:06.920]And lots of countries really started to militarize drones
[00:10:10.930]even more than they had done before.
[00:10:13.000]So then basically that point onward,
[00:10:16.040]everyone wanted one and they're used widely,
[00:10:19.960]and this is why even just 10 years ago if I said drone,
[00:10:23.810]people would think of the Predator drones
[00:10:25.450]that were flying in Iraq and Afghanistan
[00:10:28.500]and all of these military environments.
[00:10:32.750]But what I haven't actually answered
[00:10:34.590]is how did we get from this,
[00:10:40.230]these large fixed-wing aircraft
[00:10:43.880]to these little things that can fly around your head?
[00:10:48.390]So any ideas on what enabled this transition
[00:10:52.380]from these big things to these little things, yeah?
[00:10:55.614]Lighter weight lithium batteries.
[00:10:57.760]Yeah, so, okay, so lighter weight, yeah,
[00:10:59.970]lighter weight lithium batteries, so yeah.
[00:11:02.238]This thing is like a traditional aircraft, lots of fuel.
[00:11:07.660]This thing, yeah, runs on batteries.
[00:11:09.870]Okay, what else, yeah?
[00:11:12.493]Change in the flying (speaks faintly).
[00:11:15.975]Change in the flying what?
[00:11:17.155]Changes in the flying style.
[00:11:20.653]Yeah, so the flying style, yeah, so how you could fly,
[00:11:23.610]so controlling these, you can fly one of these in this room
[00:11:28.410]because you can hover it, right?
[00:11:31.400]That one, well, you could fly over the States.
[00:11:35.220]Yeah, another idea?
[00:11:36.249]I'm gonna guess
[00:11:37.310]higher resolution cameras.
[00:11:40.668](speaks faintly) High resolution cameras
[00:11:44.087]in the model now (speaks faintly).
[00:11:45.960]Yeah, so high resolution cameras, certainly.
[00:11:49.340]It makes them more fun because you can actually do something
[00:11:52.470]and see what's happening and things, yeah, another idea?
[00:11:55.579]Smaller computer systems
[00:11:56.974]and programs capable of making them stable.
[00:11:58.893]Yeah, so smaller computer systems
[00:12:00.970]capable of making them stable.
[00:12:02.470]And I'll say a lot of these things,
[00:12:04.120]so how many of you have one of these
[00:12:05.780]in your pocket or your bag, right?
[00:12:07.660]Okay, or how many do not, anybody?
[00:12:11.460]Okay, so really this is the thing
[00:12:15.620]that made this transition possible.
[00:12:19.330]And part of it was batteries, sure,
[00:12:20.805]we're wanting to pack more batteries into these things,
[00:12:24.150]but the other really critical aspect of this
[00:12:26.330]was the IMU in here, this inertial measurement unit,
[00:12:30.710]and what that is is there's the gyroscope
[00:12:34.500]and the compass and the accelerometer
[00:12:37.690]that knows when I'm holding it this way,
[00:12:40.600]it's holding this way and when you rotate it,
[00:12:42.500]the screen rotates nicely for you.
[00:12:44.590]And when you wanna play your augmented reality game,
[00:12:48.350]it knows exactly how you're tracking in the environment.
[00:12:51.300]That's this IMU, this inertial measurement unit,
[00:12:54.620]and because of consumer electronics,
[00:12:57.750]and in particular phones, it went from things
[00:13:00.340]that were these million-dollar systems,
[00:13:04.840]so actually these IMUs, they were used on submarines.
[00:13:08.510]They were the size of a room
[00:13:10.240]and they were millions of dollars
[00:13:11.840]and they allowed the sub to know exactly where it was
[00:13:14.740]under water better than it could otherwise.
[00:13:18.380]But because of consumer electronics
[00:13:20.180]and in particular phones, these IMUs got shrunk to a point
[00:13:23.240]where you could put it on an actual vehicle.
[00:13:27.130]And why that's important is because,
[00:13:29.460]so we've known how to fly these types of aircraft,
[00:13:32.250]traditional aircraft for a long time
[00:13:33.617]'cause you have your ailerons and your rudder
[00:13:38.160]and your flaps and all that, and they're pretty stable.
[00:13:41.320]So if you've flown a plane, at least a kinda classic plane,
[00:13:46.890]not maybe whatever, the seven,
[00:13:49.439]whatever Boeing Max planes would crash by themselves,
[00:13:53.670]but if you fly an older plane,
[00:13:55.800]if you let your hands off the controls,
[00:13:57.440]it'll still stay in the air and fly.
[00:14:00.280]That's not true of these little guys because in part,
[00:14:04.000]and I have a picture here but also a vehicle,
[00:14:09.530]the thing is is that in order to have this move,
[00:14:12.360]what happens is that to be able to move sideways,
[00:14:19.100]what happens is that if we wanna go this way to your left,
[00:14:23.530]these two propellors have to actually slow down a little bit
[00:14:26.920]so that the whole vehicle tips.
[00:14:28.800]These have to speed up a little bit
[00:14:30.650]so that it doesn't fall out of the sky
[00:14:32.670]and then it gets tipped
[00:14:33.600]and you can think of some physics diagrams
[00:14:36.050]and then it'll slide over to the side, and same thing.
[00:14:39.550]The thing is is that it actually has to actively control
[00:14:42.760]these very fast because if there's some disturbance
[00:14:46.960]and this propellor slows down a little bit,
[00:14:48.700]it's gonna move this way.
[00:14:50.810]So the flight controllers on these things
[00:14:53.220]in this little box here are doing this
[00:14:56.830]hundreds or thousands of times a second,
[00:14:59.080]measuring exactly its orientation in space,
[00:15:02.460]its attitude, it's called,
[00:15:04.180]and adjusting the speeds of the motors just as fast
[00:15:07.960]so that it can stably hover here.
[00:15:11.350]And that's really the biggest reason why we've been able
[00:15:14.900]to go from these really expensive military-type vehicles
[00:15:20.400]to these consumer electronics.
[00:15:24.490]Now, personally, there are a lot
[00:15:27.660]of good military applications of these
[00:15:30.990]but I personally am motivated by all the other things
[00:15:35.610]that these types of systems can do.
[00:15:37.530]So we can do a lot more than just waging a war
[00:15:41.190]and much more than just taking pictures.
[00:15:45.805]And this is really the focus of my work,
[00:15:48.420]is how can you actually make these types of systems interact
[00:15:51.410]with the environment and really solve problems
[00:15:56.030]in some other scientific domains
[00:15:59.120]by solving problems in the robotics domain.
