March 20, 2024 • 29 mins
In the words of the NASCAR legend Richard Childress, “Once you’ve raced, you never forget it…and you never get over it.” This is the most fitting quote to explain the near obsession so many students and engineers competing in the IAC develop. The intoxicating call to the racetrack easily proves there is a racing driver in all of us.
Today, on the inside track, we sit down with team leader Stephanie Meyer of Autonomous Tiger Racing at Auburn University to explore this idea. At first, she didn’t feel the pull of racing in the ways that more traditional gearheads would, but as we will find out, working on the Luna car sparked the interest that has kept her on the track. Later, she discusses how to overcome the various levels of anxiety that come with sending out a car to race, even if you’re not sure the car is ready to compete.
Today, on the inside track, we sit down with team leader Stephanie Meyer of Autonomous Tiger Racing at Auburn University to explore this idea. At first, she didn’t feel the pull of racing in the ways that more traditional gearheads would, but as we will find out, working on the Luna car sparked the interest that has kept her on the track. Later, she discusses how to overcome the various levels of anxiety that come with sending out a car to race, even if you’re not sure the car is ready to compete.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
I'm finding it very difficult to leave the racing because
there isn't a professional racing equivalent to go into just yet.
Speaker 2 (00:06):
And I love my car.
Speaker 1 (00:08):
I can't leave my car. You know, her name is Nova.
She is very much so my baby as well as
the baby of my other team members, and I just
I can't imagine walking away, and so it's kept me
at the university for longer.
Speaker 3 (00:26):
The world around us is changing faster than ever before.
Ideas once only imagined in science fiction are becoming a reality.
Throughout the course of our amazing twenty three episode season,
we'll speak to some of the greatest minds in robotics
and artificial intelligence to discuss the groundbreaking work that's fueling
it all. I'm your host Ryan Marine, joined me and
my co host Paul Mitchell, the presidents of the Indian
(00:48):
Autonomous Challenge and see why we call this the Inside Track. Today,
on the Inside Track, we sit down with team leader
Stephanie Meyer of Autonomous Tiger Racing at Auburn University to
explore this idea. At first, she didn't feel the poll
of racing in ways that more traditional gearheads would, but
as we will find out, working on the Nova car
(01:09):
sparked the interest that has kept her on the track. Later,
she discusses how to overcome the various levels of anxiety
that come with sending out a race car to compete,
even if you're not sure the car is ready. So
it started at the beginning. What brought you to this project?
What drew you to this team there at Auburn University.
Speaker 1 (01:33):
Yeah, that's an excellent question. So I actually got started
into robotics several years ago in my undergrad I was
never a big racing person, not even a big car person,
so I never would have assumed that I was going
to be working with these high speed cars, let alone
leading the team one day. But yeah, I fell in
love with the robotics competitions. Back in my undergrad I
did what we call asme design competitions. Back in the day,
(01:56):
they used to be things like designing concrete beams to
hold a certain amount of way. Luckily these days they're
more robotics based. And then I got to know a
bunch of grad students through that project, and one of
them said, there's an excellent robotics grad student lap down
in Auburn, Alabama. Let's go down there. They'll take you
and give you a stipend and you get to work
on robots for a few more years. And I said,
(02:17):
industry robots, industry robots, let's go through the robots grad degree.
And then a few years into that, after working with
just general autonomous vehicles and robots, one of my colleagues
in the research lab actually found the autonomous competition here
and he said, hey, you're a perception expert, a laser
censor expert. We need one of those. You want to
join the team. And again, at this point, I'm not
(02:38):
very into racing. I was like, well, I kind of
just want to do normal robotics, but I'll help you
a little bit. And then next thing you know, I'm
doing a lot of the work and like just dead
into it along with the rest of them.
Speaker 3 (02:48):
So the hook was said. At that point, hook was.
Speaker 2 (02:50):
Said, it's amazing.
Speaker 3 (02:51):
What is it about robotics? Why was that something that
grabbed your interest?
Speaker 1 (02:56):
Right, that's a really good question as well. So I
started off just really being interested in programming. I think
what really got me about programming is how systematic it is.
It's really easy just to organize all of your thoughts,
but then you can also double check your thoughts because
you run the compiler, you run the program, you look
for the errors that have happened. And that's all well
and good when you see the numbers and the letters
(03:17):
on the screen. But it's even more amazing when you
see something move because of it.
Speaker 2 (03:20):
And that's robotics.
Speaker 1 (03:21):
That's the logic, the code that actually makes something happen.
Speaker 3 (03:25):
What was the first robotics project that you worked on.
Speaker 1 (03:28):
The first one I was adjacent to was an obstacle
course robot. It had to climb let's see, it went
through a water pit, a sand pit so it could
collect up all the sand after it got wet, and
then with sandy tires, it had to climb three stairs
and deliver a resource package. So this was rescue robots
that delivered like a package of rice or beans into
(03:50):
a hole at the top of the stairs, and then
it had to go back. I was really more watching
all of the more experienced engineers working on this project,
but I definitely caught the competition bug with that. The
first one that I was actually hands on with was
more of a manufacturing robot. This one was take a
flat sheet of paper, have the robot turn it into
some sort of projectile like a paper airplane, and then
(04:11):
test it, and then you got extra points for your
robot if it could shoot farther than the rest of them.
Speaker 3 (04:15):
Oh, that's really neat, and then you have to probably
decide what's the best projectile.
Speaker 1 (04:19):
Right exactly, it was completely open and of course folding
a paper airplane probably gives you the biggest distance, but
that's also going to be more complex machinery. We ended
up just having the paper rolled into a tube and
then crushed, so we make this really compact puck. And
then if you've ever played with hot wheel tracks, you
know they have the wheels spinning like this that your
(04:39):
car gets shot through.
Speaker 2 (04:40):
Yes, we just shot the puck through that. It worked
pretty well.
Speaker 3 (04:43):
My kids love that, so that's right in my wheelhouse definitely. Now,
all those sound like awesome robots, right, but none of
them are million plus dollar race cars exactly.
Speaker 2 (04:54):
So when you first.
Speaker 3 (04:55):
Saw this Indy Autonomous Challenge car and understood what it
was that you were going to be working with, what
was your reaction?
Speaker 1 (05:02):
Honestly, it's a bit intimidating. It's one thing to take
a few pieces of plywood and stick them together into
a frame for your pup crushing robot. It's another thing
to work with such precise engineering. You don't want to
mess this up and let alone, we're running it a
whole lot faster. There's not a whole lot of risk
with the puck robot or with the stair climbing robot.
If it falls off the track, it's fine. If we
(05:23):
fall off the track, we're not nearly as fine and
as very expensive. So the stakes are so much higher here,
but that does push you to be more precise in
the work.
Speaker 3 (05:31):
Sure, so you get brought in initially just to help
with your one area of expertise, but now it's grown
to the point where this is your team. You are
leading this team of students that are working on this project.
Speaker 2 (05:45):
Exactly take us through that process.
Speaker 1 (05:47):
So these cars have a perception stack that includes cameras
which everyone knows what those are, radar, which most people
are pretty familiar with, and then ldar, which are sort
of a laser version of a radar. Both radars and
light are sending out electromagnetic pulses into the world just
to see where everything is around it, and the light
ar gives you a lot denser view that's a lot
more similar to a camera with all of its pixels
(06:09):
than the radar does. So light oar is my area
of expertise sending out. It sends out those pulses, gives
you a really dense scan of locations of things in
your environment, and then you need to process that data
to determine out of all of these points of locations,
what's a car, what's a wall, what's the track. So
I'm here just to do that, And I started off
getting invited up to Indianapolis at the very first dev
(06:32):
whenever we had these car physically physical cars being put together.
