TED设计:Vijay Kumar: The future of flying robots

In my lab, we buildautonomous aerial robotslike the one you see flying here.Unlike the commercially available dronesthat you can buy today,this robot doesn't have any GPS on board.So without GPS,it's hard for robots like thisto determine their position.This robot uses onboard sensors,cameras and laser scanners,to scan the environment.It detects features from the environment,and it determines where it isrelative to those features,using a method of triangulation.And then it can assembleall these features into a map,like you see behind me.And this map then allows the robotto understand where the obstacles areand navigate in a collision-free manner.

What I want to show you nextis a set of experimentswe did inside our laboratory,where this robot was ableto go for longer distances.So here you'll see, on the top right,what the robot sees with the camera.And on the main screen —and of course this is sped upby a factor of four —on the main screen you'll seethe map that it's building.So this is a high-resolution mapof the corridor around our laboratory.And in a minuteyou'll see it enter our lab,which is recognizableby the clutter that you see.


But the main point I want to convey to youis that these robots are capableof building high-resolution mapsat five centimeters resolution,allowing somebody who is outside the lab,or outside the buildingto deploy thesewithout actually going inside,and trying to inferwhat happens inside the building.

Now there's one problemwith robots like this.The first problem is it's pretty big.Because it's big, it's heavy.And these robots consumeabout 100 watts per pound.And this makes fora very short mission life.The second problemis that these robots have onboard sensorsthat end up being very expensive —a laser scanner, a cameraand the processors.That drives up the cost of this robot.

So we asked ourselves a question:what consumer productcan you buy in an electronics storethat is inexpensive, that's lightweight,that has sensing onboard and computation?And we invented the flying phone.


So this robot uses a Samsung Galaxysmartphone that you can buy off the shelf,and all you need is an app that youcan download from our app store.And you can see this robotreading the letters, "TED" in this case,looking at the cornersof the "T" and the "E"and then triangulating off of that,flying autonomously.That joystick is just thereto make sure if the robot goes crazy,Giuseppe can kill it.


In addition to buildingthese small robots,we also experiment with aggressivebehaviors, like you see here.So this robot is now travelingat two to three meters per second,pitching and rolling aggressivelyas it changes direction.The main point is we can havesmaller robots that can go fasterand then travel in thesevery unstructured environments.

And in this next video,just like you see this bird, an eagle,gracefully coordinating its wings,its eyes and feetto grab prey out of the water,our robot can go fishing, too.


In this case, this is a Philly cheesesteakhoagie that it's grabbing out of thin air.


So you can see this robotgoing at about three meters per second,which is faster than walking speed,coordinating its arms, its clawsand its flight with split-second timingto achieve this maneuver.In another experiment,I want to show youhow the robot adapts its flightto control its suspended payload,whose length is actually largerthan the width of the window.So in order to accomplish this,it actually has to pitchand adjust the altitudeand swing the payload through.But of course we wantto make these even smaller,and we're inspiredin particular by honeybees.So if you look at honeybees,and this is a slowed down video,they're so small,the inertia is so lightweight —


that they don't care —they bounce off my hand, for example.This is a little robotthat mimics the honeybee behavior.And smaller is better,because along with the small sizeyou get lower inertia.Along with lower inertia —

(Robot buzzing, laughter)

along with lower inertia,you're resistant to collisions.And that makes you more robust.So just like these honeybees,we build small robots.And this particular oneis only 25 grams in weight.It consumes only six watts of power.And it can travelup to six meters per second.So if I normalize that to its size,it's like a Boeing 787 travelingten times the speed of sound.


And I want to show you an example.This is probably the first planned mid-aircollision, at one-twentieth normal speed.These are going at a relative speedof two meters per second,and this illustrates the basic principle.The two-gram carbon fiber cage around itprevents the propellers from entangling,but essentially the collision is absorbedand the robot responds to the collisions.And so small also means safe.In my lab, as we developed these robots,we start off with these big robotsand then now we're downto these small robots.And if you plot a histogramof the number of Band-Aids we've orderedin the past, that sort of tailed off now.Because these robots are really safe.

The small size has some disadvantages,and nature has found a number of waysto compensate for these disadvantages.The basic idea is they aggregateto form large groups, or swarms.So, similarly, in our lab,we try to create artificial robot swarms.And this is quite challengingbecause now you have to thinkabout networks of robots.And within each robot,you have to think about the interplayof sensing, communication, computation —and this network then becomesquite difficult to control and manage.So from nature we take awaythree organizing principlesthat essentially allow usto develop our algorithms.The first idea is that robotsneed to be aware of their neighbors.They need to be able to senseand communicate with their neighbors.

