Driverless Cars: The Early Concepts

Wouldn't it be awesome if your car took you where you needed to go without you having to touch the wheel? In this episode, we look at the history behind the driverless car concept.

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

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:04):
Get in touch with technology with tech Stuff from how
stuff works dot com. Hey there, and welcome to tech Stuff.
I'm your host, Jonathan Strickland. I'm an executive producer with
how Stuff Works, and I heart radio and I love
all things tech and I mentioned a few weeks ago
that I was planning a suite of episodes about driverless cars,

(00:27):
including the history of developing them, the challenges we face
in implementing them, the potential benefits that autonomous vehicles could have,
and how long it might be before we see truly
driverless cars deployed on a wide scale basis, also whether
or not they'll hit the road before they're ready. Now, today,
we're going to start that journey by talking about a

(00:49):
time when the idea of driverless cars was mostly science fiction. Also,
I think it's important to understand that the development of
self driving cars is not a linear your story with
a solid narrative. It's not a leads to be leads
to see a lot of people have worked on developing
the technology independently, sometimes from very different perspectives and philosophies,

(01:13):
but all heading towards a common goal. These projects can
sometimes overlap each other in time without having any direct
connection between them, so you could have independent studies happening
throughout the world. So I'm just gonna cover some of
the big, really important ones, but just no, there were
a lot of people working on this at various times
throughout our history. I'm gonna stick as close to a

(01:37):
chronological timeline as I can, but it will involve jumping
around a little bit in time, just because otherwise I'd
be saying, meanwhile, blah blah blah was happening, and then
so and so was doing this, and it was just
turned into a really hackneyed kind of like radio mystery.
I don't want that to happen. So picking a starting

(01:57):
point is actually pretty darned tricky too, because the technology
of autonomous cars spans lots of different fields. You could
go into all sorts of technologies and talk about the
development and how the evolution of those technologies eventually found
their way into autonomous cars, and if I dug too

(02:18):
deep in that rabbit hole, this series would last maybe
thirty episodes. We'd be talking about the development of the automobile,
we'd be talking about development of cameras. I kind of
like how Wired did it? In a piece that was
published in twenty sixteen titled quote a brief history of
autonomous vehicle technology end quote. And as that name suggests,

(02:41):
that article went beyond self driving cars, and the first
item on that list actually dates all the way back
to fifteen hundred or thereabouts, and a famous turtle named Leonardo.
I'm sorry, tari Is tar is telling me that's the throng. Leonardo.
Leonardo was not actually a turtle, but apparently some sort

(03:02):
of renowned renaissance artist and proto scientists. Okay, Leonardo da Vinci. Okay,
all right, I was was way off base there, Thanks Harry. Anyway,
the story of his invention comes to us courtesy of
the Codex Atlanticus, Folio eight twelve r, in which da

(03:23):
Vinci describes the creation of a cart that would use
springs and gears. The springs would store energy, and the
gears and the cogs would drive the wheels and steering mechanisms,
so that this cart, once wound up and quote unquote programmed,
could travel and steer on its own power. So you

(03:43):
would change these settings around, and that would allow the
cart to not just travel across a distance, but also
turn left or right, depending upon the settings that you
had selected. I imagine this was largely done through various
gears that once it hit a certain point of its revolution,
it would engage the steering mechanism, But honestly, I don't

(04:06):
know the full details. Uh If da Vinci ever built
one of these for real z s. The working model
did not survive the passage of time, but modern enthusiasts
have attempted to recreate this self driving cart, which we
now believe was intended for use in theatrical productions. It
wasn't meant to do practical work on like manual labor

(04:29):
or anything like that. It was meant to be part
of a show that you would have this cart go across,
perhaps as an effect, maybe it's carrying some sort of
piece of scenery or an actor or something. But Da
Vinci's drawings and designs, we're not step by step instructions,
and so any recreation of this invention has to be

(04:50):
done with a healthy dose of interpretation on the part
of the builders. They have to start kind of filling
in the gaps and making, you know, best guesses. But
some people, such as the staff at Wired, point to
this as an early example of a vehicle that can
drive itself. Though to be fair, da Vinci's design required
some humans to pre set the device so that it

(05:13):
would do what it was supposed to do. So you
couldn't just, you know, tell the cart I need you
to cross upstage left. You would actually have to quote
unquote program it. Now, no discussion about the early days
of driver less cars would be complete without acknowledging the
accomplishments of someone known as Francis P. Houdina. He is

(05:39):
credited in various articles as being an electrical engineer for
the Army, and once he left the service, he founded
a company called the Houdina Radio Control Company. I'll have
more to say about Francis in a second, but let's
get to the heart of the matter as far as
autonomous cars go. Back in ninet, the Houdina Radio Control

(06:01):
Company engaged in a major publicity stunt. It happened on
July when a car dubbed the Lendrick in Wonder, though
sometimes it is credited as the American Wonder, roamed the
streets of New York City without a driver behind the wheel.
In fact, there was no one in the car whatsoever.