All of the teams came together in Indianapolis and we
spent several months in twenty twenty one just all of
us living in the area getting to know each other,
coming into work in the garage every day, first on
just a single car, and then as the other one
started becoming available, splitting off more into our individual individual regions.
(06:53):
And so I got to see the entire process, you know,
while I'm up here just hanging out with the team,
hanging out with the other teams unlight our drivers getting
ready for the actual light our work. And during the
first season, for the first several months, for me, it
was just business as usual, just writing code, just testing
it on the computer. It's just research. I've done this
(07:14):
for years. And then that moment when Will Bri and
our team lead at the time, came up to me
and said, Stephanie, you're up. We have to pass another
car in a few days. I'm like, this is my
software all of a sudden, because until we had to
interact with other vehicles, it was just the controller, it
was just a localization in the GPS, And now my
perception has to actually see these other cars in order
(07:36):
for us to pass them. If my software fails, we
could very likely crash into this vehicle and not only
hurt our car but the other person's car, which.
Speaker 2 (07:44):
Would just be a disaster.
Speaker 1 (07:45):
Nobody's crashed yet at this point. Ours could be the
first because of my software. And that it was terrifying,
And in that moment I was like, I'm not ready.
The software is not ready. Will I need another month,
But you don't get another month. It's going to happen
in three days, whether you're emotionally prepared for it or not.
Speaker 3 (08:02):
It has to be like feeling like a parent or
something right you've taught your child, in this case, your
your robotic race car, all these lessons and now you
have to see if those lessons actually sunk in. What
was it like when it actually accomplished the goal that
you set out to accomplish.
Speaker 2 (08:17):
Oh, it was.
Speaker 1 (08:18):
It was some kind of high, for sure, because I've
got all this nerves just really built up, and I'm
expecting it to fail the moment we passed the other
car for the very first Really, yeah, I'm straight up like, oh,
this one error I saw that one time that I
thought I fixed is going to come back.
Speaker 2 (08:31):
Isn't it.
Speaker 1 (08:31):
We're just going to drop that point and it's going
to slam right into it.
Speaker 2 (08:34):
Oh, I'm terrified.
Speaker 1 (08:35):
But then it does it, and it does it another lap,
and it passes another lap and it's fine. And afterward
I'm like, oh my goodness, it worked. I can't believe
it actually worked, because no matter how much testing you do,
you know you're always going to doubt your masterpiece.
Speaker 2 (08:47):
Yeah, you don't know until it's put to the test.
Speaker 1 (08:49):
Right exactly, But just that immense joy when it does work.
It's incredible.
Speaker 3 (08:54):
Going back to the sensor, something I found really interesting.
You have the three the camera, the radar, the line,
each of them are great at certain things, maybe not
so good at other things, and it's finding a way
to mesh these all together. And it makes me think
of what the human driver deals with, whether it's on
the racetrack or just on the street. You're perceiving with
your eyes, your ears. There's a sense of feel as
(09:16):
well that goes into that, but you have to find
a way to turn that into computer terms, right, right,
in a way that your car is going to understand.
Speaker 2 (09:25):
That's right.
Speaker 1 (09:27):
That reminds me of a very interesting topic that keeps
coming up as well. People ask us, have you talked
with actual race car drivers? Have you talked with engineering
teams that help these drivers analyze the same sorts of
data that we're looking at out on the track, And yeah,
we did. We talked with a lot of them back
before we even saw any of these cars, before we
placed our hands on them, and we tried to make
(09:47):
it useful. We tried to take the information about the
wind's going to change how your vehicle controls, and you
have to feel how the cars moving and then anticipate
how you need to change the steering so you keep
the slip and take that turn in this pitarticular way
on the ims, because that's the way to get around
your opponents and really take it well. And we're jotting
down notes and memorizing everything that they say.
Speaker 2 (10:10):
At the beginning, we.
Speaker 1 (10:11):
Actually did not know how to translate that into machine terms.
Because that's exactly what we've come to the conclusion of
answering these questions and thinking back retrospectively is our job
now is to figure out how to translate that. And
we didn't know at first. That's the project, that was
the learning that we had to do. And I think
we're definitely more at the point where we can look
(10:33):
back at that advice and say, I know exactly how
that fits into my system. That is a tweak on
this controller gain that is utilizing this sensor better, and
it's really cool to see that growth over the few
years we've been doing this.
Speaker 3 (10:44):
The thing I've taken away being around this project over
the last few years that I was not expecting is,
as you talk about, you are breaking down what a
human driver is doing to its component parts finding a
way then to translate that into the machine language. And
that to me, when you start to see what mean,
(11:05):
what's set a human race car driver to one side,
even the human driver on the street. Sure, the amount
of information that you have to process to do something
that we do every day, and we don't think any
more about when you do break it down in the
way that you do. All of a sudden, I came
away with this greater appreciation for the task that the
human brain is able to accomplish.
Speaker 2 (11:25):
Oh, of course, yeah.
Speaker 1 (11:26):
The human brain is the most incredible computer we have,
and we don't have a compute system that's even close
to the processing power of the human brain yet, which
kind of brings us to another part of the autonomy problem.
Speaker 2 (11:36):
We need to simplify what the human brain does.
Speaker 1 (11:38):
We need to find what are the most essential parts
of that program that's going on up here, and that
has been one of the keys in the racetrack because
we have such limited test time because the tracks are
high demand and incredibly expensive, and we're also sharing this
track time with the eight other teams, so we need
to be very good at simplifying our system down to
(11:59):
a way that we can work with it in this
type timeframe. We need to be able to see exactly
where the problem was when the problem occurs, and then
adjust the code in the hour between our next shot
up and thirty minutes out on track. So simplification of
what the human brain.
Speaker 3 (12:12):
Does without giving away industry secrets. Can you give us
a couple bullet points of when you do boil it
down to its most basic components, what kind of things
are you identifying?
Speaker 1 (12:23):
Let me think about that for just a moment, from
an incredibly academic standpoint, like thinking about all the complicated
research algorithms that we read papers on and we're trying
to compete with and build upon the cutting edge of
the research when we do publish. As an academic, we
started out with a lot of those ideas, let's have
(12:43):
incredibly complicated control logic cutting edge, and then we find
that that makes it much more difficult to adjust our
system during test days. It makes our system less flexible.
If it's tuned in perfectly, if the assumptions that you
make about the weather of the day are perfect, if
your censors are giving you quality enough data for you
(13:05):
to run this really high edge algorithm, theoretically you'd get
the better result out of it, But on average, we
need to simplify down to something that's been published several times,
and you wouldn't be able to get a paper out
about it anymore.
Speaker 2 (13:18):
So we've boiled it down to.
Speaker 1 (13:22):
Yeah, we've basically taken it from the complicated research level
algorithm down to just textbook level. Yeah.
Speaker 3 (13:29):
Well, fair enough, And I know you don't want to
give away anything. This is incredibly intense competition between teams,
but also there is a great deal of camaraderie. That's
the thing I've noticed is well being around the teams.