So this video illustrates the basic idea.You have four robots —one of the robots has actually beenhijacked by a human operator, literally.But because the robotsinteract with each other,they sense their neighbors,they essentially follow.And here there's a single personable to lead this network of followers.So again, it's not because all the robotsknow where they're supposed to go.It's because they're just reactingto the positions of their neighbors.


So the next experiment illustratesthe second organizing principle.And this principle has to dowith the principle of anonymity.Here the key idea is thatthe robots are agnosticto the identities of their neighbors.They're asked to form a circular shape,and no matter how many robotsyou introduce into the formation,or how many robots you pull out,each robot is simplyreacting to its neighbor.It's aware of the fact that it needsto form the circular shape,but collaborating with its neighborsit forms the shapewithout central coordination.Now if you put these ideas together,the third idea is that weessentially give these robotsmathematical descriptionsof the shape they need to execute.And these shapes can be varyingas a function of time,and you'll see these robotsstart from a circular formation,change into a rectangular formation,stretch into a straight line,back into an ellipse.And they do this with the samekind of split-second coordinationthat you see in natural swarms, in nature.

So why work with swarms?Let me tell you about two applicationsthat we are very interested in.The first one has to do with agriculture,which is probably the biggest problemthat we're facing worldwide.As you well know,one in every seven personsin this earth is malnourished.Most of the land that we can cultivatehas already been cultivated.And the efficiency of most systemsin the world is improving,but our production systemefficiency is actually declining.And that's mostly because of watershortage, crop diseases, climate changeand a couple of other things.

So what can robots do?Well, we adopt an approach that'scalled Precision Farming in the community.And the basic idea is that we flyaerial robots through orchards,and then we buildprecision models of individual plants.So just like personalized medicine,while you might imagine wantingto treat every patient individually,what we'd like to do is buildmodels of individual plantsand then tell the farmerwhat kind of inputs every plant needs —the inputs in this case being water,fertilizer and pesticide.Here you'll see robotstraveling through an apple orchard,and in a minute you'll seetwo of its companionsdoing the same thing on the left side.And what they're doing is essentiallybuilding a map of the orchard.Within the map is a mapof every plant in this orchard.

(Robot buzzing)

Let's see what those maps look like.In the next video, you'll see the camerasthat are being used on this robot.On the top-left is essentiallya standard color camera.On the left-center is an infrared camera.And on the bottom-leftis a thermal camera.And on the main panel, you're seeinga three-dimensional reconstructionof every tree in the orchardas the sensors fly right past the trees.Armed with information like this,we can do several things.The first and possibly the most importantthing we can do is very simple:count the number of fruits on every tree.By doing this, you tell the farmerhow many fruits she has in every treeand allow her to estimatethe yield in the orchard,optimizing the productionchain downstream.

The second thing we can dois take models of plants, constructthree-dimensional reconstructions,and from that estimate the canopy size,and then correlate the canopy sizeto the amount of leaf area on every plant.And this is called the leaf area index.So if you know this leaf area index,you essentially have a measure of how muchphotosynthesis is possible in every plant,which again tells youhow healthy each plant is.By combining visualand infrared information,we can also compute indices such as NDVI.And in this particular case,you can essentially seethere are some crops that arenot doing as well as other crops.This is easily discernible from imagery,not just visual imagery but combiningboth visual imagery and infrared imagery.

And then lastly,one thing we're interested in doing isdetecting the early onset of chlorosis —and this is an orange tree —which is essentially seenby yellowing of leaves.But robots flying overheadcan easily spot this autonomouslyand then report to the farmerthat he or she has a problemin this section of the orchard.

Systems like this can really help,and we're projecting yieldsthat can improve by about ten percentand, more importantly, decreasethe amount of inputs such as waterby 25 percent by usingaerial robot swarms.

Lastly, I want you to applaudthe people who actually create the future,Yash Mulgaonkar, Sikang Liuand Giuseppe Loianno,who are responsible for the threedemonstrations that you saw.

Thank you.


来自:VOA英语网 文章地址: https://www.veryv.net/html/20180914/Vijay-Kumar-The-future-of-flying-robots.html