(06:24):
There was someone standing on the sideboard of the car,
the running board so they were perched on the outside
where they could reach in and grab the wheel, but
no one was actually sitting in the driver's seat Ellie,
and no one was apparently in control, at least not
directly behind the wheel. But as the name of the
company would indicate the line, Rick in Wonder was a

(06:46):
radio controlled car like a toy r C Cards, just
instead of being a toy, it was a full sized automobile.
Circuits connecting to motors would control the movements of the gearshift,
the accelerator, the brakes, the steering wheel. Actual control of
the vehicle came from an apparatus inside a following car,

(07:08):
So you'd have a car operated by a regular human
driver sitting in the driver's seat. In the passenger seat
would be the operator for the radio controlled car that
would travel in front of them. So the car was
not truly driverless because the driver did exist. It's just
the driver was in another vehicle. While researching this episode,

(07:30):
I found a New York Times article that described the event,
and it doesn't sound like it was a complete success.
From that article, these are quotes. A loose housing around
the shaft to the steering wheel in the radio car
caused the uncertain course as the procession got underway, as
John Alexander of the Houdina Company, riding in the second car,

(07:52):
applied the radio waves the directing apparatus attached to the shaft,
and the other automobile failed to grasp it prop really.
As a result, the radio car careened from left to
right down Broadway, around Columbus Circle and south on Fifth Avenue,
almost running down two trucks and a milk wagon, which

(08:12):
took to the curbs for safety. At Street, Houdina lunged
for the steering wheel, but could not prevent the car
from crashing into the fender of an automobile filled with cameramen.
It was at Fort Street that a crash into a
fire engine was barely averted. The police advised Judina to
postpone his experiments, but after the car had been driven

(08:35):
up Broadway, it was once more operated by radio along
Central Park drives. And uh, here's that bit about Francis
Judina I was mentioning earlier. You may have noticed that
the name bears some passing resemblance to that of Harry Houdini,
who was alive at that same time. He was the
famous escape artist and magician Now, according to the story,

(09:00):
and there are at least some court documents to back
this up, the post office was sometimes in the habit
of delivering some of the mail meant for the Houdina
Company to Houdini the magician, including bills, and Houdini was
not crazy about getting the bills from some other company

(09:21):
and he felt that it was encroaching upon his name.
So he marched on over to the Houdina Radio Company
headquarters and a scuffle broke out after he started raising
a fuss. The head of the office, a guy named
George Young, filed charges against Houdini for disorderly conduct, but
on the day that Houdini was to appear in court,

(09:42):
Young failed to show up, and so all charges were dismissed.
Now there's a guy named Dean Carnegie who posted a
few years ago that he had been contacted by the
son of the person who called himself Francis Houdina, and
that Judina was a pseudonym. And further, he says that
the son of Houdina revealed that this scuffle was all

(10:05):
just a publicity stunt, that Houdini had thought it up
and they had all worked on it together. I do
find it odd to have gone through all this trouble
to take on a pseudonym, establish a company, create a
demonstration of a radio controlled car, and all in an
effort to set up a big PR stunt with a magician.

(10:28):
It seems to me like that's an awful lot of
trouble to go through before you get to a point
where you can hold this PR stunt. But I guess
if someone was going to do it, it would be Houdini.
I am highly skeptical that it was in fact a
publicity stunt, only because, as I say, it's an awful
lot of trouble to go through in order to do it,

(10:50):
and it wouldn't necessarily bring good publicity, I would imagine.
I just thought I would share the story because when
I was doing the research, it kept popping up. Now,
this Houdina radio controlled vehicle was not the only one
of its kind. There were some other people who also
built full size radio controlled apparatus for cars. There were

(11:12):
a few in the twenties and thirties, and in general
people would start calling these phantom cars because they appeared
to be driven by an invisible phantom, but all of
them were actually controlled by a remote driver. Those controls
might not resemble those of a typical automobile, but they
would end up controlling motors that would affect the automobiles

