字幕 (515)
0:00- [Derek] There are
robots the size of bees,
0:03others that can jump on water,
0:05and some that are powered
0:07by tiny combustion engines
the size of a penny.
0:11One day they could work in swarms,
0:13they could save your
life, or even spy on you.
0:18We got access to the best micro
robotics labs in the world.
0:21- Whoo, that flipped.
0:22- [Derek] To learn
how these robots work
0:25and what are they for.
0:27This video is sponsored by Onshape.
0:30This is a tiny yellow submarine.
0:33Under water, it can move around
0:35by flapping these miniature
wings nine times per second,
0:39but you can use those same wings
0:41when the submarine is out of the water,
0:43only now you have to flap
them 250 times a second
0:48So, this robot can do both.
0:53but since it weighs only 175 milligrams,
0:55about the mass of two Cheerios,
0:57surface tension is a problem.
1:00- That's a consequence of
physics at a smaller scale.
1:03The surface tension is like a wall
1:04that blocks the transition process.
1:07- This happens because water molecules
1:10Groups of these molecules
pull in all directions,
1:13but at the surface there's no water above,
1:15so the pull is only
sideways and downwards.
1:18This imbalance creates
strong, cohesive forces
1:21that compress the surface
into a tightly packed layer,
1:24making it difficult to break.
1:27This is the same effect
1:28that lets water striders walk
1:29on the surfaces of ponds and lakes.
1:33This other robot weighs only 68 milligrams
1:36and by using a spring mechanism
that mimics a flea's leg,
1:39it can jump without breaking
the water's surface,
1:42just like a water strider.
1:44(upbeat music)
(water bubbles)
1:46It's like there's solid ground below.
1:51Now that's great if you wanna
stay on top of the water,
1:54but this barrier can also be a problem
1:56if you want to go underwater.
2:01the submarine splits water
into hydrogen and oxygen
2:05and then stores these gases
in a buoyancy chamber.
2:08It does this because the
wings are super fragile.
2:11If they started flapping while the robot
2:13was still trapped under water,
they would break right off.
2:17So, the buoyancy from the gas
helps bring the fragile parts
2:21of the robot out of the water,
2:24but the robot is still stuck
2:25in that top layer of surface tension.
2:28So, a sparker inside the
chamber ignites the gas,
2:32and the explosion breaks
the surface tension
2:34and shoots the robot 30
centimeters into the air.
2:39And once it's free, this robot can fly.
2:44This robot found a different way
2:45to break through the surface tension.
2:47It uses these large water
repellent copper pads on its feet
2:50to walk on the water,
2:53but when it needs to dive beneath,
2:55it applies 600 volts to those pads,
2:57which creates a positive charge
2:59that attracts water molecules to it
3:01and breaks the hydrophobic barrier,
3:03and that allows it to sink on command.
3:05(gentle music)
(water gurgling)
3:08Then, once submerged,
it can walk underwater.
3:13Both of these robots were
made by Dr. Kevin Chen at MIT.
3:17- We're looking at the flight room,
3:18and this is where we do all
of our flight experiment.
3:21As you can see, it has
motion capture cameras.
3:26is one of the only places in the world
3:27where robots this small attempt flight.
3:29- [Henry] Okay, so because
this robot's so small,
3:31it is such low inertia, right?
3:33So, you're saying that
it could flip faster
3:34than any other drone in the world?
3:36- Beyond 7,000 degrees per second.
3:38I mean, you can actually hit the button.
3:40- You'd let me hit the button?
3:44- Okay.
- Three, two, one.
3:46(robot buzzing)
(triumphant music)
3:50(defeated music)
(robot buzzing)
3:52- We know who's fault it is.
3:55- [Derek] Getting these
robots flying is tough.
3:58I mean, they're the size of bees,
4:00so the internal mechanisms
have to be even smaller,
4:03like the parts of a watch.
4:07- [Derek] Components
have to be precise
4:08to within five microns.
4:10That's a 10th the width of a human hair.
4:14we have those very big
flies zipping by in the lab,
4:17and I was making the statement of,
4:18"Oh, they're just showing off."
