Rovers Get Around

What's the perfect way for rovers to get around? Why settle on the wheel? We look at the challenges of designing rover mobility systems.

Learn more about your ad-choices at https://www.iheartpodcastnetwork.com

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:00):
Brought to you by Toyota. Let's go places. Welcome to
Forward Thinking. They've even welcome to Forward Thinking, the podcast
that looks at the future and says moving right along.
I'm Jonathan Strickland, I'm Lauren Obama, and Joe McCormick. You guys,

(00:22):
I have like a super awesome, cool description of a
DARPA challenge. I want to tell you a DARPA challenge. Yeah,
so you know DARPA. That's that that that big old
facility or agency I should say, in the United States
that ends up funding lots of stuff for essentially the
Department of Defense. Without DARPA, we wouldn't necessarily have things

(00:43):
like autonomous cars that are coming out. Google's autonomous car
wouldn't be a thing because a lot of people who
worked on that originally developed autonomous cars as part of
a DARPA challenge. So DARPA issues these challenges out to
pretty much any entity that can compete in them and says,
if you're able to do you know, this particular task

(01:04):
in this particular way with something, then when you will
be awarded a certain kind of prize. So in this case,
the Darker challenge, I wanted to talk about has to
do with the robotics. They don't all necessarily deal with robotics,
but this one specifically does. So here's the challenge. You
first have to build a robot that is capable of
moving around and sensing an environment, be able to leave

(01:25):
a building, climbed down a set of stairs, go to
a golf cart, get in the golf cart, and drive
the golf cart to a different location in dynamic situation.
So there could be traffic, there could be pedestrians, and
the robot has to be able to navigate around these
things without causing harm to itself or others, you know,

(01:45):
three laws. And then it gets to a new location,
gets out of the golf cart, goes into another building,
climbs a different set of stairs, then has to break
through a wall, and then find a fire hose, pick
it up and point it at a target. Now, does
it have to turn the fire hose on? Not necessarily,
at least that's not how it was described to me.

(02:06):
So the whole point of the exercise is to really
push the limits of what we can do in robotics.
I mean, that's the case with all the DARPA challenges.
It's it's meant to be something that really makes engineers
and scientists work extra hard in order to achieve it,
because these are extraordinarily difficult things to ask of anyone,

(02:26):
but also very useful and and extraordinarily difficult. Absolutely, like
I couldn't do that, Like, I mean, the kool aid
man would be able to break through the wall apart,
but not I think that the rest could be a challenge.
This is exactly yeah. Yeah, I did not say that
the robot had to yell out oh yeah after coming
through the wall. So I think I think it's the
only obvious. I think you kind of have to have
it do that. But so anyway, the winning group I

(02:49):
think gets like two million dollars if they win this, uh, which,
in the grand scheme of things, when you think about
how much work goes into making something like this happen,
is probably getting pay it about about the you know,
it's probably maybe even taking a loss, who knows. But
the idea is that you'd have to have a robot
that's strong enough to break through a wall. It would
have to be able to maintain its balance by going

(03:10):
up and downstairs. It would have to be able to
be environmentally aware, so it had to be able to
to uh navigate around obstacles, have to be able to
actually navigate period, and have to have some form of
of maps inside of it, our GPS that could know
how to get from location A to location b. Um.
There are a lot of different elements to this challenge,

(03:30):
and none of them are easy. I have a question,
does it have to be a two legged robot? Or
can it be a six legged robot that goes downstairs
and drives a golf cart? It does not have to
be a two legged robot. A lot of the teams
are focusing on making a human like robot because a
lot of the stuff we happen to operate, say like
golf carts, happens to be made for human shaped things

(03:51):
and not six legged uh insect bots. But there's nothing
specific from what I understand because I actually asked that
same question. I said, I said, is it does it
have to be like a bipedal robot? And I was
told no, it doesn't have to be. In fact, if
you can figure out a way of engineering around that,
then you could do that. And in fact, that's kind
of what leads me up to this topic I wanted

(04:11):
to talk about. Is just the idea of how robots
get around places. I mean, that's that's just one aspect
of all the difficult things we we just mentioned, but
it's still all in its own, hugely challenging. Yeah, the
idea is robot locomotion. Locomotion is different from just motion.
So robot motion is something we've been working on for
a long time, making an arm that welds a car

(04:34):
door or you know, or a cartesian robot that moves
over something moving along under it on a conveyor belt
and it can punch holes in it or something like that.
But locomotion is talking about getting from one place to another,
which is not necessarily as easy as it sounds, right,
And in fact, if you look at a lot of
the early robots used in manufacturing, those robots that were

(04:57):
not meant to move, they weighed tons and tons and
that you set them down. In fact, only were they
not meant to move, they were not meant to be
near people, like human interaction was not a thing because
these were giant, powerful machines that could do serious injury
to someone if they got in the way. But now
we're talking about a world where our robots are coming
into closer contact with us, and also that we want

(05:19):
to send off to other places, how do we make
sure they can get around? Right? And so robot locomotion
is a big topic, and I think we could do
a bunch of episodes about locomotion generally, I think today
it would be cool to focus on one really specific
challenge and maybe one of the hardest jobs in robot locomotion,
which is getting around on other planets. Right, So we're

(05:41):
talking about rovers here, the idea of some sort of
robotic entity that can move around on the surface of
a planet and it's receiving commands all the way back
here from Earth. In some cases, I mean we'll actually
really in pretty much all cases, you have to have
a lot of autonomy built into the robot because from
the time it comes into contact with something to the

(06:04):
time when someone can react to it and send a
message back to the time when it can then react
to the message can be half an hour. Like within
the case of of Mars, I think it's something like
fourteen minutes each way. Well that that is one potential
time because earthen mars are never Earthen mars are not
always at the exact same distance from each other. Fourteen minutes.