[00:16:02.050]And I'm proud to say that now,
[00:16:06.690]I do air, earth, water, and fire, so I cover the elements.
[00:16:14.040]And this is not just me, this is that group of people
[00:16:16.690]sitting in the middle there
[00:16:17.810]who actually do all the real work,
[00:16:19.930]the graduate students and undergrads here in the NIMBUS Lab.
[00:16:23.960]So we do work on actually, so, my remote is working.
[00:16:34.920]So we have a project on actually digging holes
[00:16:38.270]in the ground, so actually flying out
[00:16:41.440]and digging a hole in the ground and putting a sensor there
[00:16:46.328]and I'll talk a little more about that.
[00:16:47.720]We have, and that's, Adam is sitting there,
[00:16:51.810]he works on that, you can ask him all the questions later.
[00:16:54.610]Ashraful over there built this crazy thing
[00:16:59.100]which we lovingly call the trumpets on the drone.
[00:17:03.750]The reason for this is actually,
[00:17:04.950]we are working with atmospheric scientists
[00:17:07.700]who want to have very precise measurements
[00:17:10.240]of all the atmosphere parameters
[00:17:12.550]as you're going up and down in the atmosphere.
[00:17:15.230]The challenge is that you have these big fans
[00:17:18.580]on this vehicle so if you put sensors below them
[00:17:21.790]or on top of them or just right on the side,
[00:17:24.630]the downdrafts from it actually interfere with the sensing.
[00:17:27.760]So this takes air from outside of the downwash of the rotors
[00:17:33.950]and pulls it past the sensor using the propellors on there.
[00:17:39.050]We also have work on water, so we actually can fly out
[00:17:45.600]and collect water samples from afar
[00:17:50.190]and we have work on actually starting fires.
[00:17:55.810]And what I wanna do is actually talk a little bit
[00:17:58.660]about a couple of these projects.
[00:18:00.190]So I'm gonna talk about actually why and how we start fires
[00:18:05.550]and kind of talk a little bit about that development process
[00:18:08.660]of how we went from, why would we wanna do this
[00:18:12.600]to actually developing this system and commercializing it
[00:18:17.353]and then I'll also talk a little bit at the end
[00:18:19.190]about why we wanna dig holes in the ground
[00:18:21.690]to put sensors there.
[00:18:25.210]So hopefully you have a little bit of an idea now
[00:18:27.280]of kind of the evolution of drones,
[00:18:30.210]and particularly these military ones.
[00:18:32.950]It's been something that people have worked on
[00:18:35.510]since the start of aviation, but again,
[00:18:37.530]it's those phones in your pocket that really enabled
[00:18:39.860]kind of the widespread use of these systems.
[00:18:43.890]So now I wanna talk a little bit about fire
[00:18:45.900]and why we might actually care.
[00:18:48.840]So wildfires are not a huge issue here in Nebraska.
[00:18:54.300]They do occur and they have definitely destroyed
[00:18:57.500]lots of property and things here,
[00:19:00.230]but all over the world they're a major problem
[00:19:05.700]and often these are in remote areas
[00:19:07.820]where you wanna basically try to put them out
[00:19:11.160]potentially before they get to the neighborhoods
[00:19:15.625]that we don't wanna burn.
[00:19:19.030]So there's a significant use of aviation.
[00:19:21.110]So there are helicopters, airplanes,
[00:19:23.840]all sorts of things used to do this.
[00:19:26.120]Now, what you'll notice here
[00:19:27.180]is that this one is dropping water.
[00:19:30.960]This one is actually dropping fire.
[00:19:33.160]So why might you wanna actually start fires to fight fires?
[00:19:41.145]If it's an area that's already burned,
[00:19:42.274]it can't burn more.
[00:19:43.940]Yeah, so if you, so the person said basically
[00:19:47.787]you wanna burn an area in advance of the fire
[00:19:50.280]because then it can't burn anymore.
[00:19:53.020]So it turns out that for firefighters,
[00:19:55.260]like the firefighter you probably think of with a hose
[00:19:59.350]trying to put out house fires and things,
[00:20:01.340]for these types of wildfires,
[00:20:02.890]if the flame is taller than your head,
[00:20:05.580]it's almost impossible to put out by using water alone.
[00:20:09.400]Now, you can use chemical retardants and things,
[00:20:11.260]which they do, but if you can remove the fuel
[00:20:15.110]in advance of the fire,
[00:20:16.400]that's a really good way to put it out
[00:20:18.960]and that's called prescribed fire or controlled burn.
[00:20:22.240]And this is not just used to actually help put out fires
[00:20:27.140]but it's used actually as a land management technique.
[00:20:30.180]So in Nebraska, we're traditionally a prairie state
[00:20:37.290]with no, I think we're the only state
[00:20:39.000]that didn't have a national forest until we planted one.
[00:20:47.420]And the reason for that is largely because fires came
[00:20:49.950]through the plains regularly.
[00:20:51.160]Every five to 10 years there'd be a fire
[00:20:54.140]that would burn through and a lot of the plants
[00:20:57.050]and animals actually adapted
[00:20:59.810]to having fire as part of their life cycle.
[00:21:03.020]So what we have now that we've kind of stopped using fire
[00:21:06.360]is that there are all these cedar trees
[00:21:08.380]that love to grow in Nebraska because great soils,
[00:21:14.260]perfect, lots of sun, but it turns out
[00:21:18.210]that one of the better ways to manage them
[00:21:19.890]is actually by having regular fires
[00:21:21.430]to actually kill the trees when they're young.
[00:21:24.450]So there's a lot of use of prescribed fires
[00:21:26.590]to actually help restore the habitats that we have.
[00:21:32.410]Also this is something where you can use fire,
[00:21:35.940]for instance, in Colorado or California
[00:21:38.270]where there are major wildfires,
[00:21:40.537]if you set a small fire during the wet season,
[00:21:43.950]that's a way you can reduce the fuel so that when it's dry
[00:21:47.860]and a fire starts, it'll be less severe
[00:21:50.530]because a lot of that fuel's been burned out in advance.
[00:21:53.250]And just to give you a little bit of perspective
[00:21:55.520]on kinda the cost of these,
[00:21:58.950]the federal government alone spends over $3.7 billion
[00:22:04.610]on direct costs on actually fighting fires,
[00:22:07.827]these types of wildfires.
