The self-driving future started with a competition nobody won

The self-driving future started with a competition nobody won

The self-driving revolution started with this car and a competition in the desert that nobody won. Some people now claim that nothing has happened in autonomous driving. They are about to be disrupted.

Mathias Sundin
Mathias Sundin

Share this story!

This is a premium article, open for everyone thanks to our premium supporters.

Self-driving cars are happening now, all around us. But some people might not be seeing the forest through the trees and risk fooling themselves that a radical change is still decades away.

The time to see, and seize the opportunities of a self-driving future is now.

Recently, more and more people have started to claim that "nothing happened" with self-driving vehicles. "$80 billion has been spent on self-driving cars with nothing to show for it," writes Lloyd Alter at Treehugger.com

It's a great example of one of Peter Diamandi's Six D's.

The Six D’s are the phases a product goes through when it is digitized.

The first d is digitalized, the second is deceptive. Although a rapid development is underway, the product has not yet become good enough to have a breakthrough, or even a practical application, which becomes deceptive.

The world's first digital camera.

The digital camera is a good example.

It was invented in 1976 by Steve Sasson at Kodak, but it wasn't until 2002, when Casio Exilim launched, that digital cameras got their breakthrough, seemingly overnight. While development had been rapid since the 1970s, the product wasn't good enough to become a widely adopted consumer product.

In 1988, Fujifilm built what was intended to be the first commercial digital camera, but it never made it to market because its specs were simply too bad. With only 0.4 megapixels and a memory capacity of only 5 images, the product was too expensive for consumers to justify.

It could then have been said that "nothing has happened" since 1976. The camera sucked then, and still sucked in 1988. But the 1988 version was almost 4000 percent better than Steve Sasson's camera. Deceptive!

4000 percent better, but still not good enough. That's where we are now with self-driving cars. They have come a long way since the competition that started the coming revolution.

The 2004 prize competition for self-driving cars - which nobody won

DARPA is an American research authority, perhaps best known for creating a model for the internet known as ARPANET. After the invasions in Afghanistan and Iraq, they saw the need for self-driving vehicles, that could operate in dangerous environments, but technological advances in the field had not progressed far enough.

Experiments and projects had been going on since the 1920s, but in order to get the development to take a big step forward, DARPA announced a prize competition. One million dollars in prize money for the team that managed to get a self-driving vehicle to drive 24 miles in the Mojave desert in the shortest time.

Prize competitions are a smart way to push development forward. If you manage to set a difficult, yet achievable goal, and a sufficiently large sum of money, you get a mix of competitors from several different disciplines with different ideas and approaches, which together invest more money than you spend in prize money.

The reason Charles Lindbergh flew across the Atlantic when he did was to win $ 25,000 in The Orteig Prize. 16 times more money than the prize money was invested by all competitors and became the starting point for a strong expansion of the airline industry.

The Orteig Prize was the inspiration for the Ansari X Prize, where $ 10 million was awarded to the first private space flight, in 2004. Today, XPRIZE is a foundation that conducts prize competitions in several areas.

Red Team's self-driving car, Sandstorm.

There was great interest in DARPA's Grand Challenge, and hundreds of teams were formed.

Among these were Carnegie Mellon, who had been involved in a project in the 1980s where the result was the world's first self-driving vehicle. Their Red Team was led by Red Whittaker, professor of robotics at CMU. They rebuilt a Humvee and equipped it with a light radar, LIDAR, other kinds of radar and cameras and when they rolled it out one early morning in March 2004, many thought that if anyone was to win, they were it.

Off we go - straight into a bale of hay

Red Team’s vehicle, named Sandstorm, rolled off the road and drove directly into a bale of hay. At first it seemed like the car was unaffected. But after a few kilometers Sandstorm began to behave a little bit strange and seemed to think that it was half a meter farther to the right or left than it really was.

A number of posts had been cut down and in a sharp curve, Sandstorm drove off the road and became stuck. One of the wheels reached the ground and when the computer noticed that the car was not moving forward it increased the gas, with the result that the tire started to burn.