(06:25):
That's either the maximum or the minimum or the average.
That number occurred for a reason, covering all the basis,
you know, that was those specifically when the Curious Curiosity
rover was landing on Mars. At that point, Earthen Mars
were no longer at their closest points. Um they you know,
it's it's one of those things because the planets are

(06:46):
always in motion, you have to you have to plan
out these launches very specifically in order to conserve as
much fuel as possible because as we know, fuel adds weight.
But all of that aside, you know, even getting all
of those challenges, which we could do a full episode
just on the of getting the Curiosity or over to Mars,
there's still the challenges of how does the rover move
around once it lands on the planet. Right, So I

(07:08):
think today we should just talk about a few different
potential designs for how rovers get around, and before that,
maybe we should talk about some general concerns. Um by
no means is this list exhaustive, because there are tons
of concerns that go into designing a robot. But think
there are some good basic groups. Yeah, yeah, right, So

(07:30):
number one I think is something that you might not
think is important when it comes to somebody like NASA
that has a government funding. Um. But actually it's very
important even to space exploration, which is just cost and
ease of design, specifically to space exploration, which we've seen
received budget cuts several times of the last few d decades. Really, right,

(07:51):
so pretty much all engineers have to consider this. You know,
what are the cost of the things you're using, the
cost of testing them, the cost of designing them, and
how feasible is it to make this thing work in
time for launch. Uh. There's also issues of mobility like
how fast can it go, how definitely can it move
around like it's turning radius and things like that. There's stability, um,

(08:15):
which you could address two ways. One you could make
a robot that is very resilient to falling over or
resilient hasn't it doesn't fall over, Or you can make
one that doesn't mind falling over. Yeah. This makes me
think of those remote controlled cars that they that came
out when I was a kid. The these these were
designed in such a way that the car flipped over,

(08:36):
you could continue controlling it. Uh, it would just mean
that some of your controls would be reversed because the
car itself was designed so the wheels were a greater
diameter than the right And I've and I've seen some
like military industrial robots that operated on a very on
a similarly small scale, um in a similar way. And

(08:57):
note here that that making instrumentation attached to your robot
that doesn't mind getting tumbled all around might be one
of the trickier parts of this particular Uh yeah, sure
you might be able to get a robot to land
on the planet and roll around a lot, but if
you are designing a robot that can do that on
its back, it probably doesn't have a lot of delicate
equipment on it. Yeah. Uh yeah, So there's there's that.

(09:19):
There's also the integrity of the robot as a as
a structural unit. You don't want it to get stuck
or to get broken easily. Uh. There's energy efficiency, which
is a big deal. When you're shooting something far away
from the Earth, you can't plug it in to recharge it. Yeah.
We haven't got a Tesla supercharged station on the Martian
soil yet, and I don't think they have any plans

(09:39):
in the near future to address that. No, So you
either need like, really, come on, you're you're with SpaceX
for gonness sakes. You need like a protected mobile power
source that's gonna last a long time, like like curiosities,
uh little nuclear diddy um, or you need something like
solar panels to continuously absorb power from the sun. But

(10:01):
but even with something like solar panels on mars Um,
you've got some problems because solar panels number one or
not that efficient. You're not going to be generating a
whole lot of useful power. That's limited. Number two. They
can get stuck out of alignment with the Sun, or
they can become covered in dust. They're big and delicate
and clumsy. You could even have your robot in a

(10:23):
region of the planet where as it goes through its
orbit it is no longer, you know, getting that much
sunlight at all, which is what happened to a previous
Martian rover that that is sadly no longer with us.
Uh though Hey, a bonus of rover design is that
these things should ideally move pretty slow in general, like
like the Curiosity, for example, can go about four centimeters

(10:45):
per second that's like point one three ft per second
um and that low speed helps you um maximize the
engine's output for for strength. It's likely you can get
a lot more torque there than uh, and that's what
you're concerned with, because you may be climbing an incline
or going over a rock, and that's that's when you

(11:06):
need a lot of torque. But you don't necessarily need
to go fast. It's not like, as far as we know,
no Martians have been chasing any of these. It's not
like this rover needs to go point one four ft
per second or it'll explode. Right Well, I mean the
thing about is luckily not on Mars. Luckily that cuts

(11:28):
down on a lot of scientific problems. From what I understand,
I'm campaigning hard to get Kiano Reeves sent to Mars.
Oh I've actually started a Kickstarter Yeah yeah, I reached
funding soon. But then there's also just the concern of
how do you get the rover to its destination, which
is a much bigger problem than most people would probably realize. Like,
you can't just put in our c car in a

(11:50):
cannon and shoot it at Mars. You can, but it's
not gonna it's not gonna make it there. It has
to be something that you can put into a space
veha goal and something that you can deposit on the
foreign surface, whether it's you know, Mars or the Moon
or wherever you want to take it without damaging the
thing so that it can move around. This is not

(12:11):
easy at all. Again, the Curiosity Rover really proved that
it was. It was one of those plans. When I
first heard it, I thought, well, that's crazy. Yeah, it
had some crazy rocket parachute thing. Well, because the Martian
atmosphere is not is not uh viscous enough really to
have a parachute be very effective. I mean, it was
able to slow the descent a little bit, but not

(12:32):
like it would here on Earth, and it had to
have rockets retro rockets essentially to slow its descent or
else it never would have been able to make it
down onto the surface safely. But also I mean pretty
wicked shocks in order to to to let the wheels
support its weight from a from a even the gentle drop.
And so we're we're talking about like again, it was

(12:52):
one of those where this whole thing had to happen
autonomously because the rover was going to enter the atmosphere
land on the plant's surface in less time. Than it
would take for it to send a message back to Earth.
So we were when we were waiting to hear about
whether or not the Curiosity Rover had set down safely,
it had actually done that fourteen minutes previously. So we

(13:15):
found out and we all celebrated, but the actual event
had happened fourteen minutes earlier. It's pretty incredible stuff when
you think about it. And not only that, but you
really have to make sure that your design is compact
so it can fit within the confines of a spacecraft.
It's not like we have a a you know, the
ability to put any shape of spacecraft up in space.