[00:22:10.400]The overall cost is many, many times that
[00:22:14.450]and of course there are billions of dollars
[00:22:16.710]of economic damage associated with these as well.
[00:22:19.600]And roughly 20% of the costs
[00:22:22.360]with these direct costs are aviation-related.
[00:22:25.860]And it's also about 1/4 of the deaths
[00:22:29.020]in wildfire fighting are related to aviation,
[00:22:35.594]mainly because there are aircraft that crash
[00:22:40.110]because they're flying really slow and low over these fires
[00:22:42.480]to try to put them out.
[00:22:45.770]So this is one of the goals of this aerial ignition work
[00:22:50.670]that we're doing, is trying to actually move this person
[00:22:55.590]who's in the middle of these hills here with a,
[00:23:00.370]it's hard to see in this picture
[00:23:01.330]but he's driving an ATV with one hand
[00:23:04.240]while holding a drip torch in the other hand
[00:23:06.680]to start this prescribed fire
[00:23:08.580]with fire completely surrounding this person.
[00:23:10.810]So, okay, some people think this is fun.
[00:23:14.970]I don't, but we wanna try to actually get that person
[00:23:20.340]out of harm's way and also give them increased capabilities
[00:23:24.320]and the way that we actually came up with this idea
[00:23:28.580]to do aerial ignition with drones
[00:23:31.090]is actually totally different.
[00:23:32.960]So we were actually working on water sampling
[00:23:36.510]and I think this is one of the reasons
[00:23:38.940]why I like working with a really wide range of people
[00:23:42.200]because we were working on this project
[00:23:43.400]where this guy was tossing out these nets
[00:23:46.220]to look for the veligers, the little baby zebra mussels,
[00:23:49.660]to see how the infestation of zebra mussels
[00:23:53.400]in this lake in Nebraska was spreading
[00:23:55.870]and whether they could manage it.
[00:23:58.510]And this was not at all staged,
[00:24:00.510]but it's actually a challenging job
[00:24:03.831]and this is where we develop this,
[00:24:05.690]we were using this water sampling system on a drone
[00:24:08.970]that could basically precisely control its own altitude
[00:24:14.710]and precision and pump water up through a little filter
[00:24:19.560]to be able to also look for those veligers.
[00:24:23.190]So this had nothing to do with fire
[00:24:25.320]except that we happened to be working on this project,
[00:24:28.340]myself and Sebastian as the robotics people.
[00:24:31.763]We were working with Craig Allen
[00:24:34.220]in the School of Natural Resources
[00:24:36.410]and he said, "Hey, we just hired this guy, Dirac,
[00:24:41.107]"and he's doing all this cool stuff with fire ecology.
[00:24:45.467]"We should have lunch
[00:24:48.010]"because maybe there's some ways to collaborate."
[00:24:49.410]So just from working on this one project,
[00:24:52.730]we connected with somebody else, had lunch,
[00:24:54.380]started talking about, okay, what's hard,
[00:24:56.640]what's interesting, and eventually they both were saying,
[00:25:02.027]"Well, it's really hard to start these fires.
[00:25:03.937]"Can we use a drone to do it?"
[00:25:05.167]And our first response was like, "No way,"
[00:25:07.170]because at the time, the FAA was saying
[00:25:10.910]you're not supposed to actually fly these things
[00:25:13.650]unless you have lots of permissions
[00:25:15.940]and we didn't wanna deal with it,
[00:25:16.980]but we kept running into each other
[00:25:19.700]and we were convinced to actually try this.
[00:25:22.420]So what we were able to do was actually quickly develop
[00:25:26.440]a prototype in just a week of work or less
[00:25:31.010]that uses little ignition spheres,
[00:25:32.520]so they're like ping-pong-sized balls
[00:25:35.160]and they're used by these helicopter systems
[00:25:38.420]to actually start these prescribed fires.
[00:25:40.760]So we used those to try this out
[00:25:44.080]and we were able to basically validate the concept
[00:25:46.670]that we could start these fires.
[00:25:49.590]We could do it with some control.
[00:25:53.460]One of the critical things is that we realized
[00:25:55.570]there were actually problems in robotics
[00:25:57.960]that we had to address to do this.
[00:25:59.570]So how can you actually do this safely and reliably
[00:26:03.200]when the person is not there to fix things if it goes wrong?
[00:26:09.360]So we decided that we actually wanted to work more on this
[00:26:13.137]and we put together a team of computer science,
[00:26:16.350]mechanical engineers and electrical engineers,
[00:26:18.720]in addition to the fire ecologists
[00:26:23.996]and other people on the team,
[00:26:26.220]and worked for another nine months
[00:26:30.240]actually evolving different systems
[00:26:33.455]to the point where we could actually try to do this indoors.
[00:26:38.050]So anybody know where this is?
[00:26:41.760]Anybody ever been in here?
[00:26:44.392]Okay, AJ who flies there, yeah.
[00:26:47.240]Okay, besides the NIMBUS Lab people, yeah,
[00:26:48.840]so this is on East Campus, the horse arena there,
[00:26:52.390]Warren Arena, so we used that to actually test it out
[00:26:55.381]and show how we could actually start these fires
[00:26:58.880]that are using a drone
[00:27:02.870]and trying to actually replace helicopters like that
[00:27:08.190]or at least augment them
[00:27:10.470]because you have to be a certain kind of person
[00:27:13.850]to wanna fly a helicopter
[00:27:15.390]with fire hanging below you over fire.
[00:27:20.060]So we did all this work.
[00:27:21.220]We worked on doing this indoors,
[00:27:23.620]trying to figure out how we could coordinate it,
[00:27:26.260]work on a lot of the robotics problems as well,
[00:27:29.515]and then we wanted to do this outdoors.
[00:27:33.660]It turns out that it took us nine months
[00:27:36.550]of research to get to that point.
[00:27:38.510]It took us another probably five months
[00:27:40.170]to get through all the permissions
[00:27:41.640]to actually be able to do this outdoors.
[00:27:43.750]So UNL, we had to convince them that this was a good idea.
[00:27:49.580]That was relatively easy but then we had to convince
[00:27:52.340]the FAA that we should do this.
[00:27:55.140]It turns out that the FAA said, "Well,
[00:27:57.107]"we don't know what to do about this,
[00:27:58.157]"but maybe it's a weapon so we should talk
[00:27:59.837]"to the Bureau of Alcohol, Tobacco, and Firearms
[00:28:03.107]"and make sure that they're okay with this."