As much as each team wants to win, and they
clearly do, there is an element where it seems like
everyone's in this together and they're all trying to paddle
(13:50):
in the same.
Speaker 1 (13:51):
Direction, of course, and I think part of that comes
from those early days back in twenty one when we
just had the one car and we needed to.
Speaker 2 (13:57):
Figure out to turn it on reliably.
Speaker 1 (13:59):
We need to figure out how we get all the
base communication software going before you even start putting your
competition stack on top of it.
Speaker 2 (14:05):
All of that was done collaboratively. All of the.
Speaker 1 (14:08):
Teams were there together and learning these cars for the
first time together, and I think a lot of that
has stuck through the years. It's one of my favorite
parts of the competition. I may not have fallen in
love with it as much without that sense of community,
because I do like to come back and see all
of my friends on the other team, and I'm sad
when they graduate and they leave and go off to
big careers, and it's great to see the same faces
(14:29):
that have stuck around.
Speaker 3 (14:30):
You did fall in love with it, though, and that's
actually what I'm really interested in here too, because you
came at this from an interest in robotics, not necessarily
in motorsports or in competition of this kind. So how
quickly did you start to build that appreciation for what
you were about to do and get that feeling of
infatuation with this project.
Speaker 1 (14:51):
It was probably a slow boil over those months of
being together, and I don't think I realized it until
it was almost over. It was almost October. I think
it was twenty third when we had that first competition.
I started to realize I might not see some of
these people again, the people from Munich and from Milan
and other places around Europe. It's possible I could have
(15:12):
never seen those guys again. And we had built up
a decent friendship, you know, a community bond, and so
I was like, oh, I need to make sure I'm
trading communications and spending time with them now at the end,
and just that a little bit of fear of not
having that experience anymore, and then the joy whenever the
competition did actually continue on the further seasons, as like
(15:33):
I have to come back, I have to keep.
Speaker 2 (15:34):
Being a part of it.
Speaker 3 (15:35):
Well, you're right, at that time, we didn't know what
the future held beyond that first event at Indianapolis. But
since then there have been several further competitions, and the
way that this has progressed the rapidity with which it
has progressed, the complexity of the task at hand, going
from time trials with a minimal amount of obstacle avoidance
to a true passing competition where you're sharing the track
(15:58):
with another entity that you don't have the data on.
You don't know what that car is going to do.
You have to have a car that's flexible enough in
its own algorithm to be able to adapt to that.
Just how much more challenging has this proposition become as
we get deeper and deeper into the in the autonomous challenge.
Speaker 1 (16:18):
I think because I've been with it since the beginning,
so I've been growing kind of along with it. Sure
it feels a lot like the same level of complexity
from the beginning, because it was a struggle back then
just to get everything started. And now that we have
that baseline, I don't think about the baseline stuff working
anymore because that's all good and now it's just that
same level of extra work to hit the next level.
(16:40):
So to me, it doesn't feel like it's more complicated,
even though I do. If I take a second and
think about it, I realize we are achieving pretty incredible stuff.
We certainly couldn't have pulled off what we're doing these
days with the road courses and the passing competitions.
Speaker 3 (16:53):
One thing I'm always interested in when I try and
describe this to other people who work in the motorsports industry.
It takes a little bit to kind of get them
on board, but then they start to become interested. Are
you feeling or do you experience the same thing? And
what is it like when you try and explain to
friends at home or family members or whatever, what it
is that you're working on. What is that process like
(17:15):
to try and get them to wrap their head around
exactly the level of complexity and the level of cool
that you're working with.
Speaker 1 (17:22):
Sure, yeah, that makes me think of all of the
times that we've been lucky enough to show the car
off to the general public, and we've taken it to
a few racing events where actual indie races are going on.
We got to speak to people who are fans of
traditional racing, and it definitely comes up with the question, Oh,
what's going on with this car?
Speaker 2 (17:41):
Is it a robot? You know?
Speaker 1 (17:42):
Is it remote control? Everyone thinks it's remote control. We
have to explain what autonomy is and then they say, oh,
that's taking all the fun out of it. That's the
most common thing we get after what is it? I
don't understand? And then that's no fun? You know?
Speaker 2 (17:56):
Who am I going to root? For sure?
Speaker 1 (17:58):
And So what I've been trying to do in response
to those sorts of questions is I fell in love
with it, as we've mentioned several times. So how can
I get that same feeling to be shared out with
the audience? How do I make them see it the
way that I do? And so I try to talk
with these people who are asking us these questions about
how much work it is for these students behind it.
Don't think of the car as an individual by itself,
(18:22):
think of all of the people who are putting their
passion and effort into it. M hm.
Speaker 3 (18:26):
And do you find that that resonates, that message resonates
what the folks you're talking with.
Speaker 2 (18:30):
I think it does a lot of people.
Speaker 1 (18:33):
A lot of people approach the car, our particular car
because it has the big Auburn University symbol on it,
and then they just walk away with these smiles on
their faces about this is super cool that my university
is doing this, that Auburn kids are doing this.
Speaker 3 (18:46):
I think that actually is one of the neat hooks
of this competition. It's not like it's the automotive industry
that it's millions and millions of dollars that's out here
doing this. It is college students, undergrad but mostly graduate students.
But these institutions that are pursuing this, why do you
think this makes sense for an institution like Auburn to
be involved in a competition like this one and not
(19:09):
just once, but ton till through that investment now for
many years.
Speaker 1 (19:13):
It fits in very nicely with the research that my
graduate lab was doing already. To me, it feels like
a very similar project. Of course, it's not the same
private sort of bringing in a bunch of money. It's
really more public spending some money, but we get to
do a lot of hands on work, which is one
of the deficits of graduate labs in general, as it
tends to be more theoretically focused. Sure, another thing is
(19:36):
the longevity of it, the repetition, or how it continues
to grow, because other projects that we have tend to
just have a very short time span. You get done
some very well defined singular problem, you deliver the results,
and then you move on to the next part of
either your career or the next project for the grad lab.
Whereas this project is able to build upon itself. We
(19:56):
have something that starts from just a little seed and
then grows and you can see it becoming something better
and better as you continue to work on it.
Speaker 3 (20:03):
Yeah, it's clearly at the cutting edge. I mean, you
just have to look at the car and that message
sinks in. I want to go back to something you
said earlier. You mentioned when you got into this you
didn't have a particular interest in cars. You didn't have
a particular interest in racing. But if I'm reading between
the lines correctly, perhaps your perception changed a little bit
after dealing with.
Speaker 1 (20:23):
This right, I'm definitely getting a lot more interested. I
don't know a lot of the driver names, and a
lot of the terminology from the different leagues is still
a little.
Speaker 2 (20:31):
Bit beyond me. Well, you're not alone in that fair, very.
Speaker 1 (20:34):
Fair, but just being around all of the other people,
because there's plenty of other people involved in this who
were avid race fans. You are avid I'm going to
tear my car apart and rebuild it type people, and
it's just been fantastic learning from them and getting to
fall into that same interest that they have, seeing how
it can be very interesting where I wouldn't have thought
of it if I weren't in this environment.
Speaker 3 (20:54):
Sure, well, your story is fascinating because you have been
there from the ground level. You've been there as this
project has built to the point now that you're leading
your own team. So what's next for you? How do
you use this as a jumping off point to accomplish
the goals and dreams that you've set off for yourself.