(13:36):
It's all dependent upon the rockets that you use, so
you need to make sure that you're really conserving that space.
Just like you don't want to have a whole lot
of extra weight or and you don't want any extra
weight if you can avoid it. You don't want to
take up too much space either. Yeah, so I guess
we should start with something like the Curiosity Rover and
the other rovers that we've sent to Mars with the

(13:58):
obvious one wheels. I mean, this is the obvious way
to get a machine to move around, and there are
pretty good reasons that it's the go to system for
rover locomotion right now. Yeah, they're they're simple for one,
one of the simplest machines that's out there. In fact,
it's you know, you to look at the the inclined plane,

(14:19):
the the lever well, the wheel and axles right up there. Yeah,
so that's the simpler you get, the better. I mean,
that's one of the greatest strengths, right is because it's
such a simple design, there are fewer parts to break down, right,
So that's a big, big advantage. Yeah, it's just it's
less complicated robotically. I mean, when you think about something

(14:40):
like a robotic leg, think about all of the different
movements and software calculations that have to take place just
to move a leg from one place to another, let
alone taking into account all the different types of terrain.
Wheels work on a lot of different kinds pretty well.
Especially since we're talking up out rolling at a very

(15:01):
controlled rate. It's not like we're talking about trying to
you know, do uh wheelies across the surface of Mars
all that would be awesome. Um, it's so we're not
worried so much about speed. We're worried more about the efficiency.
We're worried about that power and for a lot of
them the terrain we're looking at, wheels work just fine. Yeah,

(15:22):
if you're just trying to drive across a pretty flat plane,
wheels are going to be great. Yeah, Curiosity actually has
some really cool wheels. Yes, it does rather. Yeah, it's
got some indentations in its wheels which allow it when
it when it rolls over uh, soil, the regular it
actually leaves behind little, um little patterns, and by maneuvering

(15:47):
the robot in very specific ways, engineers on Earth can
command the Curiosity Rover to roll around in such a
way that these little uh, these little markings, these little
raised portions from the indentations in the wheels end up
spelling out Morse code messages. And so they've sent messages.
You know, they've left messages for things like you know,

(16:07):
putting in the initials JPL, things like that. Um, so
it's cute stuff like that. Although they've also used it
in order for the Curiosity Rover to to leave a
mark behind that's just says this is your starting point. Okay,
so like navigational stuff. Yeah, it's not just for you know,
it's not just not just the equivalent of IBM manipulating

(16:29):
individual atoms to spell out it's it's its name. Yeah. Um. Also,
I think wheels are generally about as energy efficient as
you can get. Now, if you are an engineer or
robotics expert who would love to correct me on that, please,
I invite that. But it seems that a wheel is
is peak energy efficiency to me, it's pretty compact. Yeah,

(16:52):
A wheel turning is a much lower Uh. It's it's
much less energy to turn a wheel than say, manipulate
a robotic limb that might have multiple servos and joints
in it. Right. Well, I'm just thinking about like the
different mechanical energy required to, say, uh, ride a bike
across a flat plane a certain distance. You're getting into

(17:15):
some gyroscopic elements that don't apply when you're when you're
turning the wheel at a speed of point one three
ft per second. But I understand what you're saying, uh,
like if you you know, because uh, one of the
things about a bicycle is that when you're moving at
a certain speed, you have this gyroscopic effect that actually
keeps you upright and helps you maintain balance. And uh,

(17:38):
that would also quote unquote make things easier. So it's
a little different, but uh, in general, yes, I agree, Um,
there are some disadvantages. Oh yeah, Well, for one thing,
wheels can get stuck especially Yeah. So in the spring
of two thousand nine, the Spirit Rover, which by the way,
I do not want to paint the Spirit Rover as

(18:00):
a failure. The Spirit Rover long outlived its planned mission
and it went all over the place. It It did
a great job and it inspired lots of people. I mean,
the NASA did a great job with getting the word
out to the public about the Spirit, about the Curiosity
as well, where it excited a lot of people about
the space program, which is one reason it was kind

(18:21):
of sad when, like I started to say, in the
spring of two thousand nine, the Spirit became stuck. It
was going across an an area called Troy on mars
Um and its wheels became stuck in loose soil. And
so for a long time after that, basically NASA was
just trying to figure out a way to plan and
escape action for the vehicle, but it never got out

(18:44):
of the loose soil. Should have launched an enormous wooden
horse to land right in front of it they could
climb into and that would have been fine. I read
my history, your your history, the idea. Okay, so that's
not the only kind of surface that that wheels have
trouble sometimes with what else have we got right? You

(19:06):
could say, what if you needed to get over a
pile of loose rocks, Well, in that case, wheels could
have a lot of trouble. Yeah. Or if it's a
if it's really really rocky terrain, wheels may not be
able to uh may not be large enough to get
over them. There's also what if there's a crack in
the ground larger than the diameter of the wheel. I
mean that's then you just get stuck. Yeah, if the