(11:35):
regular operations, and according to various accounts, some of these
would be controlled by follow cars, sometimes by someone on
the street, sometimes by someone following in a low flying aircraft.
It all just depends upon the account. In nineteen thirty nine,
at the New York World's Fair, General Motors presented an

(11:55):
exhibit called Future Rama And it was not a and
aimated series by Matt Graining. It was something else. It
was a vision of the far off future of nineteen
sixty and the General Motors exhibit was a ride. Essentially,
visitors would get on these cars, these chair cars, and

(12:17):
be pulled through this exhibit, where there was an enormous
scale model. One part of this vision, which encompassed lots
of different thoughts about the future, was the Motorway of
the future, and the vision included a sort of driver
assist system in cars. They described an electro magnetic breaking

(12:39):
system that would engage if one car were to get
too close to the car ahead of it, and cars
would be traveling down lanes that would have raised walls
on either side, sort of like guiding slots. So it
sounded like GM was pretty sure cars would still be
under the control of human beings even in the far
off future world of nineteen six team, but there would

(13:01):
be some automated elements that would find their way into
vehicle operation. And these days we would say that that
vision of the future was pretty much on the money,
except for the part about it being nine, because it
took a little longer than that. Norman bell Getz, who
designed this exhibit for General Motors, described a couple of
possible methods for controlling traffic in the future using radio waves,

(13:25):
and in one he envisioned a system that would include
numerous broadcast towers along the side of the road, and
so cars would maintain contact with these broadcast towers. But
another version, he suggested the possibility of an electrical conductor
embedded in the road itself for direct control, and that
would be a method that a lot of different people

(13:47):
would look into over the next few years. Today we
would say that such an enormous system is probably unlikely
because of the huge investment it would require an infrastructure
that would come along with it. But back in the
nineteen thirties, the US highway system was still developing, it
was still being built across the country, so it's probably

(14:07):
seen as more of a possibility. After all, we were
already connecting distant parts of the country to each other,
so couldn't we just go to the extra effort to
wire all of that in some way. I have more
to say about this version of autonomous cars and how
that evolved into what we think of as an autonomous
car today, but first let's take a quick break to

(14:28):
thank our sponsor. In nineteen thirty six, a magazine called
Modern Mechanics published an article about a different method for
autonomous control of a vehicle. This version would include building
cars that had special photo cells on them to detect

(14:51):
specific frequencies of light. The car it self would project
this light. It would have a projector of some sort
on the front of the car, and then in the
road itself would be steel mirrors that could reflect the
light back at the car, and the photocells would pick
up that light. The article pointed out that cars were
already approaching the limits of human reflexes. This idea that

(15:15):
we were starting to drive faster than we could react. Um,
I guess, uh you know, good old uh Jack of
the Pork Shop Express would say he never, he never
drives faster than he can see. But we know that
human reflexes have their limits. So this article mainly focused
on the link between the photo cells and the steering

(15:36):
mechanism for the car. That leaves a lot of questions unanswered,
such as how the car would accelerate or break, would
there still be a human responsible for those operations, and
the only thing that would be taken off the plate
of the human driver is steering the car. One also
wonders exactly how any one car would deal with the
presence of other cars on the road that are similarly outfitted. Right,

(16:01):
So if you have a whole bunch of cars that
use this technology and it's dependent upon light being reflected
back at them, what happens when you start getting the
signals from one car picked up by another, or what
happens if you're driving on a really sunny day or
in really bad weather like fog or rain. But the
point is people were already thinking about alternatives to radio

(16:23):
controls even in the thirties. In nineteen fifty three, Arthur
mac Barrett created what he called a driverless vehicle he
mounted a wire in the ceiling of a warehouse, and
he had a specially outfitted vehicle. In this case, it
was a towing tractor that could follow that wire, and
in later versions he would bury the wire within the

(16:45):
floor of a facility. And he called this system the
guide O Matic. Now this was not meant for cars necessarily,
but it was based on similar principles to some of
the more sci fi ideas that were proposed in the
nineteen thirties. So we started seeing it actually being put
to use in the fifties, and in fact, this kind
of system is still used to this day, but obviously

(17:08):
with a much more sophisticated equipment than what was available
in the fifties. In Disney Studios produced a segment titled
Magic Highway USA as part of The Wonderful World of Disney,
and the piece included some humorous gags about what the
future of driving could be, just based off jokes really,
but after that was a more measured look, a more