[00:28:07.120]And then we identified a place
[00:28:09.966]at Homestead National Monument just south of here
[00:28:12.700]and it's a national park site
[00:28:14.060]and Department of Interior had to get involved,
[00:28:15.930]and I'm very proud to say that this aircraft
[00:28:18.950]was approved by NASA as airworthy.
[00:28:22.310]So I've not gone to space yet or sent anything to space
[00:28:24.690]but NASA has approved some of our systems,
[00:28:27.770]'cause that's actually who the Department of Interior uses
[00:28:30.770]to certify aircraft as airworthy.
[00:28:34.030]So I have a big stack of papers with permissions and stamps
[00:28:37.580]but what that let us do is actually test this system
[00:28:40.350]in the wild in this prescribed burn environment.
[00:28:43.920]So this is a relatively small burn
[00:28:47.850]and here you can see the vehicle and it's a little hex-rotor
[00:28:55.860]which is not a whole lot bigger than this thing,
[00:28:59.990]dropping these ignition spheres
[00:29:01.750]and here you can see some of the motion there
[00:29:04.650]where it's puncturing these balls and then dropping them.
[00:29:08.830]And there are chemicals in here,
[00:29:10.270]you have to inject another chemical
[00:29:11.720]to get it to start a chemical reaction to ignite,
[00:29:14.800]and here you can see, going back over a little bit later,
[00:29:18.790]the fire that was started with this.
[00:29:21.410]So this was really one of the first outdoor validations
[00:29:24.743]of this system and it was about 20 acres we burned.
[00:29:30.400]Here's a picture showing kind of where we dropped
[00:29:33.440]these ignition spheres and we learned a lot,
[00:29:36.430]as we do every time we go out
[00:29:38.470]into the field and try new things.
[00:29:41.160]And that led to actually a couple more evolutions of systems
[00:29:46.870]into this one which you can't see too well,
[00:29:49.730]but it went from one that's the size of this little one
[00:29:52.310]I was holding up that could hold about a dozen
[00:29:55.080]of these little ignition spheres to this one
[00:29:58.360]that Ashraful who's sitting there helped design
[00:30:00.662]that could hold hundreds of these ignition spheres.
[00:30:04.370]And this is a much larger vehicle
[00:30:07.480]that can fly for a lot longer.
[00:30:10.920]And we also, at the same time
[00:30:12.900]we're looking at how we can make this more autonomous
[00:30:16.380]and in a way to actually both improve the capability
[00:30:21.260]and make it easier to operate.
[00:30:23.150]So this is where, again, this was a lot of fun
[00:30:28.020]to actually learn from old textbooks
[00:30:31.280]how you should do these burns under different conditions.
[00:30:34.180]So we actually read about a lot of these kinda fire models
[00:30:38.500]and how fires spread and there are basically anything
[00:30:42.070]you can think of will influence how fire moves,
[00:30:44.350]which is why it's so hard to model.
[00:30:45.730]So winds, moisture, the fuel type,
[00:30:50.320]when the last time it rained,
[00:30:53.420]all of these factors influence it.
[00:30:55.660]What we were able to do is actually incorporate that
[00:30:57.890]into an automated planning system
[00:31:01.160]that could take these types of, you know, with these,
[00:31:04.827]you can say, "Okay, I just wanna burn this area,"
[00:31:07.067]and it can automatically plan a good ignition
[00:31:10.680]given the parameters that you give it.
[00:31:15.040]It also lets you do things while you're actually flying.
[00:31:21.830]If you go up high, you can get those great aerial images
[00:31:25.450]to see, okay, how well does what we're trying to do match
[00:31:29.410]what we were doing
[00:31:32.600]and we had the capability so that you could just correct
[00:31:35.570]this fire model as it went and it would improve over time
[00:31:38.480]so that you could just continually improve the performance
[00:31:42.500]so that you have better situational awareness
[00:31:44.460]of what's happening in the whole area.
[00:31:46.970]Also things like being able to automatically re-plan.
[00:31:50.670]So if you initially thought
[00:31:51.780]you were just doing this simple fire
[00:31:53.547]but then it turns out that this was a really flammable area
[00:31:57.430]and burned faster here,
[00:31:59.710]the system could automatically actually re-plan
[00:32:02.280]so that you would both stay safe
[00:32:05.090]and save time on your fire planning.
[00:32:10.260]So this led to a burn we did
[00:32:15.440]in Nebraska Indian Caves.
[00:32:19.810]And if you haven't noticed in the spring
[00:32:26.600]when there's lots of smoke on the horizon,
[00:32:28.700]we normally notice it because Kansas is burning
[00:32:32.140]and we get all the smoke.
[00:32:34.040]But you'll see that there's actually a lot of use
[00:32:35.510]of prescribed fire in the state
[00:32:37.670]for this type of land management.
[00:32:40.370]So this is a state park where they actually,
[00:32:43.390]and sorry, this is not the best projector here
[00:32:46.980]in terms of lumens or whatever,
[00:32:49.770]but we tested this at a much larger scale at the state park
[00:32:54.440]where they do these burns regularly, again,
[00:32:57.520]for kind of maintaining the natural habitat in this area.
[00:33:06.605]And you can see just how even these little ignition spheres
[00:33:10.140]can start these fires, and in this case,
[00:33:12.970]they're not trying to burn all the trees down.
[00:33:15.060]They're just trying to burn kind of the undergrowth
[00:33:18.840]that builds up over time.
[00:33:22.470]And so there was a really large area
[00:33:25.490]that we were able to start this prescribed fire on
[00:33:29.450]and there you can see the kind of before and after,
[00:33:31.397]and if you look really closely,
[00:33:33.040]you can see the drone flying up there
[00:33:36.408]and we burned a whole lot of this area,
[00:33:38.250]including some areas where they said,
[00:33:42.180]basically they weren't willing to send people to
[00:33:44.030]just because it was steep areas that were too dangerous
[00:33:46.930]to actually put fire on while a person was nearby.
[00:33:50.763](coughs) Excuse me.
[00:33:54.970]So that's kind of how this project has evolved in the lab.
[00:33:59.650]So it started out really with a chance encounter,
[00:34:03.820]so every time you go to some random talk
[00:34:06.170]you should take this opportunity to think about, oh,
[00:34:09.410]what do I know that could maybe apply there
[00:34:11.470]and is there some crazy idea
[00:34:13.420]that could actually move society forward in these ways?