(19:28):
wheel gets to the point where it's no longer making
contact with the ground, you could be really seriously stuck.
Or you're just talking about inclines. If the incline is
steep enough, the wheel may have to work harder than
the machine is capable of generating. I have to generate
more power than it's capable of doing, and then you
can't go any further. Or if you happen to be

(19:50):
descending an incline, you might end up losing control, right,
but despite these disadvantages, I don't want to give the
impression that I think wheels are a thing of the past.
I think wheels are awesome and it's amazing what you
can do with wheels and smart engineering, like the like
the rock or Bogeye system that they use for the
suspension of the drive on the Curiosity rover. Right, these

(20:14):
are the Curiosity. So it's got six wheels and each
can be independently controlled, and each receives torque directly from
the engine UM. The axles for each wheel are also
independent UM. Each is connected to this UM suspension device.
It's sort of like a leg like device. Yeah, so
you've got a lot of maneuverability there. You've got a

(20:35):
lot of ability to uh to to give yourself the
propulsion you need to get around or over obstacles. It's
pretty ingenious. Yeah. If they're actually videos online you can
go look up that are pretty cool. The jp L
researchers trying to design the system and they're testing it
in the lab, driving it over these little humps and
it's it's the most inspiring crawling over a hump you've

(20:58):
ever seen at the speed of a few centimeters per second. Yea,
so those are wheels. Let's talk about legs. Legs. How
come we don't have rovers with legs? Okay, So legs
are great. I love them, love me, love me. My
gams are great, fantastic. Well, they're really great for biological
organisms like us, absolutely. I mean they let us get around, uh,

(21:22):
let us get into trees and stuff out of trees. Both. Yea,
there's the less climb things like we can climb stairs,
we can get over rocks, we can do lots of
cool things that are really useful. We can jump, we
can hop, it's so nice outside. Sorry I got a
little distracted, but yeah, we were able to use our

(21:43):
legs to do a lot of things. Are very versatile
is what I'm getting down. Too great for all kinds
of terrain. Right, So in that sense, they're fantastic and uh,
and so there's a lot of work being done building
robotic legs that have a lot of versatility to them. However,
that work takes a lot of time, because holy cow,

(22:03):
is it hard to do. Yeah. I was just thinking
that I can build um a wheel and axle with
tinker toys, but I can't build legs. Yeah. The whole
part where they're incredibly useful. Is because they have all
of these different points of articulation. Yeah, and those points
of articulation give us that freedom exactly so that we're

(22:26):
able to, uh, to tackle various obstacles in different ways.
But in order to create a robot that has the
same degrees of freedom, you have to build in a
lot of different systems for that to work. And then
not just we're not just talking about mechanical systems, which
are already that's complicated enough, right, to build a mechanical
leg that can do that can move like a leg

(22:48):
that one is not taking crazy amounts of power to operate,
and two can still can still support the weight of
the rest of the robot. But then you also have
to build in the software sums that you are alluding
to earlier, Joe, the software that can control the movements
of this leg, that can can detect changes in the

(23:09):
environment and react to it so that the robot maintains
its balance and is able to continue forward. Uh. This
is not easy to do, right. I'd imagine with legs
it's more difficult to program stability. Like with a wheeled vehicle,
it's just got static stability, So you just sit it
down somewhere and it's stable. Um, if you have a

(23:32):
creature with legs that's going to going to be lifting
legs and stuff like that, it seems like there's more
potential for it to fall over. You have to shift
your weight every single time you take a step, and
it's all incredibly automatic to most of us. But I mean,
but that's why babies fall down all the time while
they're trying to work that thing out, or puppies or
adorable bust and Dynamics robots. Yeah. So so one thing

(23:55):
we can probably get all the way is that bipedal
robots that would be you know, used for space exploration
probably not gonna happen anytime soon. There's no reason. I
don't know why you need that now. I mean, the
only reason I can argue is that people would probably
have develop uh an emotional attachment to such a thing
because it resembles us enough. But then we've seen people

(24:17):
develop strong emotional attachments to the spirit and the curiosity rovers,
so it's not like that's a prerequisite. No. When I'm
imagining a rover with legs, I'm imagining probably a hexapod,
like a like a six legged rover kind of like
the Curiosity rover has six wheels, um, and I would
imagine that something like that would probably have a more

(24:40):
limited and narrow focus of locomotion, Like it would be
something if you wanted to climb a mountain on Mars
with a rover, that seems like that might be useful
for that kind of thing. It's got these billy goat
scrabbling legs, right that would like you see on real goats.
That would be that would be uh, definitely a special

(25:01):
case right there where because you're talking about a task
that is particularly difficult for a wheeled rover. You would
need to find something some alternative, right and uh, but
you wouldn't use that same legged rover to roll across
the expanse of Mars because it would be less efficient
unless you could attach wheels to the ends of those legs.