4:31- That flipped.
- There we go.
4:40- So these bots fly but not like birds.
4:43I mean, they don't soar.
4:45Instead, they have to use
a whole lot more energy,
4:47flapping their wings
hundreds of times per second.
4:51So why do they do that?
4:53Well, it comes down to
this scale phenomenon.
4:57Larger objects typically
have less surface area
5:00relative to their volume,
5:02and that's important.
5:04Let's just approximate a flyer by a cube.
5:07Let's say it's 10 centimeters on a side.
5:10Well then that would have
a volume of 10 by 10 by 10,
5:14or a thousand cubic centimeters,
5:16and it would have an area
of 10 by 10 by six sides,
5:19600 square centimeters.
5:21So the surface area to volume
ratio would be 0.6 to one.
5:26But now imagine we have
a much smaller flyer
5:28that is just one cubic
centimeter in volume.
5:31Well, its surface area is going
to be one by one times six.
5:36That is six square centimeters.
5:38So that's gonna be 10
times the surface area
5:42It's gonna have a surface
area to volume ratio
5:46Now why is that so important?
5:48Well, it's because drag
depends on surface area.
5:51So if you have more surface area to volume,
5:53well you're gonna have a lot more drag,
5:56and also, at that small scale,
5:58you'll be much lighter
relative to that drag.
6:01So you're not gonna have as much inertia,
6:03so you'll get pushed
around more by the air,
6:06so you can't just soar
through it like a bird.
6:10and other insects flap their wings a lot.
6:14What they're doing is
generating swirls of air
6:17above the top of the wing,
6:18and those vortices create
low-pressure zones.
6:22When combined with the high
pressure below the wing,
6:24that is what generates lift.
6:26So they're pushed up into
that low pressure region
6:29by flapping their wings back and forth.
6:35This robot was inspired by
seeds from a maple tree.
6:40Their unique shape creates
the same swirling vortices
6:43above the seed's leading edge,
6:48they spin and generate
surprisingly high lift.
6:52These seeds are still just falling,
6:55but if you add miniature electric rotors
6:58to the ends of each wingtip on this robot,
7:00then it can generate enough lift to fly.
7:03(gentle music)
(robot whirring)
7:08But this robot isn't quite insect scale.
7:12It actually weighs about 50
times more than the RoboBees
7:17So to power something that small,
7:19you can't just use electric motors.
7:21I mean, the magnets and coils
don't scale down effectively
7:24to such a small size.
7:29So to power the first RoboBees,
7:31they had wings driven by special crystals
7:33called piezoelectric crystals.
7:36By applying a voltage across the crystal,
7:38they contract slightly,
7:40but only around 0.1%,
7:42not nearly enough of a
deflection to make a robot fly.
7:46So roboticists designed a chassis
7:48that mechanically amplifies
the motion 30 times.
7:52If you then turn the voltage on
7:53and off 120 times per second,
7:56the RoboBee flaps its wings and flies.
8:01But there is a downside
to piezoelectric crystals,
8:04which is, they're fragile.
8:06Even a small impact to the
wings and the crystal cracks,
8:10and the RoboBee stops working.
8:12So at MIT, they are building
their RoboBees differently.
8:17- They're so confident that
it will survive being dropped
8:20that they're gonna let me
throw it off of a building.
8:22So, I mean, here we go.
8:27(laughing)
Okay, we gotta go see it. Come on.
8:29(gentle music)
(henry laughs)
8:34I mean, it looks pretty good.
8:37Like how does this thing
survive? I have no idea.
8:40- [Derek] Well, these
robots have a secret ingredient.
8:43- So put into scale.
- Okay.
8:46- [Derekr] Instead of using
piezos to drive the wings,
8:48these bees use soft polymers.
8:50They effectively work like tiny muscles.
8:55- [Derek] They take a polymer
8:56and they coat each side
with carbon nanotubes
8:59that creates two effective
conducting plates.
9:02So if you apply opposite
charges to these plates,
9:05that pulls them together,
stretching out the polymer.