Speaker 1 (21:12):
That's an excellent question, and I think it is a
wonderful jumping off point. There's been so many different industry professionals,
either from the sponsors of the competition, or just any
industry that's just interested in coming to the races. Who
have been talking with myself and a lot of the
other students, and they're always trying to hire us. They say,
what are you doing next? When are you graduating? So
(21:32):
it'd be really easy to go from this and get
just amazing research or technical career for myself. I'm finding
it very difficult to leave the racing because there isn't
a professional racing equivalent to go into just yet.
Speaker 2 (21:45):
And I love my car.
Speaker 1 (21:46):
I can't leave my car. You know, her name is
Nova Miss she is very much so my baby, as
well as the baby of my other team members, and
I just I can't imagine walking away. And so it's
kept me at the university for longer, which is another
thing that the university appreciates.
Speaker 2 (22:02):
I'm sure they do.
Speaker 1 (22:03):
Master students stick around to PhD just to get more
time with the car.
Speaker 3 (22:07):
We've talked a lot about how this program has helped
you develop, and that's a worthwhile end unto itself. But
there's also of course going to be some trickle down
from this that could potentially see some of this technology,
some of the research that you're doing enter into our
daily lives from an AI perspective, whether that's in automotive
(22:28):
or elsewhere. So, having been hands on with this now
for several years, what do you see as the most
likely outcomes in terms of some of this AI entering
our day to day lives here in the short term future.
Speaker 1 (22:42):
Sure, I think one of the most immediate benefits to
the consumer market is going to be how we've been
working with this technology. So a lot of this technology
hasn't either been put in on automobile yet or it
hasn't been put in such extreme conditions. And so we're
testing these sensors for the sense building companies, We're testing
the compute units and in that sort of collaboration with
(23:05):
these sponsors, making a better unit, making something that is
more suited for this sort of environment. And those products
are going to become publicly available. They're going to go
into the OEA manufacturers of Autonomous and aid ass which
is assisted of vehicles. So I think that's going to
be the most immediate is the sensor improvements, the compute
improvements that are being tested out in these vehicles. And
(23:27):
then probably the next one up is going to be
the people who.
Speaker 2 (23:30):
Do leave this and go into industry.
Speaker 1 (23:32):
They're going to be taking all of the learning that
they've done, all of their experiences and their research ideas.
Speaker 3 (23:39):
AI is a huge talking point just in society today.
There's a lot of hand ringing about the effects of
AI for good or for ill. Sure, when you see
things like this written, or when you talk to the
family members, they maybe don't work in this field, but
they probably have opinions on it. First of all, what
are you hearing and what are your opinions on the
(24:00):
potential positive and negative impacts that AI might have in
our near term future.
Speaker 1 (24:06):
Sure, as far as true artificial intelligence, which is a
very statistically learned model, we basically look at it as
a black box. You put a bunch of information in,
it figures out how to process that information, and then
it spits a result out. Whenever I talk to people
about AI, I expect this doomsday opinion. I expect hesitance.
But surprisingly, even from the youngest person to the oldest
(24:29):
member of society, they come up again and again thinking
this is really neat. Look at how much you can
accomplish using this AI. Particularly thinking of some of the
language processing models that are out there today. There are
so many people that just you sit them down with it,
they start to play with it, and then they're like, Oh,
I can use this to make this easier, I can
design my recipes, I can figure this out. And it's
(24:51):
just it's really encouraging to see across a wide swath
of the society that people are seeing the good in it,
seeing the good that it can do.
Speaker 3 (25:00):
Is it something that lends itself to getting hands on
with it, even at the most fundamental level, to start
then to grasp what the positive outcomes might be.
Speaker 2 (25:09):
I think so yeah.
Speaker 1 (25:11):
Letting people just access some sort of an AI model
and play with it, I think does make it more familiar,
makes it more personal, and can take away some of
that fear level.
Speaker 3 (25:20):
It does feel like we're at the cutting edge of
something here in society with AI starting to become integrated,
and if you look at historical long lens it does
seem like typically when new technology starts to enter into
a culture, especially revolutionary new technology, there can be some
negative impacts. So with that said, we've talked about the
(25:45):
benign aspects of AI, but how about the malignant aspects?
What things have you concerned as we look into the
short term.
Speaker 1 (25:52):
Sure, and we can look into the automotive industry to
kind of think about that a little bit. The automotive
industry tends not to trust AI by its because these
are black boxes. It's very difficult to guarantee that they're
going to give you the safe result that you need.
It's very difficult to analyze what they're even doing to
process this data, and so we tend to like to
fuse it with more classical type approaches. This is where
(26:14):
you take a scientist or an engineer and have them
actually explicitly put their ideas down into how you should
manipulate the input data into the output data. They are
directly causing that program to give a result. You fuse
the intentional programming with the statistical programming in order to
get the high performance of the statistical programming. But the
(26:34):
guarantees and the safety and the control of the classical
programming really interesting.
Speaker 3 (26:39):
As always, anything we haven't touched on yet that you
think people would like to know for those who maybe
are coming to this that are new they've not been
following the end Autonomous challenge, or even for folks who
know a lot about it. But is there something that
you've taken away from this that you'd like to make
sure others are aware of.
Speaker 1 (26:57):
Let me think about that for a moment, because I've
thought a bit about.
Speaker 2 (27:01):
Women and engineering. Oh yeah, sure, quite a bit.
Speaker 1 (27:04):
Because I've been approached by people and they're like, are
you with the racing. I'm like, I'm leading the team,
and they're always amazed, which is a good thing to
a degree, But I think that there shouldn't be that amazement,
you know, it should be normalized. So something as visible
as this is really gives a great opportunity for women
in engineering, women in motorsports, and women and intelligent fields
(27:27):
in general. You know, people who are capable of doing things.
It gives exposure to us, and it also normalizes it
because we don't want people to be amazed that there's
a woman doing this engineering. We want to be people
to be amazed that people can do this engineering.
Speaker 3 (27:42):
Do you find do you get a reaction from young people,
young women, young girls in particular, when they find out
what you're doing, and do you feel like you're having
an impact?
Speaker 2 (27:52):
I do.
Speaker 1 (27:52):
I've actually had several one on one interactions with younger
females where they'll come in with their dad to see
their car, the car that dad might be a potential sponsor, etc.
We might see them just in the big public open events,
and sometimes there has a tant to approach the car
and they need to be encouraged, or sometimes they'll just
come right up and put their hands on it. And
(28:12):
I tend to try to ask them, do you want
to lead a race team when you're older?
Speaker 2 (28:16):
And like, yeah, say yes, say yes.
Speaker 1 (28:19):
And I like to we do have a remote control mode,
even though these are fully autonomous cars, which is more
for testing and for show, I like to put the
controller in their hand and let them wiggle the wheels
when it's on display, and you know, become hands on.
Don't be afraid to touch the technology, don't be afraid
to ask questions, and to consider that this can be
for you as well. And I like to think that
just having that momentary interaction does change the perspective. It
(28:43):
says you don't have to stay away from it because
you're a girl. This is something girls can do too.
Speaker 3 (28:47):
A wonderful message for sure, A great one to end
on as well. Stepanie, thank you so much for joining us.
I'm so excited to see what the future hold. It's
not just for you, but your entire team as well.
Speaker 2 (28:56):
Yes, thank you so much for talking with me. It's
been fun.