(25:23):
So what do you know, Yeah, there's actually a cool
rover that does just that. Yes, things so creepy. So
so this is called the Athlete, the all terrain hex
limbed Extraterrestrial Explorer, And this is it's being billed as
a vehicle which you might want to uh separate in

(25:45):
your mind from totally robotic rover. But but nonetheless. Okay,
so these things, if you've never seen a picture of them,
they look like giant robotic insectoid wheelers like from from Oz.
I was going to make that same compare Harrison that
it was like the wheelers from Return to ozs, which
is that's what they look like. Um. But yeah, so

(26:07):
six articulated legs, each of which ends in a wheel. Um,
And I mean really, the wheels are for the general
all purpose getting around, and the legs are for if
the thing encounters um any kind of difficult terrain, extreme
terrain I think is the official NASA terminology about that.
So the wheels can lock into place and act just

(26:28):
like feet or hands essentially. But yeah, this thing is
It's part spider, part wheeler from Return to Oz and
all nightmare in my opinion. Um yeah, so there's that.
And you can see why this would be useful, especially
if legs can double as arms and you need to

(26:48):
use them for army things. Uh, where's the king keep
his armies in his sleevies? Alright? So moving on? What
about this? Um, this ge go bot thing I'm seeing here?
It was an interesting thing I saw. This is not
for exploring planetary surfaces. But I did see it mentioned
in the context of space for the from the s

(27:10):
A And basically it's a crawling robot that makes use
of biomimicry of the geckos gripping fingers. So yeah, yeah,
they're they're using like slightly larger versions of movie Spider
Man's ability to climb. They've got these tiny hairs or
spines or whatever you want to call them that um
like dramatically increase its surface area, allowing it to cling

(27:30):
to a surface. At a microscopic level. Yeah, if you
if you look at a geckos little fingertips and you
were able to look at them on a nanoscopic level,
you would see all these hairs, and those hairs would
actually interact with material at an atomic level. So this
is a level up from that. We're not talking atomic,
we're talking micro Yeah, micro level, like molecular maybe, but

(27:54):
it's it's definitely yeah, they're they're definitely larger than the
nano hairs you would find on a geckos uh limbs,
but they they have the same effect. They're enough so
that the robot can support its own weight, right, And
the cool thing about these as these this yes, a
post points out is that they could work in space,

(28:15):
unlike some like sticky adhesives. So imagine you want to
fly around in your spacecraft or drive around on the
surface of Mars with some kind of large vehicle that
may require exterior repairs. This could like crawl around on
the outside of the hull and do those repairs for
you in low gravity. Sure right, and then feed more nightmares.

(28:39):
So then we've got a Boston Dynamics. I'm just gonna
give up all these things give me nightmares. No, actually,
I love Boston Dynamics is a weird mix of nightmares
and total cuteness. It's it's nightmares and dreams at the
same time. It's it's our favorite robotics company that also
gives us the supreme willies. He usually doesn't give me
nightmares until I see people kicking it and I'm just like,

(28:59):
get up, okay. So Boston Dynamics is another company that
has worked with DARPA, as we mentioned earlier, but they
they've been creating uh lots of different kinds of robots,
especially for military applications, seemingly for like transport. So there's
the Big Dog and the wild Cat and These are
four legged robots that they have programmed to run around

(29:21):
or walk around in different kinds of terrain. And these
things are really cool because of what they can do
in terms of legged locomotion. Like they're they're they're pretty advanced.
One of them, I remember, I don't remember which one
it was. Now, is that the big dog that you
just alluded to. Someone from the company goes up to
it and shoves it with his foot and then it's

(29:43):
already walking on a slippery i c. Surface. It starts
to stumble and scrabble and like you would imagine an
actual animal, and then it collects itself and continues to
move I mean, albeit on this slippery surface. So and
it does so apologetically, as if to say, please, sir,
don't kick me again. We were all still upset about

(30:04):
this video. It is deeply impacted. If you want to
hear a lot more about Buston Dynamics and their robotic animals, um,
we did a podcast and a video, both released on
February nine, called Robot Pets and Robot Pets of the Future, respectively.
If so so, go check those out if you want
to hear a lot more about them. Now, I don't
know if robots like this would have such a great

(30:27):
use on a foreign planet. Maybe, but I I would
have a few questions. One of them would be, so
if you want to look at the advantages of of
the stumbling robot that you know gets its balance after
you've shoved it, I wonder number one, what use would
that be on Mars unless you're expecting something to push it?

(30:47):
And then and then number two extra nightmares, thanks Joe,
is something I suspect but don't know for for sure.
I think when you would design legs for stability, it's
important to take into account the force of gravity, which
would vary depending on the rover's location. So if you
take something that's been designed to stumble and get its
balance on Earth, I don't know if that would still

(31:08):
work on like a low gravity and you would need
to mechanically engineer the legs in such a way so
that it could support the weight of the robot in
whatever in environment it was going to operate in. So now,
if it's going into a lower gravity environment, that's actually
you don't have to worry about making them stronger. Obviously,
then in software you would probably have to adjust exactly

(31:29):
how how much power it would send to all the
different servos and motors that are involved in those legs.
I imagine it could be done um, largely algorithmically, but
I I that would just be hell on testing. Yeah
it would. I mean, I'm not even sure that would
be a problem. It just seems like it might be.
I think the biggest the biggest question would be what
would you need that kind of robot to do? And

(31:51):
the only thing I could I could imagine is if
you wanted to explore a region of Mars Mars where
you knew that the the the ground was going to
be very rocky, very difficult to traverse otherwise, and that
you would need this ability for a robot that could
study itself in case it stepped on something that gave way, right,
rather than say a wheeled robot that might end up
becoming part of a uh rock slide, maybe a legged

(32:15):
robot that could catch itself. But um, you know that's
that's complete just speculation. Okay, but here's another option. Let's
go a little bit back in the wheel direction. So
we we've seen there might be some good uses of legs,
but it might be a more narrowly focused kind of thing. Yeah,
what about tank treads continuous tract locomotion. Yeah, the tracks

(32:38):
are good for things like again, navigating really uh difficult terrain.
I mean, that's why we put them on tanks, because
tanks would be driving across areas that needed you needed
to have a lot of power, so there needs to
be a lot of torque there. It also needed to
be able to go over difficult areas. So it has

(32:58):
some advantages, but those might be outweighed by the disadvantages. Well,
I would imagine that they are better at loose soil, Yes,
you wouldn't have to worry. Yeah, so they are better
at navigating through that. However, they also require more power.
You have a lot more moving parts, so that is
already an issue, right, it's not. It's not as simple

(33:20):
as wheel and axle. So the more moving parts you have,
the more opportunity there is for a mechanical failure, you know.
So there's that. There's also the fact that you have
usually a track of some sort that could become de
threaded from that, and then in that case, you you've
lost your ability to to push yourself forward and you're
it's just like getting stuck in the loose soil. Except

(33:41):
in this case you just lost the tread. Um they're
able to turn, uh well, depending upon the nature of
the treads, they're able to turn fairly tightly, but it
takes a lot of energy to do it as opposed
to a wheel. So turning anything that's on the least
track systems is a laborious process, requires a lot of power.