9:08But if like charges are applied
to both plates, they repel,
9:12and so the polymer shrinks
9:15and if we roll up layers
like this into a tube,
9:17we can amplify the force they generate.
9:20It stretches up to 25% of its length.
9:24By cycling the voltage
hundreds of times per second,
9:27these muscles drive the RoboBees wings.
9:32- When you shrink down to smaller scale,
9:35your fly wing frequency goes up higher.
9:37So we are at the 400 hertz range.
9:39- Which is right in between
a honeybee and a mosquito.
9:41- Yes.
(group laughs)
9:44- [Derek] This flexible
muscle can take bumps
9:46and scrapes and keep working,
9:48but if it's pierced by a needle,
9:50the carbon nanotubes get pulled in
9:53and then the plates touch,
9:54causing a short circuit that
renders the muscle useless.
9:57But the scientists have even
found a way around this.
10:00When high current is cycled,
10:02the carbon nanotubes that
are touching burn off
10:05and so the muscle self-heals.
10:08Kevin and his team even invented a process
10:10to perform laser surgery on the robot.
10:13- You're creating smaller defect
10:15around a very, very big defect
10:17and then by isolating the small defect,
10:19you're using the small defect
to isolate the big defect.
10:21So that was what we call
10:22the laser-assisted clearing process.
10:25- One robot was really
tested to its limits.
10:28Its artificial muscle was
pierced by cactus needles
10:31and hit by a laser beam
(laser sizzling)
10:33and it could still fly,
10:36but these muscles are energy intensive
10:39and for robots at this scale
that have to be so light,
10:41there's no room for extra batteries.
10:44Luckily, there is another
way to get around.
10:48- [Henry] That's a jumping, flying robot.
10:51- [Dr. Muller] This RoboBee
conserves energy by hopping.
10:54This tech was used on another drone
10:56at the City University of Hong Kong.
10:59Normally this drone can
only fly continuously
11:03but with the hopping attachment,
11:05it can keep moving for 50 minutes,
11:07nearly 10 times longer.
11:11Scientists believe this
could be even more effective
11:13in low-gravity, low-air
resistance environments like Mars.
11:17So it would be perfect for
an ingenuity version 2.0.
11:21But microrobots are
already being used today.
11:24(gentle music)
(airplane whooshing)
11:27Every day planes complete
hundreds of thousands of flights
11:30and most of them have
multiple turbine engines.
11:33Now a crack in a turbine
can be catastrophic,
11:36so manufacturers inspect
them every 3000 flight cycles
11:41but inspections cost tens
of thousands of dollars
11:44and can take a whole day.
11:46That's where this cockroach-inspired
robot from earlier,
11:50It's incredibly fast,
11:52it can run 10.5 body lengths per second.
11:55Speaking in relative terms,
11:57that's faster than a
horse, and it's versatile.
12:00Its special foot pads can apply a voltage
12:03to polarized metal surfaces,
12:05creating an opposite
charge underneath its feet
12:08and that's how it's able
to stick to metal surfaces,
12:11similar to a balloon sticking to a wall
12:13after you rub it on your hair.
12:15Rolls-Royce and Harvard
are working to put HAMR
12:17inside of engines to
inspect for turbine cracks
12:22And since its mass is so small,
12:24adhesion forces are much
stronger relative to its weight.
12:28So HAMR can get into some tight spaces
12:31and that can be pretty useful.
12:34(sirens wailing)
One of the first times
12:35that robots were deployed
in an emergency situation
12:38was during the 9/11 search
for survivors at Ground Zero.
12:42Unfortunately, they didn't
turn out to be that helpful.
12:45They were big and expensive
and they'd get stuck.
12:49Three different types inspected
eight sections of rubble,
12:52but none found survivors.
12:55So an ideal rescue robot
12:57should be able to navigate tight spaces,
12:58withstand damage and debris,
13:00operate across varied environments,
13:02and be inexpensive enough
to be replaced if destroyed.
13:06- The material cost is actually quite low
13:08for making the robot.