Sandstorm remained standing on its belly, just over 11 kilometers after the start, with almost 23 miles left to the finish.

Another competitor also got stuck. A third was fooled by a small rock and stopped. A fourth couldn't accelerate enough to get up a hill. TerraMax, a 15-ton monster truck, was tricked by tumbleweed that blew in front of and behind the car. Although it could easily drive over them, the car believed they were immovable objects, and remained standing.

These were the most successful vehicles.

Team CajunBot ran into a concrete barrier right at the start. Same for a high school team from Palos Verdes. Team ENSCO drove off the road and rolled over, after only Β 200 meters.

One of the more odd competitors, a self-driving motorcycle, fell over as soon as it drove away. They had forgotten to turn on the gyroscope. (It had been built by Anthony Lewandowski, who became a well-known name when his self-driving truck company, Otto, was sued by Google, his former employer.)

Ghost Rider.

It turned out that Sandstorm's 11 kilometers was the best any vehicle managed to perform. At the finish, DARPA's manager and a large group of journalists were waiting. What do you say when no one has even come close to winning?

Well, next year we will do it again, with twice the prize money.

New attempt in the desert - with success

Prior to DARPA's second Grand Challenge, a new team had joined.

Sebastian Thrun also had a background at Carnegie Mellon, but he was now at Stanford, leading the university's AI lab.

Unlike Red Team, Thrun and his team saw self-driving vehicles as essentially a software problem.

Red Team was more focused on the hardware and spent considerable time making it work optimally. Thrun and his Stanford Racing Team bought products on the market and spent most of their time trying to get the computer software to work optimally.

One day, Google Founder, Larry Page suddenly knocked on the door of Sebastian Thrun's office. He had heard that they would compete in the competition and had countless questions about how close a self-driving car was to becoming a reality. Both he and his co-founder, Sergey Brin intended to come and watch the competition. Little did Thrun know then that that visit would change his life.

In October 2005 it was time for competition number two. This time Red Team had two cars in the competition and both actually reached the goal. So did three other vehicles, one of which belonged to the Stanford Racing Team. Stanford's car rolled over the finish line after seven hours, thus beating the Red Team's two cars, which finished second and third.

Sebastian Thrun and his focus on software had paid off, and DARPA awarded them two million dollars.

Next challenge: An urban environment full of moving objects

DARPA was happy, but not completely satisfied. The competitors had made great progress, but driving on a desert road with fixed obstacles does not mean that the cars would manage in many real-world environments.

Therefore, a third competition, the DARPA Urban Challenge, was announced.

Now the contestants would need to navigate in an urban environment, full of moving objects and other vehicles and manage to drive 10 miles in under six hours, and at the same time follow all traffic rules.

Red Team decided to run again and now started a collaboration with General Motors. Their car became a Chevy Tahoe named, Boss. Thrun's team also ran again, sponsored by Google. Brin and Page reappeared on race day and this time had a large part of Google's leadership with them.

Again, Carnegie Mellon's Red Team (now Tartan Racing) and Thrun's Stanford Racing were favorites. At the start, suddenly the GPS receiver on Red Team's vehicle, Boss, stopped working. The competition team let another team start, but the people from Carnegie Mellon failed to make it work. Another team started, and the panic began to spread.

DARPA's boss, Tony Tether, looked at the situation, looked up from Boss at the giant Jumbotron just behind them.

"Turn that off," he said. Shortly thereafter, the GPS receiver began to function and Boss was able to roll away a few minutes later.

The first car to roll over the finish line was Stanford's Junior, at 4 hours and 29 minutes. Boss was still out on the track, but had also started. When it crossed the finish line, it did so in 4 hours and 10 minutes. However, that alone was not enough for victory; the cars also had to follow the traffic rules. Results were reviewed overnight and in the morning DARPA Chief Tether stepped onto the stage to present the result.