Speaker 3 (29:03):
Thanks for joining us this week. On the inside Track,
that was Stephanie Meyer of the Autonomous Tiger Racing team
at Auburn University. As she so aptly explained, the passion
for racing isn't something that could be measured or predicted.
It's something you just have to experience to understand. I'm
your host, Ryan Marine. Join us next time on the
(29:24):
inside Track, where we speak with Simon Hoffman of the
Tomb Autonomous Motorsports team at the Technical University of Munich
to discuss the more granular aspects of building out the
brains inside their autonomous car.
I'm finding it very difficult to leave the racing because
there isn't a professional racing equivalent to go into just yet.
Speaker 2 (00:06):
And I love my car.
Speaker 1 (00:08):
I can't leave my car. You know, her name is Nova.
She is very much so my baby as well as
the baby of my other team members, and I just
I can't imagine walking away, and so it's kept me
at the university for longer.
Speaker 3 (00:26):
The world around us is changing faster than ever before.
Ideas once only imagined in science fiction are becoming a reality.
Throughout the course of our amazing twenty three episode season,
we'll speak to some of the greatest minds in robotics
and artificial intelligence to discuss the groundbreaking work that's fueling
it all. I'm your host Ryan Marine, joined me and
my co host Paul Mitchell, the presidents of the Indian
(00:48):
Autonomous Challenge and see why we call this the Inside Track. Today,
on the Inside Track, we sit down with team leader
Stephanie Meyer of Autonomous Tiger Racing at Auburn University to
explore this idea. At first, she didn't feel the poll
of racing in ways that more traditional gearheads would, but
as we will find out, working on the Nova car
(01:09):
sparked the interest that has kept her on the track. Later,
she discusses how to overcome the various levels of anxiety
that come with sending out a race car to compete,
even if you're not sure the car is ready. So
it started at the beginning. What brought you to this project?
What drew you to this team there at Auburn University.
Speaker 1 (01:33):
Yeah, that's an excellent question. So I actually got started
into robotics several years ago in my undergrad I was
never a big racing person, not even a big car person,
so I never would have assumed that I was going
to be working with these high speed cars, let alone
leading the team one day. But yeah, I fell in
love with the robotics competitions. Back in my undergrad I
did what we call asme design competitions. Back in the day,
(01:56):
they used to be things like designing concrete beams to
hold a certain amount of way. Luckily these days they're
more robotics based. And then I got to know a
bunch of grad students through that project, and one of
them said, there's an excellent robotics grad student lap down
in Auburn, Alabama. Let's go down there. They'll take you
and give you a stipend and you get to work
on robots for a few more years. And I said,
(02:17):
industry robots, industry robots, let's go through the robots grad degree.
And then a few years into that, after working with
just general autonomous vehicles and robots, one of my colleagues
in the research lab actually found the autonomous competition here
and he said, hey, you're a perception expert, a laser
censor expert. We need one of those. You want to
join the team. And again, at this point, I'm not
(02:38):
very into racing. I was like, well, I kind of
just want to do normal robotics, but I'll help you
a little bit. And then next thing you know, I'm
doing a lot of the work and like just dead
into it along with the rest of them.
Speaker 3 (02:48):
So the hook was said. At that point, hook was.
Speaker 2 (02:50):
Said, it's amazing.
Speaker 3 (02:51):
What is it about robotics? Why was that something that
grabbed your interest?
Speaker 1 (02:56):
Right, that's a really good question as well. So I
started off just really being interested in programming. I think
what really got me about programming is how systematic it is.
It's really easy just to organize all of your thoughts,
but then you can also double check your thoughts because
you run the compiler, you run the program, you look
for the errors that have happened. And that's all well
and good when you see the numbers and the letters
(03:17):
on the screen. But it's even more amazing when you
see something move because of it.
Speaker 2 (03:20):
And that's robotics.
Speaker 1 (03:21):
That's the logic, the code that actually makes something happen.
Speaker 3 (03:25):
What was the first robotics project that you worked on.
Speaker 1 (03:28):
The first one I was adjacent to was an obstacle
course robot. It had to climb let's see, it went
through a water pit, a sand pit so it could
collect up all the sand after it got wet, and
then with sandy tires, it had to climb three stairs
and deliver a resource package. So this was rescue robots
that delivered like a package of rice or beans into
(03:50):
a hole at the top of the stairs, and then
it had to go back. I was really more watching
all of the more experienced engineers working on this project,
but I definitely caught the competition bug with that. The
first one that I was actually hands on with was
more of a manufacturing robot. This one was take a
flat sheet of paper, have the robot turn it into
some sort of projectile like a paper airplane, and then
(04:11):
test it, and then you got extra points for your
robot if it could shoot farther than the rest of them.
Speaker 3 (04:15):
Oh, that's really neat, and then you have to probably
decide what's the best projectile.
Speaker 1 (04:19):
Right exactly, it was completely open and of course folding
a paper airplane probably gives you the biggest distance, but
that's also going to be more complex machinery. We ended
up just having the paper rolled into a tube and
then crushed, so we make this really compact puck. And
then if you've ever played with hot wheel tracks, you
know they have the wheels spinning like this that your
(04:39):
car gets shot through.
Speaker 2 (04:40):
Yes, we just shot the puck through that. It worked
pretty well.
Speaker 3 (04:43):
My kids love that, so that's right in my wheelhouse definitely. Now,
all those sound like awesome robots, right, but none of
them are million plus dollar race cars exactly.
Speaker 2 (04:54):
So when you first.
Speaker 3 (04:55):
Saw this Indy Autonomous Challenge car and understood what it
was that you were going to be working with, what
was your reaction?
Speaker 1 (05:02):
Honestly, it's a bit intimidating. It's one thing to take
a few pieces of plywood and stick them together into
a frame for your pup crushing robot. It's another thing
to work with such precise engineering. You don't want to
mess this up and let alone, we're running it a
whole lot faster. There's not a whole lot of risk
with the puck robot or with the stair climbing robot.
If it falls off the track, it's fine. If we
(05:23):
fall off the track, we're not nearly as fine and
as very expensive. So the stakes are so much higher here,
but that does push you to be more precise in
the work.
Speaker 3 (05:31):
Sure, so you get brought in initially just to help
with your one area of expertise, but now it's grown
to the point where this is your team. You are
leading this team of students that are working on this project.
Speaker 2 (05:45):
Exactly take us through that process.
Speaker 1 (05:47):
So these cars have a perception stack that includes cameras
which everyone knows what those are, radar, which most people
are pretty familiar with, and then ldar, which are sort
of a laser version of a radar. Both radars and
light are sending out electromagnetic pulses into the world just
to see where everything is around it, and the light
ar gives you a lot denser view that's a lot
more similar to a camera with all of its pixels
(06:09):
than the radar does. So light oar is my area
of expertise sending out. It sends out those pulses, gives
you a really dense scan of locations of things in
your environment, and then you need to process that data
to determine out of all of these points of locations,
what's a car, what's a wall, what's the track. So
I'm here just to do that, And I started off
getting invited up to Indianapolis at the very first dev
(06:32):
whenever we had these car physically physical cars being put together.
All of the teams came together in Indianapolis and we
spent several months in twenty twenty one just all of
us living in the area getting to know each other,
coming into work in the garage every day, first on
just a single car, and then as the other one
started becoming available, splitting off more into our individual individual regions.