13:10The human labor is high,
13:12but in terms of the material, right,
13:13couple of dollars per robot,
13:15but it's really not that much.
13:17- So the idea is to deploy swarms
13:20of insect-sized microrobots
13:22to search for survivors in disaster zones.
13:24(dramatic music)
(siren wailing)
13:29But I understand when I say swarm,
13:31you might get a little worried.
13:33I mean, swarms of miniature killer robots
13:35are straight out of dystopian sci-fi.
13:39Think the hunter-seeker from "Dune"
13:41or the killer robot bees
from "Black Mirror."
13:43- You might be familiar
13:44with that like famous
"Black Mirror" TV episode
13:46where all like the bees.
13:47- Yeah, yeah.
- Yeah.
13:49everybody that I had ever
met in my entire life
13:51sent me a text message and was like,
13:53"Hey bro, you seen this?"
13:56- [Derek] But this
idea isn't so farfetched.
13:58In the early 2000s, bees were dying off.
14:01- It's called
colony collapse disorder.
14:03- Congress is holding hearings,
14:05even the Vice President has been briefed.
14:06- In fact, the whole
RoboBee project started
14:08with the goal of replacing the bees.
14:11Thankfully that idea didn't last long.
14:14- Bees can do much better jobs
14:16in terms of pollination than
those robots much more cheaply.
14:21you need a huge colony of
bees to do those effectively.
14:26Also, from an environmental
protection perspective,
14:28I think it doesn't make sense
14:29to replace bees with robotics bees
14:33from a cost-effective perspective
14:35and also from the
perspective of, you know,
14:37if you have so much money,
14:39why you making those bees
than protecting the real bees.
14:42- Okay, so they won't replace the bees,
14:44but I can still easily imagine a world
14:46where these same robots
14:47that are supposed to help in a disaster
14:49are secretly being used to spy on me.
14:52(tense music)
(robot buzzing)
14:54I mean, it's a bug that would
literally look like a bug.
14:59(shower water dripping)
That's terrifying.
15:00- Is there any fear from you
15:02about what they could be
used for, like ethically?
15:05- We really focus on
the fundamental science
15:07and solving the fun technical problems.
15:09And as a society in general,
15:11we all should think about collectively
15:14how to prevent those new
technology from doing harm.
15:17- But we're getting a
bit ahead of ourselves.
15:20I mean, most of the robots we've seen
15:21aren't able to spy on us.
15:23In fact, they're not
even fully autonomous.
15:26- We have offboard sensing
from those cameras,
15:30you have offboard power from those,
15:32and offboard computation.
15:34What you see today is
everything is offboard.
15:36But hopefully in five years,
15:38then we can combine both sensing autonomy
15:40and that's the longer-term goal.
15:42- Harvard's RoboBee has
managed short bursts
15:45of untethered autonomous flight.
15:47So it's fair to say we aren't that far off
15:50from robot insects
operating freely around us.
15:54Still, there is a limit
15:55to how far these robots
can go on just batteries.
16:00Batteries need shielding
to prevent damage,
16:02short circuits, and leaks.
16:04And the thing is, as
batteries are scaled down,
16:06this shielding has to stay
about the same thickness.
16:09So that means smaller batteries
16:11become increasingly inefficient.
16:14And that's ignoring that
the energy-to-weight ratio
16:16of batteries is just fundamentally lower
16:18than that of chemical fuels.
16:21At the insect scale,
every milligram matters.
16:25- We just said, like let's
just sail past all of that
16:27and just use a video game cheat code
16:30and just power our robot
16:32with the smallest explosions possible
16:34and put two tiny internal
combustion engines on board it,
16:40(upbeat music)
(robot tapping)
16:41And it sounds like a combustion engine,
16:42which is probably my favorite part of it.
16:47- [Dr. Muller] Cameron's
penny-sized engine runs
16:49on a constant stream
of methane and oxygen.
16:52This is fed into a chamber
where it's ignited by a spark
16:56and so it combusts,
releasing a burst of energy.
16:59The hot gases rapidly expand,
17:01pushing against a
flexible polymer membrane
17:03that acts like a piston.