(34:03):
And also if you've designed the treads so that they're
really really grippy, it may not turn so well because
turning actually means that you have to create a skid
right the it can't turn like a wheel thing where
you can just start gradually introducing a turn. So it
starts to curve over to the left or to the right.
It has to start skidding itself to the left or

(34:26):
right and then continuing forwards. It's not this natural kind
of smooth curve and um. And again, if the material
that you made the track out of, or if whatever
it's traveling on is too grippy, then it can get stuck,
not being able to turn left or right. Uh. This
may just be because of the continuous track vehicles I'm

(34:46):
familiar with, like bulldozers and tanks and stuff, But I
would also tend to imagine that a design like that
is probably heavy. Yeah, I'm a I'm not a I'm
not an engineering expert with this kind of thing. Um,
But but but I think the answer here is yes,
is okay. Think think about a wheel system. All the
mechanics that you need for it are a couple of
axles and wheels have a relatively small surface area. Um.

(35:08):
Part of the draw of continuous treads is that they
have that larger surface area, but that's necessarily um, more
material and therefore more weight. They're also driven by two
or more wheels themselves, so you have at least as
much interior mechanics as a wheel system. Um. However, I
mean that that's the kind of thing that, although this
applies to wheels as well, material science is going to

(35:29):
help us improve in the future. So this leads me
to a question for you, Joe, what if we could
design a robot that on its own could propel itself
without any sort of limbs or wheels. It actually moves
itself some other way. So it's it's not it's not

(35:52):
gripping the ground with legs, it's not turning wheels, it's
not using tracks. The whole robot itself moves as a
unit in some way. Jonathan, I want you to imagine something.
Paint a picture with your words there, Joe. Okay, imagine
a big old beach ball. Okay. Now imagine it is

(36:13):
on a beach, tumbling over the sand blown by the wind. Okay,
Now imagine it is full of scientific instruments that measure
things about the environment. So far, typical day at the beach.
Now imagine it's on Mars. Okay. Not. I posed this
exact same scenario in a blog post. I wrote about

(36:34):
something that is actually being designed that the Mars tumbleweed designs. Uh.
And this is an idea that that engineers have been
talking about for a long time, the idea of a
wind driven inflatable rover for use on Mars. It's pretty
ingenious because you no longer have to dedicate a huge
amount of work to creating a power system that can

(36:55):
propel your your robot, because the robots being propelled by
whatever wind happens to be present at the place where
it's going. You can then dedicate whatever power systems you
have just to the scientific instruments. And that conserves a
lot of space and a lot of weight. Yeah, and
so there are a lot of potential advantages to inflatable

(37:15):
rovers that blow with the wind. Number one, Yeah, like
you were saying, you don't really need a drive system.
You might want to have some kind of internal steering mechanism,
but that would like you can have a counterweight that
would still take advantage of the wind as the propulsion,
but it would be kind of like a rudder on
a ship, be steering in one direction or another. Um.

(37:36):
But then again when you add that, you'd be adding weight,
which would be cutting into what the wind can provide
in terms of moving you. UM. And so one problem
this would solve is that it would be very easy
to land a vehicle like this and you wouldn't have
to worry about it breaking upon impact with the surface.
You just drop it down and it set down like

(37:57):
a beach ball. So you can just imagine it being
blown along on the surface, but not necessarily alone. You
could also, potentially, I would think, use a system like
this in a swarm search strategy. So if you wanted
to send a whole bunch of beach balls up to
Mars and they could fan out over the landscape and
explore lots of different things, which would give you redundancy

(38:19):
as well. If one of them breaks down or get stuck,
it's not such a big deal. If you have a
big inflatable beach ball that's exploring Mars, you can set
it down. How would you do that just partially deflated? Yeah,
take some of the air out of it, and then
it sits down become stationary. Yeah. See your description of

(38:39):
the swarm of of of these things just makes me
think of every concert I went to in the eighties,
just the air would be filled with these things, which
just makes me think that we'd really be giving the
Martians a great time. But I think it's cool the
I mean, they're obviously. The biggest disadvantage you could argue
is that, since you don't ultimate control the direction that

(39:02):
this thing can go in, you're very much limited by
whatever way the wind happens to be blowing. You're literally right. Well,
you're also very limited in terms of the instrumentation you
can carry, because it needs to be extremely lightweight in
order to work. Like the Curiosity can take all kinds
of arms and machines that do different tests. You would

(39:23):
you would have much fewer options with something like this.
You'd have to really pick what you want to go
on there, and you'd have to find a way to
make it very small and light and furthermore capable of
working within this this beach ball capacity. I mean, I'm
not sure if you would be able to send a
little scooper thing out into the sale right point, you
might be able to have some sort of semi permeable