(17:31):
thoughtful look at the future of driving. Still had some
pretty far out ideas in it. Now it did include
some practical stuff that has in fact come to pass.
For example, the idea that signage is going to need
to be larger and simpler so that motorists traveling at
great speed can read and understand the road signs in general.
That did come to pass. It also predicted a road

(17:54):
system that would be able to retain heat to keep
it dry even in snowstorms. That has not happened. It
does sound a lot like some of the smart road
systems that were being peddled no no pun intended around
a few years ago. You may remember those the smart
highways that we're supposed to be made up of photo cells,
solar cells essentially, and they could soak up light, generate electricity,

(18:19):
and even be warmed so that snow and ice woulden
form on them. Those have not really panned out so well,
at least not in any widespread application. However, some of
the other bits in that magic Highway section included predictions
similar to GPS and rear mounted cameras on cars, so

(18:40):
we do have those, although the one that was proposed
in the piece was more of a full time rear
view camera, so instead of having a rear view mirror,
you would have a rear view screen that you would
consistently looked to for information about what's going on. Behind
the car. Some of the other predictions did not pan
out in any way, shape or form, such as tunneling

(19:03):
by atomic energy. Yeah, the actual special suggested that we
use an enormous like atomic ray cannon essentially that would
melt a hole through the side of a mountain when
we needed to build a tunnel for a highway. That
clearly has not happened. But the special then goes on
to suggest that in the future will get in our

(19:25):
family vehicles and with a push of a few buttons,
which in the specials depicted as physical slighter controls kind
of like you would see on a on a stereo
or a soundboard, we would select our destination and an
electronic system incorporating the vehicle and the road itself, so
it would be a system that has both internal components

(19:46):
in the car and external components in the environment would
take care of everything else. So again this vision hinges
on that sort of smart highway concept, the idea that
a lot of this work is being done by the
infrastr ructure, not just the vehicle. Disney was just one
company to promote this kind of idea. America's Electric Light

(20:07):
and Power companies ran an advertisement in the Saturday Evening
post in the nineteen fifties with an illustration that showed
the stereotypical nineteen fifties American family depicted as it was
at that time in the medium, which is to say
it was a white upper class or upper middle class
at least family. There was a father, mother's son, and daughter,

(20:29):
so that kind of stereotypical family. All four of those
family members are inside a car that has sort of
like that big glass bubble kind of approach, sort of
what you would see in something like the Jetsons, and
they're all facing inward toward each other. A couple of
them are playing dominoes, they're having conversation. No one's having
to drive right, the car itself is doing it. And

(20:50):
the ad talks about how the electric age will lead
to automation and efficiency in all sorts of areas, including
stuff like flat TV screens and vehicles control by quote
electronic devices embedded in the road end quote. Now keep
in mind again this is in the post World War
two era. This is an era in which America's industry

(21:11):
was a key component of national identity. It was part
of what people thought of when they were asked the
question what is it to be American industry and innovation
were very much, really important components of that identity. And
these weren't just concepts. These weren't just artists and advertisers saying,

(21:34):
let's come up with some sort of wild idea. There
were engineers who were actively building cars and test roads
to work out the actual details. Joseph Bidwell and Lawrence
Hofstad who were researchers with General motors outfitted in nineteen
Chevrolet with pickup coils to work with a road that
had embedded electrical wire in it. The coils were connected

(21:55):
to motors that could adjust the cars steering so that
can continue to follow the wire below, very much much
like the guide Oh Manic that talked about earlier. Meanwhile,
over at our c A another smarty pants was working
on this challenge. This would be Vladimir Zwarakin, which some
of you people may know as one of the pioneers
who played a really big part in the development of television.