[00:34:18.870]I think crazy ideas are great
[00:34:21.880]and I think I read some study once
[00:34:24.870]that if you're a close-knit group,
[00:34:29.934]so if my student comes to me that I work with every day
[00:34:34.000]with some crazy idea, I'm more likely to reject it
[00:34:36.580]than if some random person walks off the street
[00:34:38.490]and comes up with the same crazy idea.
[00:34:42.740]So I try to be open-minded
[00:34:45.010]when my students come up with crazy ideas.
[00:34:47.740]You can ask them after if I am.
[00:34:51.260]But it's these interactions, I think,
[00:34:53.140]where a single chance encounter meeting, you know,
[00:34:56.580]when we were doing water sampling,
[00:34:57.710]that leads to some of these results.
[00:35:00.260]And one of the things we've now looked at
[00:35:05.530]is actually commercializing this system.
[00:35:07.140]So myself and a couple of students,
[00:35:09.210]we actually started a company now
[00:35:11.560]called Drone Amplified in town where we're actually,
[00:35:14.680]we've commercialized this system
[00:35:16.380]and I like to say in the research lab,
[00:35:18.870]we do 95% of the hard work but then you go to a company
[00:35:23.000]and you realize there's still 95% of the hard work left.
[00:35:28.870]So the company is actually now taking the basis
[00:35:32.150]that we developed in the lab in actually making it
[00:35:34.420]so that it's actually a robust and reliable system
[00:35:36.620]that people can throw in the back of a plane
[00:35:40.970]and fly to a remote location in Alaska
[00:35:43.860]and be in the field for a long time
[00:35:47.330]with this and still have it work.
[00:35:51.230]But it's based on the same concept
[00:35:53.950]of the things we worked on in the lab.
[00:35:56.950]Now, okay, I've highlighted some of the students,
[00:36:00.030]but I think the real thing is that in all these projects,
[00:36:03.790]especially for the students in the room, the faculty there
[00:36:10.087]who are listed up there aren't really important.
[00:36:12.000]It's really the students who were the key drivers of this
[00:36:16.100]including, so Christian and Becca
[00:36:17.933]at the time were both undergrads
[00:36:20.830]who worked in the lab working on this.
[00:36:23.110]So I definitely encourage all of you,
[00:36:26.730]if you aren't involved in research,
[00:36:29.098]find something to do besides take classes.
[00:36:32.710]There's a lot of interesting research going on
[00:36:34.990]across campus related to this.
[00:36:38.190]So these are the people who are really the ones
[00:36:40.730]who did all the hard work.
[00:36:43.340]I just get to talk about it.
[00:36:46.450]So I wanna talk about, I guess, maybe,
[00:36:49.457]I have a little bit more time,
[00:36:50.440]I'll talk about one other project
[00:36:52.470]that we're working on that I alluded to.
[00:36:54.690]So you have this fire, we're starting fires
[00:36:58.780]but we also wanna know what's actually going on
[00:37:00.630]with the fire and all the area around it
[00:37:02.830]so one of the questions is can we actually deploy sensors
[00:37:07.710]around this fire or in other locations
[00:37:09.950]that are hard or difficult to get to?
[00:37:14.602]And the idea here is that we wanna be able
[00:37:18.280]to send out a drone to fly out to a remote location,
[00:37:25.110]dig a hole in the ground, and then put a sensor in there
[00:37:28.670]and then leave so that we can basically put sensors
[00:37:31.210]in remote locations to monitor everything
[00:37:35.140]from how fires are progressing to soil moisture in crops
[00:37:40.735]to security-related threats
[00:37:44.975]where you wanna covertly place sensors
[00:37:48.080]and I'm gonna skip all of the development,
[00:37:52.220]but this, again, is a couple of years of development
[00:37:55.880]and you can talk to Adam there as well about all this.
[00:37:59.340]But what we were able to do is actually develop this system
[00:38:04.440]where we can use a much bigger helicopter
[00:38:07.550]to carry our drone which is inside this canister here
[00:38:11.630]which is about yay high and actually carry this drone out
[00:38:17.440]into a really remote location
[00:38:19.320]because these things still have pretty limited battery life.
[00:38:23.590]So this is some experiments we did last year
[00:38:28.130]where we're flying out and then when we get to the location,
[00:38:32.560]the mother ship drops our system off.
[00:38:36.480]We get rid of the cover and then the arms
[00:38:39.270]on the vehicle spring out and we can release ourselves
[00:38:44.560]and fly off to wherever we're going.
[00:38:51.010]And when we get there, we can basically figure out
[00:38:55.220]where it's safe to land and dig a hole
[00:38:58.390]in the ground to put the sensor in.
[00:39:01.100]And this is, a whole lot of engineering goes into this
[00:39:06.650]but also a whole lot of soil science.
[00:39:10.590]So this is something where we learned a lot
[00:39:12.320]about how to dig holes in the ground in an effective way
[00:39:17.720]and then it can leave that whole kind of auger system around
[00:39:21.470]and that's a sensor that's embedded
[00:39:23.530]in that kind of big drill bit
[00:39:25.680]that can then be out there for monitoring various parameters
[00:39:30.630]and for long periods of time.
[00:39:35.270]So I just wanted to say that
[00:39:36.173]because in part it's really cool, I think,
[00:39:38.780]and a lot of effort went into it.
[00:39:41.310]The video makes it really simple-sounding
[00:39:43.820]but a lot of actual challenges
[00:39:46.820]in robotics that we had to address.
[00:39:49.190]So I think I'll, I guess, wrap up now
[00:39:53.990]and I think the one message that I'd like to get across,
[00:39:59.268]if you go away with anything,
[00:40:00.810]is that drones, flying robots, UAVs, UASs,
[00:40:05.680]whatever you wanna call them,
[00:40:07.040]have a lot of potential to do more
[00:40:09.640]than just doing this military applications
[00:40:14.044]or just flying around and taking cool pictures of people
[00:40:19.170]at the Huskers football game or something like that.
[00:40:22.000]They really, I think, will become a much more common tool
[00:40:26.130]in interacting with the world and doing things
[00:40:29.440]and you see some of this,
[00:40:31.943]there's the whole package delivery thing.
[00:40:35.290]We can talk about that.
[00:40:36.390]I'm not too much of a believer in that,
[00:40:39.710]like, swarms of these things.