17:05- So the membrane moves as the piston
17:07and then instead of having
to, like, have any sort
17:10of elaborate system
that brings it back down
17:12because it just naturally is elastic,
17:14it sort of has its own restoring force.
17:16That was our clever little innovation.
17:19- As the membrane shrinks back,
17:21it vents the exhaust gases,
allowing the cycle to repeat.
17:25Despite the continuous
flow of methane and oxygen,
17:28the fuel line never catches fire.
17:31That's because as explosions get smaller,
17:33their volume shrinks much
faster than their surface area.
17:36This causes them to lose heat more quickly
17:38to their surroundings.
17:41only a small amount of
gas burns at a time,
17:44so heat quickly escapes
into the fuel line,
17:48and stopping the flame from
traveling back up the line.
17:52With just two of these combustion chambers
17:55one for the front legs
and one for the back,
17:57Cameron can control its heading.
17:59- You can actuate just
one of the two sides
18:02at a given time if you want
18:03because both sides are operational.
18:05So if we spark in both
sides, it'll move straight.
18:07But if we just do one or
the other, it'll pivot.
18:09- [Derek] And this robot
18:10is super powerful for its size.
18:13- It weighs 1.6 grams,
18:15which is about as much
as a gummy bear weighs.
18:19It can jump like two feet
in the air approximately.
18:21It can carry 22 times its body weight,
18:23which is about what a cockroach
or a lot of beetles can do.
18:26We'll be able to put
a fuel tank, you know,
18:29microelectronics sensors,
a camera battery,
18:32and still have weight left over to go,
18:34and this thing will still chug along.
18:37That's the future. That's the goal.
18:41- Scientists have created robots
18:43that can do what some insects do
18:45and there are clear
applications for this work.
18:49But for these roboticists,
these self-labeled misfits,
18:52that's not what it's all about.
18:55- If it's about application,
18:56we should all like make a startup
18:57and try to like think about what we can do
18:59to make money, right?
19:01We think there are nice applications,
19:03like inspection and search and rescue,
19:06but I would say as a research lab,
19:08we are mostly driven by curiosity.
19:10I think that's a very
honest answer. (laughs)
19:13(quirky electronic music)
19:20who designs hardware like
these robots, mini or mighty,
19:24you know that you need to
create a CAD model first
19:27so you can prototype and
bring your ideas to life.
19:30But the problem with most
mainstream CAD software
19:33is you need a very powerful
computer and a mighty budget.
19:36That's why we reached out
to today's sponsor, Onshape.
19:39Onshape is a modern CAD
19:41and PDM system designed for businesses.
19:44And unlike any other
professional CAD system,
19:47Onshape is built entirely in the cloud,
19:49powered by Amazon Web Services.
19:51That means you don't need
expensive computer hardware,
19:54it never crashes, and
you never lose your work.
19:58It just runs in your
browser on any system,
20:01be it Windows, Mac, Linux,
or even on your phone.
20:05I can actually pull up that
mini robot model right now
20:08and start moving some
of these things around.
20:11It's also built to be collaborative
20:13so you can work alongside friends
20:14or colleagues like never before
20:16on the same design in real time.
20:19For businesses, you can
trial Onshape for six months
20:22and for students' hobbyists,
it's completely free to use.
20:25Just head over to onshape.pro/veritasium
20:28and sign up to get started today.
20:30Onshape also uses tried
20:31and tested Git-style
version control methods
20:33from software design so
you don't have to worry
20:36about sending multiple large files around
20:38and keeping track of all
the different iterations.
20:40It is all done for you.
20:42None of this emailing V2 final final
20:44for real this time one nonsense.
20:47If you work for a government agency
20:49or have contracts with the
government, I have good news.
20:52Onshape now has a plan that enables teams
20:54to comply with federal
regulations like ITAR and EAR,
20:58so nothing is holding you back.
21:00Go sign up for free at
onshape.pro/veritasium
21:03to get started engineering
for the future today.
21:06I wanna thank Onshape for
sponsoring this video,
21:08and I want to thank you for watching.