(22:18):
In fact, depending upon whom you ask, it was Warrikin,
not farms Worth who was pioneer of TV. But honestly,
it's a very complicated story, and I've talked about before
on tech stuff so I'm not gonna go into it here,
but back to driverless cars. His concept included embedding circuits
in the roads that would be able to sense vehicles magnetically,

(22:40):
and his vision had the circuits identifying the speed and
position of vehicles, which would provide information to a centralized
system that could then send out instructions to specific cars
in order to manage traffic. And his idea turned out
to be impractical for widespread deployment for autonomous cars. However,
it did becomes sort of the foundation for car sensing

(23:02):
loops that are under many intersections. They're used to help
control traffic lights. Those loops that can detect if there's
a vehicle on top of it through the electromagnetic effect
and thus send a signal to the traffic lights that
they should switch over soon so that they change the
direction of traffic. That's pretty cool. A key component in

(23:24):
many of these concepts was that the system for control
lay outside of the vehicle itself. It required some sort
of larger centralized system to handle things, and the cars
would respond to commands from that system. A car might
have some components abort it to help with this, but
for the most part, the important elements were external to

(23:44):
the vehicle. So why was that? Why were we thinking
outside the car? Well, keep in mind that before we
did not have transistors, so electronics were very large and bulky,
and even the seven transistor was not a practical component
that you would incorporate into a finished product. So it
would be a few years before transistors would really play

(24:07):
an important role in consumer technology, and mentorization was just
getting started in the fifties and sixties, so computers were
enormous machines that would take up at least a desk,
but sometimes an entire room. So driving, because it's such
a complicated task, meant it wasn't really practical to create
a fully autonomous car. The computer you would need to

(24:31):
calculate all the different decisions that would be made in
order to drive a vehicle would be bigger than the
car was. It made more sense to look outside the
vehicle for the components that would be needed for a
driverless automobile and send commands to a car that would
be more like a dumb terminal would be for a supercomputer.
Experts recognized the potential for autonomous systems. In particular, many

(24:54):
engineers believed a good system would save lives and prevent injuries.
As we became accustomed to travel link at higher speeds,
there was a legit fear that people were driving too
quickly to be able to react safely in the event
of an emergency. In nineteen sixty, Norbert Wiener, a mathematician
at m I T. He's also known as the father
of cybernetics, said, quote, by the time we are able

(25:17):
to react to our senses and stop the car which
we are driving, it may already have run head on
into a wall end quote. He was advocating for some
sort of feedback system that could react in a fraction
of the time humans can, And he had a point.
Reaction times can average between a hundred fifty milliseconds to
three hundred milliseconds, depending upon this stimuli, and that gets

(25:40):
pretty darn fast, and a hundred fifty milliseconds it's not
a lot of times, so that's a pretty fast reaction time. However,
let's say that you're driving in a car that's going
sixty miles per hour or around kilometers per hour. That
means you're traveling at eight ft per second. Even if
your reflexes are on the fast side. That hundred fifty milliseconds.
It means you travel thirteen ft before you'd even start

(26:03):
to do anything, you would see something happened to you,
and by the time you were able to start touching
the break, you've all already traveled thirteen feet. Also, if
you're traveling sixty you're in a vehicle. The vehicle has
a pretty hefty mass. You've got a lot of inertia
to deal with. Two, You're not gonna stop on a dime.

(26:24):
It's gonna take you time and therefore distance to stop.
So the thought was if we could build out vehicles
that could react for us much more quickly than we
ourselves can react, and that these vehicles could monitor conditions
that surround the car at all times, not just what's
happening at whatever you happen to be focusing on at

(26:45):
that moment. Wouldn't that be great? And we will revisit
that idea and a couple of episodes. When we start
talking about the arguments for autonomous cars, we'll also talk
about the arguments against them. I've got a lot more
to say about these early concepts and autonomous cars, but
first let's take another quick break to thank our sponsor.

(27:13):
You know it would be great if our cars could
watch after us. But the researchers, engineers, and mechanics of
the sixties, we're running into huge design challenges and progress
was pretty slow. Money for autonomous systems was running low
as well, as the automotive industry began to dedicate funds
toward developing technology that would help mitigate human error. So,

(27:34):
in other words, the tech to take human error out
of the equation, that is, to make driverless cars was
really complicated and beyond our ability to to realize at
that time. So instead companies shifted to, well, human error
is going to happen. We can't take it out of
the equation, so let's figure out how to have human

(27:56):
error make the least negative impact possible. So for that
reason we saw money instead being dedicated to the development
of other technologies, stuff like seat belts, airbags, anti lock
brake systems, and it would stay that way throughout the
nineteen sixties and nineteen seventies. It really wasn't until the
nineteen eighties that we started seeing serious work and experimentation

(28:18):
in driverless systems going again. For land based vehicles. Keep
in mind, we had had things like you know, automated
pilots for a long time, that was a relatively simple
problem to solve compared to cars. Cars continued to be