[00:40:41.200]There are even companies trying
[00:40:42.310]to actually have our personal transport,
[00:40:44.600]so instead of calling a Uber car,
[00:40:47.890]it's the Uber Lift or whatever that flies you around.
[00:40:52.994]I had to put the Terminator picture.
[00:40:54.350]I don't think we're going there.
[00:40:56.674]And we can talk about all sorts of reasons
[00:40:58.600]why I don't think that robotics will be advanced enough,
[00:41:02.530]at least within our lifetimes or our children's lifetimes.
[00:41:07.890]But really I think this is one of these things
[00:41:12.210]where there's a lot of potential.
[00:41:14.870]I love this area because there's still just tons
[00:41:17.570]of challenges in the robotics domain
[00:41:20.150]to actually make these work, be safer,
[00:41:24.180]be more reliable, how can you do this
[00:41:27.230]without having to be a foot away from it flying
[00:41:30.950]when you're operating far away.
[00:41:34.430]And I guess one of the reasons why I don't think
[00:41:38.370]these terminators will be coming to get us
[00:41:40.160]with drones any time soon is
[00:41:41.777]'cause there's a whole lot of failures involved,
[00:41:44.210]and so behind these nice videos,
[00:41:48.020]there's a lot of work that goes wrong.
[00:41:53.780]So this is an early test where we were testing
[00:41:58.184]dropping the drone from another larger one
[00:42:01.350]and it did not go well.
[00:42:06.980]So, yeah.
[00:42:16.076]Failures are part of robotics.
[00:42:18.770]Here, a motor failed to start
[00:42:20.710]and actually the software thought the motor had started
[00:42:23.470]so our software tried to take off
[00:42:27.120]and this just tells me we need not only better hardware
[00:42:30.510]but better software to make these work better.
[00:42:35.170]Fortunately, okay, we did not have the string on here
[00:42:38.040]because we thought we would go fishing.
[00:42:41.130]At the time, the FAA would not allow us
[00:42:44.360]to fly in this location,
[00:42:47.350]so we put a tether on it so it wasn't really flying.
[00:42:51.010]So it was easy to pull it out
[00:42:53.190]and spend, I don't know,
[00:42:57.970]my students spent 10 hours putting this in bags of rice
[00:43:02.650]like you do with your phone when you drop it in a lake.
[00:43:07.810]He's now actually a faculty member
[00:43:09.470]so I guess it didn't hurt him too much
[00:43:11.560]doing all the bags of rice.
[00:43:14.410]But yeah, so, robotics,
[00:43:16.090]there's a lot of really hard problems,
[00:43:18.010]a lot of opportunities, and yeah.
[00:43:22.450]I think this is all part of the learning process.
[00:43:25.800]Definitely wanna thank all the many people involved,
[00:43:29.670]especially the students, and then I have one last question
[00:43:32.760]before you can ask me questions, which is who is who?
[00:43:40.006]So there's some actual firefighters
[00:43:41.520]and then there are the computer scientists
[00:43:43.400]and robotics people and all the other people.
[00:43:46.990]So can you tell with this?
[00:43:50.070]It's not just the screen.
[00:43:53.700]Yeah, these are all the firefighters, right,
[00:43:55.040]they're all already covered in black.
[00:43:57.090]We were just in the shiny yellow ones.
[00:44:02.330]And with that, I'd be happy
[00:44:04.520]to answer any questions you may have.
[00:44:08.027]Thanks, yeah?
[00:44:10.288]How many patents do you own altogether?
[00:44:13.414]How many are on the actual drones
[00:44:14.957]and how many are on the softwares you use in the drones?
[00:44:17.506]How many?
[00:44:18.339]Patents.
[00:44:19.172]Okay, so the question is how many patents do we have.
[00:44:22.020]I think we have, and then are they on the drones themselves
[00:44:27.690]or the software or other things.
[00:44:29.173]Oh no, how many are actually
[00:44:30.443]on the, when you build a drone
[00:44:32.065]and how many are on the technology to power those drones?
[00:44:35.093]Yeah, so we have a few patents that we filed.
[00:44:38.780]Patents take forever.
[00:44:42.048]Most of them are actually on the actual payload system
[00:44:46.692]itself and the, I mean, you can't patent software
[00:44:50.810]but the idea behind the algorithms that control them.
[00:44:54.660]So actually, minus maybe this one
[00:45:00.140]which is one of the most modified drones,
[00:45:02.420]all of these are off-the-shelf vehicles, and so, yeah.
[00:45:08.180]We're not in the business of trying to build drones.
[00:45:12.880]There are certainly people who are researching
[00:45:15.020]the aerodynamics of propellors and batteries and materials.
[00:45:19.700]Yeah, our focus is really on creating new systems
[00:45:25.370]to do more things with them,
[00:45:28.050]and that's normally kind of on the payload side.
[00:45:29.909]How many patents do you have
[00:45:31.037]on the payload?
[00:45:32.448]So the question is how many pounds of, what?
[00:45:35.800]Patents.
[00:45:36.684]Patents?
[00:45:37.517]On the payload,
[00:45:38.350]the delivery system, how many patents it is, and then--
[00:45:40.930]Oh, just a few, I don't know, it's not,
[00:45:45.474]I guess I have a startup so I should care more about this
[00:45:47.900]but I also don't care.
[00:45:50.250]I prefer to publish things and do new things,
[00:45:53.210]but yeah, so was there another question, yeah?
[00:45:57.896]So with all those firefighters,
[00:45:59.005]are you usually gonna have the individual,
[00:46:01.353]I think you called it a drip line--
[00:46:02.234]Drip torch.
[00:46:03.067]Drip torch. Yeah.
[00:46:04.275]What's the difference in time
[00:46:05.816]between one of those commercial drones
[00:46:07.645]versus a crew of firefighters?
[00:46:09.511]Yeah, so that's a good question.
[00:46:10.590]So for those who might not hear it,
[00:46:12.360]the question was how much faster is it
[00:46:14.570]for a drone to go out there and start these prescribed fires
[00:46:18.730]versus somebody with a drip torch,
[00:46:21.370]and the answer is it's really complicated.
[00:46:23.320]So it turns out that when you're doing these burns,
[00:46:28.300]most of the people are actually around the fire
[00:46:32.730]trying to put it out, so these spot fires.
[00:46:35.370]So you'll start it along a road
[00:46:38.470]and so that the wind is blowing this way
[00:46:41.810]and maybe the road is here so you start it here
[00:46:43.770]and have it try to go slowly that way,
[00:46:46.970]but the wind might blow an ember across the road.