In third place was a team from Virginia Tech, which won $ 500,000. The prize money for second place was a million dollars and went to Stanford Racing and Sebastian Thrun. Thus, Boss and Red Team from Carnegie Mellon had won in their third attempt.

Now everyone started investing in self-driving cars? Not really...

In just three years, the development had gone from the cars not being able to drive more than a few kilometers on a desert road, to being able to navigate 10 km in a city environment with moving obstacles. Now, of course, all the major car companies threw themselves after the winning teams and invested billions of dollars into this revolutionizing technology.

Well, not really.

The car companies showed no interest whatsoever. Not even GM, who had sponsored the winning team. Instead, it was Google who took the lead.

Larry Page was the driving force behind the decision. To Sebastian Thrun, who him recruited to lead the effort, said: β€œIf successful, this will be bigger than Google. Even if the chance of success is only ten percent, it's worth it.”

Page and Brin designed their own challenge, and aptly named it Larry1K.

The car would manage ten different driving environments around California, each challenge 1,000 miles. Urban environments, bridges, highways, small roads, and so on. In addition, the team itself added a challenge that the cars would be able to drive 100,000 miles all by themselves. For this challenge they were given two years.

In 2010, the New York Times revealed the secret project, and three months before the deadline, they had passed all the challenges.

Today we know the project as Waymo, an independent company in the Alphabet group.

GM's development manager Larry Burns, who was behind GM's sponsorship of Boss, writes in the book Autonomy, that Google as a software company understood better than most how close self-driving vehicles were. While GM, a hardware company, despite Boss's victory, saw self-driving vehicles as science fiction. Larry Burns was recruited into Google and in 2009 came to work on their self-driving vehicle program.

It should be added, in the car companies' defense, that in 2007 and 2008 they were fully engaged in surviving the financial crisis. Both GM and Chrysler went bankrupt and had to go through reconstruction.

Now all car companies invest in self-driving

It is interesting to reflect on that time, during 2008 and 2009. It was not only the starting point for the self-driving revolution, but also for two other perhaps equally important developments.

In 2008 the Tesla Roadster was launched, the first series-produced electric car to use lithium-ion batteries.

And the following year, Uber was launched.

None of these three major innovations - self-driving, electricity, and transport-as-a-service - came from any of the large established car companies.

Now that people claim that "nothing has happened", it is worth remembering what a huge opportunity the car companies missed ten years ago. Everyone is investing in this development now, but if they had already done so, they probably would have been in the lead today.

From DARPA's three competitions came a whole community of engineers and computer scientists with extensive knowledge of self-driving vehicles. With relatively little money, a car company could have hired these people. Instead, they are now investing $80 billion just to catch up. A pretty nice ROI for the $35 million DARPA spent.

A dramatic change in the transport sector - and in society

It is difficult to come up with a product that has had a greater impact on society during the 20th century than the car.

It has shaped the way we live and work.

The oil that fuels them has made poor countries rich and powerful. Several wars have been fought over the black gold.

The car has also had a great impact on the environment in the form of emissions and dangerous particles, not to mention climate change.

Cars have given us freedom. Freedom to move people and goods over long distances for a low cost, which has created economic growth, jobs, and development.

At the same time, when we look back to the early 2000s from the future, we will be fascinated by these cars. How big, bulky and ineffective they were. Today's children will probably never drive a car, except possibly as entertainment inside a fenced area.

The car of today will be like the horse. Once upon a time the most efficient means of transportation but in the future just something you use for fun.

The pessimists cross their arms and say "look, nothing has happened". The fact-baseed optimists see how far we have already come, but of course also how complicated it is to reach fully self-driving cars.

But the change in the car's appearance, business model and function is already underway and it represents a great opportunity to change society for the better.

Those who think nothing has happened will be left in the dust from the self-driving electric cars, quickly accelerating away, when you press a button.

Check out our article of how even a few self-driving cars in the roads will reduce traffic jams.

This is a premium article, open for everyone thanks to our premium supporters.