(06:53):
And so I got to see the entire process, you know,
while I'm up here just hanging out with the team,
hanging out with the other teams unlight our drivers getting
ready for the actual light our work. And during the
first season, for the first several months, for me, it
was just business as usual, just writing code, just testing
it on the computer. It's just research. I've done this
(07:14):
for years. And then that moment when Will Bri and
our team lead at the time, came up to me
and said, Stephanie, you're up. We have to pass another
car in a few days. I'm like, this is my
software all of a sudden, because until we had to
interact with other vehicles, it was just the controller, it
was just a localization in the GPS, And now my
perception has to actually see these other cars in order
(07:36):
for us to pass them. If my software fails, we
could very likely crash into this vehicle and not only
hurt our car but the other person's car, which.
Speaker 2 (07:44):
Would just be a disaster.
Speaker 1 (07:45):
Nobody's crashed yet at this point. Ours could be the
first because of my software. And that it was terrifying,
And in that moment I was like, I'm not ready.
The software is not ready. Will I need another month,
But you don't get another month. It's going to happen
in three days, whether you're emotionally prepared for it or not.
Speaker 3 (08:02):
It has to be like feeling like a parent or
something right you've taught your child, in this case, your
your robotic race car, all these lessons and now you
have to see if those lessons actually sunk in. What
was it like when it actually accomplished the goal that
you set out to accomplish.
Speaker 2 (08:17):
Oh, it was.
Speaker 1 (08:18):
It was some kind of high, for sure, because I've
got all this nerves just really built up, and I'm
expecting it to fail the moment we passed the other
car for the very first Really, yeah, I'm straight up like, oh,
this one error I saw that one time that I
thought I fixed is going to come back.
Speaker 2 (08:31):
Isn't it.
Speaker 1 (08:31):
We're just going to drop that point and it's going
to slam right into it.
Speaker 2 (08:34):
Oh, I'm terrified.
Speaker 1 (08:35):
But then it does it, and it does it another lap,
and it passes another lap and it's fine. And afterward
I'm like, oh my goodness, it worked. I can't believe
it actually worked, because no matter how much testing you do,
you know you're always going to doubt your masterpiece.
Speaker 2 (08:47):
Yeah, you don't know until it's put to the test.
Speaker 1 (08:49):
Right exactly, But just that immense joy when it does work.
It's incredible.
Speaker 3 (08:54):
Going back to the sensor, something I found really interesting.
You have the three the camera, the radar, the line,
each of them are great at certain things, maybe not
so good at other things, and it's finding a way
to mesh these all together. And it makes me think
of what the human driver deals with, whether it's on
the racetrack or just on the street. You're perceiving with
your eyes, your ears. There's a sense of feel as
(09:16):
well that goes into that, but you have to find
a way to turn that into computer terms, right, right,
in a way that your car is going to understand.
Speaker 2 (09:25):
That's right.
Speaker 1 (09:27):
That reminds me of a very interesting topic that keeps
coming up as well. People ask us, have you talked
with actual race car drivers? Have you talked with engineering
teams that help these drivers analyze the same sorts of
data that we're looking at out on the track, And yeah,
we did. We talked with a lot of them back
before we even saw any of these cars, before we
placed our hands on them, and we tried to make
(09:47):
it useful. We tried to take the information about the
wind's going to change how your vehicle controls, and you
have to feel how the cars moving and then anticipate
how you need to change the steering so you keep
the slip and take that turn in this pitarticular way
on the ims, because that's the way to get around
your opponents and really take it well. And we're jotting
down notes and memorizing everything that they say.
Speaker 2 (10:10):
At the beginning, we.
Speaker 1 (10:11):
Actually did not know how to translate that into machine terms.
Because that's exactly what we've come to the conclusion of
answering these questions and thinking back retrospectively is our job
now is to figure out how to translate that. And
we didn't know at first. That's the project, that was
the learning that we had to do. And I think
we're definitely more at the point where we can look
(10:33):
back at that advice and say, I know exactly how
that fits into my system. That is a tweak on
this controller gain that is utilizing this sensor better, and
it's really cool to see that growth over the few
years we've been doing this.
Speaker 3 (10:44):
The thing I've taken away being around this project over
the last few years that I was not expecting is,
as you talk about, you are breaking down what a
human driver is doing to its component parts finding a
way then to translate that into the machine language. And
that to me, when you start to see what mean,
(11:05):
what's set a human race car driver to one side,
even the human driver on the street. Sure, the amount
of information that you have to process to do something
that we do every day, and we don't think any
more about when you do break it down in the
way that you do. All of a sudden, I came
away with this greater appreciation for the task that the
human brain is able to accomplish.
Speaker 2 (11:25):
Oh, of course, yeah.
Speaker 1 (11:26):
The human brain is the most incredible computer we have,
and we don't have a compute system that's even close
to the processing power of the human brain yet, which
kind of brings us to another part of the autonomy problem.
Speaker 2 (11:36):
We need to simplify what the human brain does.
Speaker 1 (11:38):
We need to find what are the most essential parts
of that program that's going on up here, and that
has been one of the keys in the racetrack because
we have such limited test time because the tracks are
high demand and incredibly expensive, and we're also sharing this
track time with the eight other teams, so we need
to be very good at simplifying our system down to
(11:59):
a way that we can work with it in this
type timeframe. We need to be able to see exactly
where the problem was when the problem occurs, and then
adjust the code in the hour between our next shot
up and thirty minutes out on track. So simplification of
what the human brain.
Speaker 3 (12:12):
Does without giving away industry secrets. Can you give us
a couple bullet points of when you do boil it
down to its most basic components, what kind of things
are you identifying?
Speaker 1 (12:23):
Let me think about that for just a moment, from
an incredibly academic standpoint, like thinking about all the complicated
research algorithms that we read papers on and we're trying
to compete with and build upon the cutting edge of
the research when we do publish. As an academic, we
started out with a lot of those ideas, let's have
(12:43):
incredibly complicated control logic cutting edge, and then we find
that that makes it much more difficult to adjust our
system during test days. It makes our system less flexible.
If it's tuned in perfectly, if the assumptions that you
make about the weather of the day are perfect, if
your censors are giving you quality enough data for you
(13:05):
to run this really high edge algorithm, theoretically you'd get
the better result out of it, But on average, we
need to simplify down to something that's been published several times,
and you wouldn't be able to get a paper out
about it anymore.
Speaker 2 (13:18):
So we've boiled it down to.
Speaker 1 (13:22):
Yeah, we've basically taken it from the complicated research level
algorithm down to just textbook level. Yeah.
Speaker 3 (13:29):
Well, fair enough, And I know you don't want to
give away anything. This is incredibly intense competition between teams,
but also there is a great deal of camaraderie. That's
the thing I've noticed is well being around the teams.
As much as each team wants to win, and they
clearly do, there is an element where it seems like
everyone's in this together and they're all trying to paddle
(13:50):
in the same.
Speaker 1 (13:51):
Direction, of course, and I think part of that comes
from those early days back in twenty one when we
just had the one car and we needed to.
Speaker 2 (13:57):
Figure out to turn it on reliably.
Speaker 1 (13:59):
We need to figure out how we get all the
base communication software going before you even start putting your
competition stack on top of it.
Speaker 2 (14:05):
All of that was done collaboratively. All of the.