(39:44):
membrane as the beach ball body, and then therefore he
could take samples of like atmosphere examples that kind of thing.
But then also there's the concern that if the ball
encounters something like a valley, for example, and then gets
to a point where it can't be blown further, it
could get stuck very quickly. That's something that we don't

(40:06):
have to worry about with these rovers, where we can
actually maneuver them around or out of natural landscapes that
would be otherwise um impossible to travel through. Right, I'm
picturing just an entire army of these of these beach
ball robots stuck in a single trench on Mars, and
just think there'd be a cartoon of a Martian walking

(40:26):
by saying, who the hell did this? Well, I want
to talk about another spherical design, Please do So this
other one I also talked about this in my my
blog post about these, it's the Joel Bot. So this
was the one that could actually jump right. Yeah, so
it's still spherical. This was designed, by the way, in

(40:46):
two thousand eight by PhD student named Rhodri ar Moore
from the University of bath uh And so imagine sort
of a spherical cage. It's just it's just an outline, right,
So yeah, think of think of your your beach ball,
where you have the seams. There's a skeleton of a beach, right,
there's nothing in between those those seams. It's just those

(41:07):
are those are bands that that form a ball like shape.
So so you're not going to have the same breeze
driven technology that that beach ball example would um. But
if if you get tension into those cage arms, you
can allow it to um, to roll purposefully and also hop. Yeah,
so imagine through the middle of this spherical cage, you've

(41:28):
got to you've got a rod that can compress like
a spring, and then once you compress it far enough,
it builds up potential energy to release suddenly and do
a hopping action like a grasshopper. And so that could
come in really handy. You can imagine if it were
to get stuck or if it were to need to
get over an obstacle. Yeah, there's something about this particular

(41:50):
design that bothers me, but I'm going to save that
because it actually applies to another robot on our list.
In that way, I can just cover them all at once. Okay, Uh,
there is something that I think is really cool. This
is probably what you're talking about here, which is the
robot based on the tin Segrety architecture. Yeah, it's called
the super ball bot, and it has been under active

(42:12):
research at NASA's Innovative Advanced Concepts Program. That's the nayak Niak.
It's like, it's like Brainiac without the brains. True, So
so i'll call it. I'll call it Brainiac. So this
isn't Superman's enemy, it would just be Man's logic. Yeah. Uh,

(42:32):
well be your Man's fair enough. Yeah. So it's under
the direction of a couple of guys named Vitas sun Spiral.
I believe I'm pronouncing that correctly. Sun Spiral does seem
to be the way it's spelled, and Adrian Ego Geno.
It's based on this tin Segrety architecture. So imagine it's
a roughly spherical shape, very roughly. Instead of seeing a

(42:57):
spherical skin or a sphereical cage, imagine you're looking at
like like a tangle of Christmas lights, except not really.
It's it's more like poles and cables. So you've got
these rods and then you've got these strings coming out
the ends of them, and just staring at it sitting

(43:17):
on the ground, it might look like some collapsed mess
of just knodded up components. But once it starts to move,
you realize that these cables connecting the ends of the
rods can relax and contract, kind of like the muscles
that control your skeleton very much. So, yeah, and so
what this thing can do is it can buy this

(43:38):
contracting and relaxing action, move itself along on the ground.
It can cause itself to tumble forward on the landscape.
It can also raise itself up a little bit by
creating that tension. So you know, it's just imagine any
kind of of structure that relies on on cables and rods,
like a tent. That's an easy one to imagine. It's

(44:00):
using tension to maintain that stability and really saying it's
like our our muscles and bones is is pretty accurate
if you think about it. Like when I when I
tried to describe this to someone, I said, all right,
we'll just imagine you have a mass of bones and
muscle tissue that's all attached together and it's still able
to relax and contract and move around that way. And
they said, that's awful. I'm like, well, you need to

(44:22):
watch these videos because even when they tend to be
really sped up, the videos are are sped up quite
a bit because this thing moves very slowly right now, Yeah,
they're still they're still learning to program its motion. And
the way they're learning to program its motion, I think
it's really cool. They're using an evolutionary simulator, so they're
basically pitting UH control schemes against each other in a

(44:45):
survival of the fittest simulated by the computer, and whichever
one ends up coming out on top is the one
they go with. So system like this can absorb shock,
and that's one big advantage. You you can drop it
on the surface of the planet without really worrying that's
gonna break. Like if you just were to drop a
rover on Mars that had wheels, I'm sure it would
just shatter. So it absorbs shock and it can fold

(45:12):
up for transport. This is a really interesting thing. It
can just kind of like fold itself into a neat
little triangle and uh, and can be deployed easily that way.
It's lightweight, it's structurally robust, and it is capable of
some very versatile motion potentially depending on how they manage
the control scheme. Yeah, the thing that concerns me, and

(45:35):
this was the other one that that made me think
about my concern so for both this one and our
cage bought friend. Uh, the issue I have is that
I can imagine a system like this getting hung up
on protrusions. So think of a really rocky surface and
this thing is trying to get through. Like if you've

(45:56):
ever warned anything that has any loops on it whatsoever,
I guarantee it's some point you've walked past something and
gotten hooked on it, and then you have to stop
and you have to dislodge it and continue on your way. Um,
or you just keep pulling, or you just keep pulling
tell rips. If you're Joe and you're just Yolo, you
just keep going, but for most of us we just

(46:17):
you know it's I said to Joe earlier before we
did the podcast, I said, I'm one of those people
who if I have something that's got loops on it,
it's gonna catch on something, probably sooner rather than later.
If I were to ever actually attempt to catch something
with a loop, I would never ever do it. Well,
hopefully Murphy's law would apply less to this robot than
it would to you, one would hope. But I'm thinking about,