(28:39):
a difficult problem. Now. One of the engineers who did
very important work in the nineteen eighties was a guy
named Ernst Dickman's from Germany. He ran a lab at
bundes Verre University in Munich, Germany, and he started out
as an aerospace engineer, so super smart guy, but he

(29:00):
had ambitions to work on creating a way for vehicles
to be able to see their surroundings and then react
to them. His work would provide the foundation for tons
of innovation in dynamic computer vision. So in the nineteen eighties,
he and his research team took a van that was
manufactured by Mercedes Benz and began to customize it for
driverless operation. Now, according to Dickmans, the university sort of

(29:24):
just let him do this because he had a reputation
of being brilliant, so they said, well, he's a smart guy,
let him do what he does. So his team refitted
this van with various systems that would be able to
control steering, acceleration, breaking. They also outfitted the vehicle with
a computer system to process information and then sensors and
cameras to gather information. So you have the sensors and

(29:48):
cameras that bring in data, send it to a computer.
The computer processes of the data and then sends commands
to the various control systems to change the behavior of
the vehicle. That's your basic concept behind the modern autonomous car.
So they incorporate technology that could detect the steering angle,
break pressure, temperature, acceleration in both latitudinal and longitudinal directions

(30:12):
and more. Uh. The camera was actually a pair of
cameras mounted on swivels that could move along two axes
in order to focus on specific points within the field
of view. And they called the experiment v A M
O R s V mores with a big V, big M,
big R. So alternating caps and lower case. Sticking a

(30:35):
camera on a car is one thing, right, anyone can
really do that. Teaching a computer to interpret images from
that camera is another thing. Entirely and in the nineteen
eighties and normally would take a computer several minutes to
analyze a single image in any meaningful way, and even
that was fairly limited compared to what we can do today.

(30:57):
So to be useful in the driving scenario needed something
dramatically better than that. A computer would have to analyze
many images per second, like ten images per second, not
one image every ten minutes. So how do you fix
that problem? That's an enormous challenge. Well, Dickmon's solution was
to limit what the car was actually looking at, and

(31:19):
so he took human eyesight as kind of a source
of inspiration. You see, we're only really able to focus
on a relatively small part of our vision. Everything else
that's in our field of view is there, but it's
not really in focus. So we concentrate on whatever we
have deemed to be important at that moment. It might

(31:39):
be traffic ahead of us, or an incoming soccer ball
kicked at our heads, or whatever it may be. So
Dickmon's thought, hey, if I limit what the computer system
is focused on and I let it ignore everything else,
then I limit the amount of data that needs to
be processed and everything speeds up as a result. So
he focused on finding shortcuts, such as programming the computer

(32:03):
to only really look at stuff like road markings and
to ignore other things. His work was dedicated to creating
an early driverless system that could function on an empty
stretch of road in the early days, so it wasn't
really important to worry about other things that your average
driver would have to worry about, like other vehicles or obstacles.

(32:25):
He was just concentrating on how can I make a
system that will reliably follow a road without having a
driver behind the wheel. He was building the foundational blocks
at that time. Dackmans also sped up the computation process
by limiting the need for the computer to save images,
so it was really just analyzing and responding to each

(32:47):
image and then conveniently kind of forgetting about them. His
techniques paid off with some early demonstrations, but they relied
heavily on predictable and reliable components like those road markings.
But if the road road markings were obscured or if
they were absent, then the car would start to drift
out of its lane. It didn't know quote unquote where

(33:08):
it was supposed to be, and it might just continue
to wander on whatever steering direction it was in before
it lost track of the road markings. And we're talking
about a five ton van that his team had been testing,
so that's potentially a real danger now. To be fair,
they were testing it on unoccupied stretches of road, so

(33:31):
at least the potential for a catastrophe was severely limited.
They worked on like the unopened stretches of the Autobahn,
for example, so it was more safe than what I'm
making it sound like, especially since this was new road,
so that the times when the road markings were not

(33:53):
detectable were rare. Dickman's work would become a part of
a huge project much bigger than just autonomous cars, called
the Eureka Framework, and Eureka is still around today, but
this was sort of a a European Union kind of
thing before there was a European Union, so it Eureka

(34:13):
is a pan European research and development funding organization. It's
meant to make sure that European nations remain competitive with
other countries, namely countries like Japan and the United States.
For several years, Dickman's and his team were working on
refining this driverless car technology, and that culminated in demonstrations