Speaker 1 (14:08):
Teams were there together and learning these cars for the
first time together, and I think a lot of that
has stuck through the years. It's one of my favorite
parts of the competition. I may not have fallen in
love with it as much without that sense of community,
because I do like to come back and see all
of my friends on the other team, and I'm sad
when they graduate and they leave and go off to
big careers, and it's great to see the same faces
(14:29):
that have stuck around.
Speaker 3 (14:30):
You did fall in love with it, though, and that's
actually what I'm really interested in here too, because you
came at this from an interest in robotics, not necessarily
in motorsports or in competition of this kind. So how
quickly did you start to build that appreciation for what
you were about to do and get that feeling of
infatuation with this project.
Speaker 1 (14:51):
It was probably a slow boil over those months of
being together, and I don't think I realized it until
it was almost over. It was almost October. I think
it was twenty third when we had that first competition.
I started to realize I might not see some of
these people again, the people from Munich and from Milan
and other places around Europe. It's possible I could have
(15:12):
never seen those guys again. And we had built up
a decent friendship, you know, a community bond, and so
I was like, oh, I need to make sure I'm
trading communications and spending time with them now at the end,
and just that a little bit of fear of not
having that experience anymore, and then the joy whenever the
competition did actually continue on the further seasons, as like
(15:33):
I have to come back, I have to keep.
Speaker 2 (15:34):
Being a part of it.
Speaker 3 (15:35):
Well, you're right, at that time, we didn't know what
the future held beyond that first event at Indianapolis. But
since then there have been several further competitions, and the
way that this has progressed the rapidity with which it
has progressed, the complexity of the task at hand, going
from time trials with a minimal amount of obstacle avoidance
to a true passing competition where you're sharing the track
(15:58):
with another entity that you don't have the data on.
You don't know what that car is going to do.
You have to have a car that's flexible enough in
its own algorithm to be able to adapt to that.
Just how much more challenging has this proposition become as
we get deeper and deeper into the in the autonomous challenge.
Speaker 1 (16:18):
I think because I've been with it since the beginning,
so I've been growing kind of along with it. Sure
it feels a lot like the same level of complexity
from the beginning, because it was a struggle back then
just to get everything started. And now that we have
that baseline, I don't think about the baseline stuff working
anymore because that's all good and now it's just that
same level of extra work to hit the next level.
(16:40):
So to me, it doesn't feel like it's more complicated,
even though I do. If I take a second and
think about it, I realize we are achieving pretty incredible stuff.
We certainly couldn't have pulled off what we're doing these
days with the road courses and the passing competitions.
Speaker 3 (16:53):
One thing I'm always interested in when I try and
describe this to other people who work in the motorsports industry.
It takes a little bit to kind of get them
on board, but then they start to become interested. Are
you feeling or do you experience the same thing? And
what is it like when you try and explain to
friends at home or family members or whatever, what it
is that you're working on. What is that process like
(17:15):
to try and get them to wrap their head around
exactly the level of complexity and the level of cool
that you're working with.
Speaker 1 (17:22):
Sure, yeah, that makes me think of all of the
times that we've been lucky enough to show the car
off to the general public, and we've taken it to
a few racing events where actual indie races are going on.
We got to speak to people who are fans of
traditional racing, and it definitely comes up with the question, Oh,
what's going on with this car?
Speaker 2 (17:41):
Is it a robot? You know?
Speaker 1 (17:42):
Is it remote control? Everyone thinks it's remote control. We
have to explain what autonomy is and then they say, oh,
that's taking all the fun out of it. That's the
most common thing we get after what is it? I
don't understand? And then that's no fun? You know?
Speaker 2 (17:56):
Who am I going to root? For sure?
Speaker 1 (17:58):
And So what I've been trying to do in response
to those sorts of questions is I fell in love
with it, as we've mentioned several times. So how can
I get that same feeling to be shared out with
the audience? How do I make them see it the
way that I do? And so I try to talk
with these people who are asking us these questions about
how much work it is for these students behind it.
Don't think of the car as an individual by itself,
(18:22):
think of all of the people who are putting their
passion and effort into it. M hm.
Speaker 3 (18:26):
And do you find that that resonates, that message resonates
what the folks you're talking with.
Speaker 2 (18:30):
I think it does a lot of people.
Speaker 1 (18:33):
A lot of people approach the car, our particular car
because it has the big Auburn University symbol on it,
and then they just walk away with these smiles on
their faces about this is super cool that my university
is doing this, that Auburn kids are doing this.
Speaker 3 (18:46):
I think that actually is one of the neat hooks
of this competition. It's not like it's the automotive industry
that it's millions and millions of dollars that's out here
doing this. It is college students, undergrad but mostly graduate students.
But these institutions that are pursuing this, why do you
think this makes sense for an institution like Auburn to
be involved in a competition like this one and not
(19:09):
just once, but ton till through that investment now for
many years.
Speaker 1 (19:13):
It fits in very nicely with the research that my
graduate lab was doing already. To me, it feels like
a very similar project. Of course, it's not the same
private sort of bringing in a bunch of money. It's
really more public spending some money, but we get to
do a lot of hands on work, which is one
of the deficits of graduate labs in general, as it
tends to be more theoretically focused. Sure, another thing is
(19:36):
the longevity of it, the repetition, or how it continues
to grow, because other projects that we have tend to
just have a very short time span. You get done
some very well defined singular problem, you deliver the results,
and then you move on to the next part of
either your career or the next project for the grad lab.
Whereas this project is able to build upon itself. We
(19:56):
have something that starts from just a little seed and
then grows and you can see it becoming something better
and better as you continue to work on it.
Speaker 3 (20:03):
Yeah, it's clearly at the cutting edge. I mean, you
just have to look at the car and that message
sinks in. I want to go back to something you
said earlier. You mentioned when you got into this you
didn't have a particular interest in cars. You didn't have
a particular interest in racing. But if I'm reading between
the lines correctly, perhaps your perception changed a little bit
after dealing with.
Speaker 1 (20:23):
This right, I'm definitely getting a lot more interested. I
don't know a lot of the driver names, and a
lot of the terminology from the different leagues is still
a little.
Speaker 2 (20:31):
Bit beyond me. Well, you're not alone in that fair, very.
Speaker 1 (20:34):
Fair, but just being around all of the other people,
because there's plenty of other people involved in this who
were avid race fans. You are avid I'm going to
tear my car apart and rebuild it type people, and
it's just been fantastic learning from them and getting to
fall into that same interest that they have, seeing how
it can be very interesting where I wouldn't have thought
of it if I weren't in this environment.
Speaker 3 (20:54):
Sure, well, your story is fascinating because you have been
there from the ground level. You've been there as this
project has built to the point now that you're leading
your own team. So what's next for you? How do
you use this as a jumping off point to accomplish
the goals and dreams that you've set off for yourself.
Speaker 1 (21:12):
That's an excellent question, and I think it is a
wonderful jumping off point. There's been so many different industry professionals,
either from the sponsors of the competition, or just any
industry that's just interested in coming to the races. Who
have been talking with myself and a lot of the
other students, and they're always trying to hire us. They say,
what are you doing next? When are you graduating? So
(21:32):
it'd be really easy to go from this and get
just amazing research or technical career for myself. I'm finding
it very difficult to leave the racing because there isn't
a professional racing equivalent to go into just yet.
Speaker 2 (21:45):
And I love my car.