[00:46:49.810]So for every one person with a drip torch here,
[00:46:53.210]there are, like, 10 people over here
[00:46:54.700]trying to find where the fire jumped to.
[00:46:58.760]But as I said, for these remote locations
[00:47:02.160]where you're not trying to start the boundary of the fire,
[00:47:06.640]where you're trying to start the interior ignition,
[00:47:08.250]it's called, that, it's really,
[00:47:12.780]sometimes you just simply can't get there.
[00:47:14.470]So we're doing work with people in Florida.
[00:47:19.080]So it turns out that they do a lot of prescribed burns
[00:47:22.980]of the Everglades to maintain that habitat
[00:47:26.260]and it's basically marshy, so you can't actually get there
[00:47:31.660]with a drip torch or you have to take a boat
[00:47:35.190]and then do little land masses on its own.
[00:47:38.240]So there, this is a great time-saving,
[00:47:43.116]and they use helicopter-based systems
[00:47:46.010]right now to do that, yeah, yeah?
[00:47:49.916]How long does do the sensors work
[00:47:52.017]and do you have to get them removed?
[00:47:54.690]Yeah, okay, so the question is
[00:47:55.640]how long do the sensors work and do you have to remove them.
[00:48:00.450]So it depends a lot on the application.
[00:48:02.820]So that thing actually has no sensor in it
[00:48:07.210]because we were just demonstrating the concept.
[00:48:12.130]A lot of these sensors can live for years on end
[00:48:16.800]if they're low-powered sensors and transmitters.
[00:48:21.480]So you can do a lot just deploying these.
[00:48:25.740]We actually do have a new NSF project
[00:48:28.950]where we're looking at deploying various types of sensors
[00:48:32.070]in rainforests and they're definitely,
[00:48:35.670]you'll want to remove them at some point
[00:48:37.140]and that's one of the things we're thinking about is,
[00:48:38.750]okay, after you put it out there can you go back
[00:48:41.720]in three years and find where it was
[00:48:44.370]and figure out how to remove it?
[00:48:45.667]'Cause you don't wanna just litter everywhere.
[00:48:51.392]Yeah?
[00:48:52.607]So I notice still a lot of concern
[00:48:54.306]about who's doing what with the drones.
[00:48:57.245]Do you have any regulations when you got them out
[00:48:59.285](speaks faintly) to who you sell them to?
[00:49:02.170]Yeah, okay, so the question is around regulations
[00:49:04.440]and are there regulations around who we sell to and things.
[00:49:11.170]So there are a lot of regulations.
[00:49:15.190]So it took six months to be able to demonstrate
[00:49:18.940]this prescribed fire work with the university.
[00:49:23.460]It took another nine months to get,
[00:49:25.640]and this is many years later
[00:49:27.310]after we did it at the university,
[00:49:28.670]it took nine months to, as a company,
[00:49:31.360]get permission to actually do these burns.
[00:49:35.240]So from a regulatory standpoint, you do have to, you know,
[00:49:39.400]if you wanna follow the rules, you have to do that.
[00:49:42.940]There aren't really rules on who to sell this to,
[00:49:45.190]although I guess the common question is, well,
[00:49:48.060]if terrorists get a hold of this, what are we gonna do?
[00:49:51.010]I think, I mean, maybe,
[00:49:54.190]I can come up with many, many other worse ways
[00:49:56.690]to use a drone to cause damage.
[00:49:59.280]This is actually relatively safe.
[00:50:01.080]You can jump on these balls all you want
[00:50:03.730]and they won't do anything too bad.
[00:50:08.435]There's another question, yeah.
[00:50:12.012]It's very fascinating
[00:50:12.847]to see all those examples.
[00:50:14.761]A lot of what I've seen is beyond the sort of sensor
[00:50:17.458]that's placed in the ground,
[00:50:18.687]communication is sort of one-way
[00:50:21.146]in terms of using the sensors.
[00:50:22.389]It's beyond sort of flying the drone.
[00:50:26.066]Are you using it for remote sensing in real time,
[00:50:28.552]collecting the remotely-sensed data or topics like that?
[00:50:32.670]Yeah, so basically real-time data collection type work.
[00:50:36.750]Yeah, so often this is our science partners
[00:50:42.010]who are doing some of this.
[00:50:43.260]So for the atmospheric sensing, for instance,
[00:50:48.050]right now we're just collecting the data
[00:50:49.850]but they're certainly interested
[00:50:50.950]in if we can get this in real time
[00:50:53.840]and try to adjust where you're sensing.
[00:50:57.740]So maybe if you were trying to find some inversion layer
[00:51:02.390]in the atmosphere, try to actually track that
[00:51:04.730]over time as it's moving.
[00:51:07.690]There's a lot of, yeah,
[00:51:08.560]definitely a lot of interesting potential in this area.
[00:51:13.800]Justin Bradley, one of the other directors in the lab,
[00:51:16.450]he also does a lot of work on multi-vehicle coordination
[00:51:20.760]and that requires real-time sensing and feedback as well.
[00:51:27.910]Any other questions, yeah?
[00:51:29.795]What should a student do
[00:51:31.558]if they wanna get involved with drone research?
[00:51:34.459]What should a student do if they wanna get involved
[00:51:36.040]in drone research?
[00:51:38.260]So, I think all sorts of things you can do.
[00:51:40.160]So there are some robotics clubs on campus.
[00:51:44.810]There's also, I think there's a drone club as well.
[00:51:49.810]You're welcome, so in the NIMBUS Lab we have weekly meetings
[00:51:53.220]that anybody can come to
[00:51:54.497]and just get to know what we're doing.
[00:51:57.890]Certainly, I think one of the things about robotics
[00:52:01.540]is you can come at it from all sorts of directions.
[00:52:04.410]So there's kind of maybe three core disciplines,
[00:52:08.610]computer science, mechanical engineering,
[00:52:10.710]electrical engineering, where it's actually be building them
[00:52:15.980]but then also sociology, there are sociologists
[00:52:19.680]who are studying the bigger questions associated with this.
[00:52:24.610]We do work with somebody in psychology
[00:52:29.250]looking at how people understand
[00:52:32.582]and think about these different ideas.
[00:52:35.740]So we did a study where we were asking,
[00:52:39.070]okay, is this atmospheric sensing drone better
[00:52:44.080]versus somebody trying to take pictures for the news.