(34:35):
that happened in nine and n one happened in Paris, France.
Dickman's had a huge challenge ahead of him. Daimler, with
whom he was working wanted his team to equip passenger
cars Daimler passenger cars with this driverless technology, and they
intended for the team to have one of these cars

(34:58):
navigate a three lane highway in public traffic, being able
to make automated changes of lane and everything while carrying real,
live human passengers, and his team would only have a
couple of years to accomplish this goal before this demo.
After a brief consideration, Dick Wan's agreed to this challenge

(35:19):
and he got to work and in October, his team
picked up several important people at the Charles de Gaulle
Airport and took them to a highway and then flipped
the automobiles, not literally, but they flipped them to autonomous mode.
And both of the cars that were used in this
demonstration still had human drivers sitting in the driver's seat.

(35:41):
They still had their hands on the wheel, but they
weren't putting any pressure on the wheel. They weren't turning
the wheel. They just had their hands there in case
something should happen, so they would occasionally take their hands
away from the wheel to show that the cars were
in fact driving themselves and that they were just there
for safety sake. His team took an altered vehicle on

(36:04):
an autonomous trip from Bavaria to Denmark that's more than
one thousand miles or seventeen hundred kilometers. The car reached
speeds of up to a hundred nine miles per hour
or a hundred seventy five kilometers per hour, so pretty
impressive for an autonomous car. Now, despite these remarkable achievements,
the technology was still too primitive for widespread use. It

(36:25):
depended heavily on predictable factors. Anything outside of that was
more of a challenge, particularly obstacle detection. They didn't build
it in to be consumer friendly, so you had these
large computer systems that were inside the vehicles themselves. So
it was an exciting advancement in autonomous car technology, but
it wasn't far enough along for consumer or practical use,

(36:46):
and so the world would have to wait a bit
longer for tech to evolve to give autonomous cars another go.
And this is related to a concept called AI winter,
which is tied into the idea of hype cycles. And
AI winter is named that way because it's considered to
be the funding equivalent of a nuclear winter, and it
describes the time when there's a growing reluctance to fund

(37:11):
AI projects. Generally speaking, this is how the pattern tends
to play out. You get some super smart people making
some cool advances in artificial intelligence, and those advances may
one day have practical application in numerous technologies, but early
on we're talking about truly experimental work that's exciting but

(37:31):
not necessarily practical at the moment. However, word of that
work gets around. Maybe the company sponsoring the research releases
a big press release that implies breakthroughs are closer than
they really are. Maybe the media picks up the story
and they run with it. Enthusiasm among the general populace grows,
and funding gets easier to secure. But as time passes

(37:54):
and it becomes clear that in reality, these sort of
things take a lot of time, and they take a
lot of work, and they take a lot of money
to make progress in fields like artificial intelligence, then people
get less enchanted. Typically, starting with the media, you get
these stories that are the equivalent of where's my flying car?
And the narrative changes from think about how the cool,

(38:15):
how cool the future is going to be too? Why
isn't the future here already? So enthusiasm for the field drops,
and then funding drops and that in turn sets back
the field even further, which delays any other big breakthroughs
in the process. Eventually, this part of the cycle comes
to an end if you're lucky, and then enthusiasm can

(38:36):
begin to build again. AI has experienced several of these cycles,
and we've also seen the same thing in other fields
as well. Virtual reality is a field that leaps to mine.
Dickman's work was really exciting, and it had even survived
one AI winter in the late nineteen eighties uh and
got all the way through the mid nineties, But at

(38:58):
that time the funding was really becoming scarce and his
work had really gone about as far as it could
go based upon the sophistication of technology at the time,
and so it kind of came to an end, and
his pioneer work was largely forgotten for many years. In
our next episode, we'll look at the resurgence of interest
in autonomous cars and how the US Department of Defense

(39:20):
got involved. But for now we're going to conclude this.
If you guys have any suggestions for future episodes of
Tech Stuff, or you've got any stories you want to
share about autonomous cars, or maybe there's some guests I
should have on the show. Anything like that, let me know.
Send me an email. The addresses tech stuff at how
stuff works dot com. You can go to tech stuff

(39:41):
podcast dot com. That's our website with all the information
about the show. In other ways to get in touch
with me, don't forget to go to our store over
at t public dot com slash tech stuff. Every purchase
you make goes to help the show, and we greatly
appreciate it. And I'll talk to you again really soon

(40:02):
for more on this and thousands of other topics. Is
it how stuff works dot com, wh