Speaker 1 (21:46):
I can't leave my car. You know, her name is
Nova Miss she is very much so my baby, as
well as the baby of my other team members, and
I just I can't imagine walking away. And so it's
kept me at the university for longer, which is another
thing that the university appreciates.
Speaker 2 (22:02):
I'm sure they do.
Speaker 1 (22:03):
Master students stick around to PhD just to get more
time with the car.
Speaker 3 (22:07):
We've talked a lot about how this program has helped
you develop, and that's a worthwhile end unto itself. But
there's also of course going to be some trickle down
from this that could potentially see some of this technology,
some of the research that you're doing enter into our
daily lives from an AI perspective, whether that's in automotive
(22:28):
or elsewhere. So, having been hands on with this now
for several years, what do you see as the most
likely outcomes in terms of some of this AI entering
our day to day lives here in the short term future.
Speaker 1 (22:42):
Sure, I think one of the most immediate benefits to
the consumer market is going to be how we've been
working with this technology. So a lot of this technology
hasn't either been put in on automobile yet or it
hasn't been put in such extreme conditions. And so we're
testing these sensors for the sense building companies, We're testing
the compute units and in that sort of collaboration with
(23:05):
these sponsors, making a better unit, making something that is
more suited for this sort of environment. And those products
are going to become publicly available. They're going to go
into the OEA manufacturers of Autonomous and aid ass which
is assisted of vehicles. So I think that's going to
be the most immediate is the sensor improvements, the compute
improvements that are being tested out in these vehicles. And
(23:27):
then probably the next one up is going to be
the people who.
Speaker 2 (23:30):
Do leave this and go into industry.
Speaker 1 (23:32):
They're going to be taking all of the learning that
they've done, all of their experiences and their research ideas.
Speaker 3 (23:39):
AI is a huge talking point just in society today.
There's a lot of hand ringing about the effects of
AI for good or for ill. Sure, when you see
things like this written, or when you talk to the
family members, they maybe don't work in this field, but
they probably have opinions on it. First of all, what
are you hearing and what are your opinions on the
(24:00):
potential positive and negative impacts that AI might have in
our near term future.
Speaker 1 (24:06):
Sure, as far as true artificial intelligence, which is a
very statistically learned model, we basically look at it as
a black box. You put a bunch of information in,
it figures out how to process that information, and then
it spits a result out. Whenever I talk to people
about AI, I expect this doomsday opinion. I expect hesitance.
But surprisingly, even from the youngest person to the oldest
(24:29):
member of society, they come up again and again thinking
this is really neat. Look at how much you can
accomplish using this AI. Particularly thinking of some of the
language processing models that are out there today. There are
so many people that just you sit them down with it,
they start to play with it, and then they're like, Oh,
I can use this to make this easier, I can
design my recipes, I can figure this out. And it's
(24:51):
just it's really encouraging to see across a wide swath
of the society that people are seeing the good in it,
seeing the good that it can do.
Speaker 3 (25:00):
Is it something that lends itself to getting hands on
with it, even at the most fundamental level, to start
then to grasp what the positive outcomes might be.
Speaker 2 (25:09):
I think so yeah.
Speaker 1 (25:11):
Letting people just access some sort of an AI model
and play with it, I think does make it more familiar,
makes it more personal, and can take away some of
that fear level.
Speaker 3 (25:20):
It does feel like we're at the cutting edge of
something here in society with AI starting to become integrated,
and if you look at historical long lens it does
seem like typically when new technology starts to enter into
a culture, especially revolutionary new technology, there can be some
negative impacts. So with that said, we've talked about the
(25:45):
benign aspects of AI, but how about the malignant aspects?
What things have you concerned as we look into the
short term.
Speaker 1 (25:52):
Sure, and we can look into the automotive industry to
kind of think about that a little bit. The automotive
industry tends not to trust AI by its because these
are black boxes. It's very difficult to guarantee that they're
going to give you the safe result that you need.
It's very difficult to analyze what they're even doing to
process this data, and so we tend to like to
fuse it with more classical type approaches. This is where
(26:14):
you take a scientist or an engineer and have them
actually explicitly put their ideas down into how you should
manipulate the input data into the output data. They are
directly causing that program to give a result. You fuse
the intentional programming with the statistical programming in order to
get the high performance of the statistical programming. But the
(26:34):
guarantees and the safety and the control of the classical
programming really interesting.
Speaker 3 (26:39):
As always, anything we haven't touched on yet that you
think people would like to know for those who maybe
are coming to this that are new they've not been
following the end Autonomous challenge, or even for folks who
know a lot about it. But is there something that
you've taken away from this that you'd like to make
sure others are aware of.
Speaker 1 (26:57):
Let me think about that for a moment, because I've
thought a bit about.
Speaker 2 (27:01):
Women and engineering. Oh yeah, sure, quite a bit.
Speaker 1 (27:04):
Because I've been approached by people and they're like, are
you with the racing. I'm like, I'm leading the team,
and they're always amazed, which is a good thing to
a degree, But I think that there shouldn't be that amazement,
you know, it should be normalized. So something as visible
as this is really gives a great opportunity for women
in engineering, women in motorsports, and women and intelligent fields
(27:27):
in general. You know, people who are capable of doing things.
It gives exposure to us, and it also normalizes it
because we don't want people to be amazed that there's
a woman doing this engineering. We want to be people
to be amazed that people can do this engineering.
Speaker 3 (27:42):
Do you find do you get a reaction from young people,
young women, young girls in particular, when they find out
what you're doing, and do you feel like you're having
an impact?
Speaker 2 (27:52):
I do.
Speaker 1 (27:52):
I've actually had several one on one interactions with younger
females where they'll come in with their dad to see
their car, the car that dad might be a potential sponsor, etc.
We might see them just in the big public open events,
and sometimes there has a tant to approach the car
and they need to be encouraged, or sometimes they'll just
come right up and put their hands on it. And
(28:12):
I tend to try to ask them, do you want
to lead a race team when you're older?
Speaker 2 (28:16):
And like, yeah, say yes, say yes.
Speaker 1 (28:19):
And I like to we do have a remote control mode,
even though these are fully autonomous cars, which is more
for testing and for show, I like to put the
controller in their hand and let them wiggle the wheels
when it's on display, and you know, become hands on.
Don't be afraid to touch the technology, don't be afraid
to ask questions, and to consider that this can be
for you as well. And I like to think that
just having that momentary interaction does change the perspective. It
(28:43):
says you don't have to stay away from it because
you're a girl. This is something girls can do too.
Speaker 3 (28:47):
A wonderful message for sure, A great one to end
on as well. Stepanie, thank you so much for joining us.
I'm so excited to see what the future hold. It's
not just for you, but your entire team as well.
Speaker 2 (28:56):
Yes, thank you so much for talking with me. It's
been fun.
Speaker 3 (29:03):
Thanks for joining us this week. On the inside Track,
that was Stephanie Meyer of the Autonomous Tiger Racing team
at Auburn University. As she so aptly explained, the passion
for racing isn't something that could be measured or predicted.
It's something you just have to experience to understand. I'm
your host, Ryan Marine. Join us next time on the
(29:24):
inside Track, where we speak with Simon Hoffman of the
Tomb Autonomous Motorsports team at the Technical University of Munich
to discuss the more granular aspects of building out the
brains inside their autonomous car.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.