(46:39):
you know, a really chaotic landscape. You could have a
lot of potential for getting especially you're you're talking without
lots of open space and then cables, so there's a
lot of areas where cables could get hooked on things.
And I could just imagine that that would end up
causing a lot of uh frustration down the line. Not
for the robot it doesn't care, but for the engineers

(46:59):
who unless we program it to care less and feel
bad about it, right like way kuld you move stupid um.
But yeah, I mean the idea that you would have
to waste time and energy to free it from whatever
it happened to get caught on. That's the to me,
that's the biggest potential um problem here is that this

(47:20):
it has the potential of getting hooked up on lots
of different stuff. I can see that, though I also
at the same time think, well, mostly what it would
be encountering I'd imagine would be like sand and then
some rocks, and there might be like some really craggy rocks.
I can see it getting hung up on, but in general,
it's not going to be running into like deer antlers

(47:40):
or protruding nails or yeah, yeah, you guys just don't
hang out in the right parts of marrows. Apparently not
think yeah, yeah, the antler valley is not the one
that I visit all that it's too it's too hell,
he's over from that one that's filled with all the
beach balls. Is that what you keep getting your loops
caught on, Jonathan is an Yeah, No, it does happen

(48:02):
more frequently than I care to admit. Al Right, guys,
do we have any other examples? Well, one is something
that I can't find in any way has been mentioned
with reference to space, so it may be completely not Germane.
But it's so cool that I just wanted to bring
it up as a possibility when we're talking about a
whole body rolling robots like these spherical robots. Is the

(48:26):
More Facts, which is a transforming hexapod robot built by
a Norwegian guy named Cary Haliverson. Yep, this one's a
really cool one. It looks like when it's when it's
completely all the limbs are pulled in, it looks like
a ball. Yeah, you know, it looks like it looks
like kind of a semi opaque whitish ball. Uh and

(48:49):
then and then things start moving. Yeah. It So it's
a spherical hybrid robot with twelve curved triangular panels. So
imagine a globe and then draw lines through it, separating
it into six curve triangles on top and six curved
triangles on the bottom, slices of pie. And now imagine

(49:10):
these little slices of pie can unfurl and turn into
the legs fold out and turn into legs like little
crab legs. So this is a robotic D twelve that's
kind of kind of except it doesn't have flat surfaces.
It's it's actually round on the outside, so it's more
like a D one hundred. If you've ever seen one
of those, you know, not the D tens. Okay, Uh,

(49:34):
it really saves you time when you have to roll percentages.
I'm just saying. So this this robot has a couple
of different ways of creating locomotion. One is that it
can unfold its legs, and it does end up having
six legs that way, the top six panels remain kind
of closed in, although they can't open up a little
bit too, but the bottom six panels open up to

(49:56):
be legs, and then it moves around kind of like
a crab. Like it's very crab, like a plastic crab.
That sounds amazing and creepy. Now, if it folds up
into a ball, it can actually just flex the legs
a little bit, pushing the panels outward and thus rolling
the ball forward, so it can either roll or crawl. Yeah,
And so what this actually reminded me of was one

(50:18):
of the rovers, one of the actual rovers we talked
about earlier athlete, which combined the leg action and the
rolling action, except in that case it was wheels on
the end of legs. Right, This would be a combination
of climbing leg abilities with whole body rolling abilities. So
you can easily imagine like using something like this to
climb up a hill, for example, to get a look

(50:42):
around the hill and see where does it want to
go from there, and then doing a controlled roll down
the other side of the hill, because that would be
far more efficient than climbing back down. You know, It's
it's not something again, like Joe was saying, it's not
something that is specifically being hailed as a future in
in Rover robot addicts, but it was it's something that
sparked our our imagination. Yeah, well, we're just saying, take

(51:05):
a look at this thing. Get a load of this
thing over here. It's pretty interesting and slightly less terrifying
than athlete was. I think athlete is perfectly cuddly. Do
you I would wrap my arms around one of those legs.
We're gonna make you watch the return to oz uh
the sequence with the wheelies in a minute, and then

(51:27):
you know or wheelers I guess is what they're called.
Oh boy, those are scary. Well anyway, that's that's generally
the types of locomotion that we wanted to look at.
There are other options as well. In fact, I've seen
some interesting robotics with that end up emulating the movements
of snakes or fish, Yeah, so there are definitely other

(51:48):
other means of locomotion, but these are the ones we
think are probably the most likely. In fact, I think
I'll be amazed if anything overtakes the wheel anytime really soon.
But it's it's fun to kind of stretch our minds
in this way. And if we ever get to a
point where some of these considerations are no longer as important,
things like you have to worry about how much power

(52:10):
the locomotion is going to consume, then who knows, I mean,
then you then you really open up the door. And
now I don't see that changing anytime in the near future,
but it's also a possibility. So I like, like Joe
said the very top of this podcast, this is just
scratching the surface even of robotic locomotion. And uh, then
there are all those other elements that are related back

(52:31):
to that DARPA challenge I mentioned at the very beginning
that we could go on and on about and we'll
probably do more episodes about some of the big problems
in robotics today and how people are going about addressing
that and and meeting those challenges. I think there there's
plenty of opportunity to talk about that but if there's
something about the future you want to hear about specifically,

(52:52):
and maybe it has nothing to do with robots at all,
let us know. Send us a message our email addresses
FW thinking at this governy dot com, or drop us
a line on Facebook, Twitter or Google Plus. You can
find us with the handle FW thinking at all three
and we will talk to you again really soon. For

(53:15):
more on this topic and the future of technology, visit
forward thinking dot com, brought to you by Toyota. Let's
Go Places