[00:52:53.457]And we found some strong correlation
[00:52:55.710]with the application area,
[00:52:56.760]so if people knew, oh, we're doing this
[00:53:00.670]so that we can better predict the next tornado
[00:53:04.020]or thunderstorm, they're like, "Oh, yeah,
[00:53:05.857]"we should definitely do that,"
[00:53:08.000]versus, well, I just want to take a picture
[00:53:12.170]of what that farmer's doing over there.
[00:53:14.720]People definitely did not like that.
[00:53:16.930]So I think there's a whole range
[00:53:19.047]of ways you can come into this.
[00:53:22.460]I guess the more direct answer is probably some of the clubs
[00:53:26.424]or connecting with the research labs and stuff.
[00:53:34.610]Other questions, yeah?
[00:53:37.790]So this is just asking
[00:53:40.691]for your thoughts on Guy Reynolds.
[00:53:43.061]He was pretty nervous
[00:53:47.111]about the whole rise of the machines thing
[00:53:49.114]and, or at least he was talking about in literature,
[00:53:51.858]there's been real concerns about the rise of the machine
[00:53:55.485]and how it would destroy human life as we know it.
[00:53:59.247]You know what I mean, destroy our humanity, (speaks faintly)
[00:54:02.915]and you have a very positive view
[00:54:05.370]of the rise of the machine.
[00:54:07.450]I wonder what you, so why are you positive?
[00:54:11.330]Are you negative? (speaks faintly)
[00:54:17.247]Yeah, so, and I did not get to go to his talk,
[00:54:22.060]but I've talked to him a little bit
[00:54:23.300]but I should probably talk more to him about this.
[00:54:28.690]So my feeling is all those failures,
[00:54:32.390]I don't know, even our technology fails, right?
[00:54:38.102]We're just not anywhere close, is my feeling.
[00:54:43.630]So in 2010 when I first came here
[00:54:45.980]and was teaching a grad seminar on robotics
[00:54:50.870]and autonomous cars were starting to be even more of a thing
[00:54:57.064]and I said, "Okay, there's no way in my lifetime
[00:55:00.397]"that these are gonna be driving everywhere."
[00:55:04.868]I a little bit regretted that for a short period of time,
[00:55:07.860]thinking that, okay, I was totally gonna be wrong,
[00:55:09.830]which, I'm frequently wrong.
[00:55:11.180]But I don't think I'm wrong about that
[00:55:13.000]because if you look at the challenges associated with it,
[00:55:21.050]'cause autonomous cars, for instance,
[00:55:22.990]are really good driving in areas
[00:55:25.500]where they have a complete map of the area
[00:55:27.740]because they've driven there so many times
[00:55:29.990]and actually there was just a talk in our department
[00:55:31.550]right before this from somebody
[00:55:34.660]at Carnegie Mellon University
[00:55:36.380]and his whole work was trying to figure out
[00:55:41.150]these strange test cases.
[00:55:43.510]So he was talking about how these cars think
[00:55:47.720]that person dressed up in a chicken suit
[00:55:49.500]is actually a chicken and not a person.
[00:55:53.570]And there are just so many examples of this
[00:55:55.880]where I think that we think our technology is super clever
[00:56:02.180]but I think they're just nowhere near
[00:56:05.010]what a human can still do.
[00:56:06.450]So I think there's always gonna be humans involved,
[00:56:10.430]at least for my lifetime.
[00:56:11.490]I think I'm safe from, I don't know,
[00:56:14.170]being a cyborg or something, but,
[00:56:17.420]or maybe that will be the solution,
[00:56:18.850]is they just implant the computers in our heads
[00:56:20.630]and we don't know the difference.
[00:56:24.609]Yeah?
[00:56:27.372]One particular concern of mine
[00:56:28.804]in seeing these drone plans was that some people
[00:56:34.539]could use that for planting land mines,
[00:56:38.651]and that doesn't seem so far off too, of course--
[00:56:43.210]Yeah, yeah, so, yeah,
[00:56:46.030]so can you use this, yeah, to plant land mines?
[00:56:49.370]You probably could, I mean, I don't know.
[00:56:53.820]Maybe this is how I go to bed at night better,
[00:56:56.210]is by thinking that I can also think about 20 other ways
[00:57:01.570]that you could put land mines out there more effectively
[00:57:05.270]and more covertly and at lower cost.
[00:57:07.007]You can just drop 'em, and we do have land mines,
[00:57:12.140]or we've unfortunately shown that you can deploy land mines
[00:57:16.050]very effectively in a lot of ways.
[00:57:19.003]And I think this is something with technology
[00:57:21.450]where you always have to, I think it's important
[00:57:23.890]that we ask ourselves the question of what does this mean?
[00:57:28.880]What are the good uses for this technology?
[00:57:30.680]What are the bad uses?
[00:57:31.590]And I think this is definitely something that's important
[00:57:34.920]to keep in mind and not just develop technology
[00:57:39.030]for technology's sake, but yeah,
[00:57:44.180]and yeah, partly I can think of many bad things you can do
[00:57:48.100]with these right now that are, yeah,
[00:57:53.140]that I hope never happen.
[00:57:55.561]Yeah?
[00:57:57.138]I build drones, I build racing drones,
[00:58:00.178]and I go on the web, and not surprisingly,
[00:58:03.740]there's somebody in there trying to,
[00:58:06.317]70 kilograms of (speaks faintly) on top of this
[00:58:10.427]to either go over and drop something,
[00:58:12.760]I'm not worried about it.
[00:58:15.238]So (speaks faintly) what you've talked about
[00:58:18.580]and what is (speaks faintly).
[00:58:20.750]Oh, yeah, no, and yeah, and I think,
[00:58:26.744]I don't know, I think it's the question
[00:58:28.150]of if you're wanting to do something nefarious like that
[00:58:31.470]and you have the resources and capability,
[00:58:35.500]I think you can do a lot of bad things with, well,
[00:58:39.370]things from 20 years ago too.
[00:58:42.100]So I think you have to find the balance between, I think,
[00:58:49.230]slowing development versus, I think,
[00:58:52.970]educating people in terms of risks
[00:58:56.790]and things as well is important.
[00:59:00.530]But I like to think people are good. (laughs)
[00:59:09.354]Okay, well, I'm happy to answer any other questions
[00:59:12.430]but with that, I thank you
[00:59:15.660]for paying attention and being here.
[00:59:17.772](audience applauding)

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

"Shaping the Future Rise of Drones"

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments