Kiwi workshop och tre elever 750x340
​Team Noa, Sanna and Elin is trying to get the Kiwi car around the track​​​
​​Photo: Lovisa Håkansson

Mission: make the car autonomous!

​“I really like cars! So, this is fun! says 15-year-old Elin, one of the many high school students who got the chance to try out how to make cars autonomous in the workshop "Get started with self-driving vehicles and AI" which took place at Chalmers during this year's edition of the International Science Festival.
It's Friday morning in May and the Science Festival is in full swing in Gothenburg. In one of the larger classrooms at Chalmers, about 30 ninth graders from Fridaskolan in Kvillebäcken are waiting for a workshop in AI and self-driving vehicles to start. The benches are placed in a U-shape and the students' attention is directed towards the small cones in yellow, blue and white that are already lined up in a kind of track curling up on the space between them.

Millie Skoglund, project assistant at the Division of Vehicle Technology and Autonomous Systems, is running today’s workshop along with Ola Benderius, associate professor at the division, and Liv Johansson, also a project assistant. Millie first came to Chalmers not even a year ago as a “Tekniksprånget” intern straight from high school. And there’s no doubt that she’s already more than comfortable in the topic autonomous vehicles.

“There’s a lot of exciting stuff here at Chalmers. But the thing that we enjoy more than anything is self-driving
vehicles! declares Millie to the students before kick starting the workshop.

A crash course in autonomous systems

First thing on the agenda: movie time! 
Scenes from a busy motorway are played on the screen, but from an unknown perspective. The students observe with curiosity. Vehicles - buses, cars and trucks - pass by at high speed. With each vehicle, a number appears. And a little further down on the screen, a diagram with a graph that moves up and down as the vehicles pass. What exactly are we watching? Millie throws the question out to her audience. No one seems willing to take a guess. 

Let’s make it a cliff hanger, she states and approaches the three objects that are placed on a table - a camera, a GPS antenna and a lidar. All sensors that self-driving vehicles need to collect data from their surroundings. The camera that can identify which vehicles are nearby, the GPS antenna that can determine the distance to oncoming vehicles and the lidar, arguably the star in the crowd - at least if you ask Millie:

“So, this one is the coolest! The lidar. It works like a radar, but it sends out millions of laser beams to be able to make very precise 3D scans of its surroundings,” she explains and goes on to revealing what the film was actually about. 

“It was a Chalmers truck equipped with sensors like these that drives every day from the harbor of Gothenburg to Borås. The numbers that popped up around the oncoming cars were a value of how certain the truck was in its assessment of what type of vehicle it encountered,” says Millie. 

The topic brings us to the next step in the process of developing self-driving vehicles: the AI part. Because it’s not enough to be able to collect data from the car’s surroundings. The autonomous vehicle must be able to understand the information, as well. Which somehow becomes the "cue" for Millie’s colleague Ola Benderius to take over. He’s a researcher focusing on self-driving systems in cars, trucks, and aircrafts.

“I develop programs that make it possible for self-driving cars to understand, interpret and make decisions based on the data that the sensors have collected. For example, if a camera on the car can detect white lines on the road, the program can make the car understand where it should drive,” he explains to the students.

Say hello to the Kiwi car


After the theoretical crash course in autonomous vehicles, it’s become high time for the students to try for themselves what it’s like to work with self-driving vehicles. The star of the show is the so-called Kiwi car. A small 3D-printed model car with black body frames and a red bumper with small glued-on eyes at the front. In the middle of the car, heaves of tangled cords in all the colors of the spectrum. And at the top, a small royal crown. The Kiwi car is part of a learning platform that Ola and his research team have been working on for several years. The purpose? To get young people to learn to program autonomous vehicles already in school.

The group is divided into smaller teams, half of which are stationed in a nearby room. Assignments are distributed. In one room, the challenge is to use a program developed for self-driving cars to get the Kiwi car to autonomously get around the track outlined with cones. But to succeed, the teams need to set the car's ability to perceive the colors of the cones correctly – making blue look like blue and white look like white - so that the car knows how to navigate among the cones. Using an iPad, the students start to pull the controls with great enthusiasm to adjust the color perception in the car's camera so that they correspond to reality.
An exciting but not entirely simple task, it should turn out.

At a table in one corner, Sanna, Elin, Noa and Carl-Johan are leaning over the iPad. They pull the controls up and down, trying to find the right levels. 

“We’re able to find the blue one but not the yellow one,” says Carl-Johan a bit frustrated.
“Yes, the yellow ones can be a bit tricky,” says Ola and tries to help the group fine-tune some more with the controls. 

The group members take turns trying to find the right color coding on the iPad. At the same time, the group - not entirely unexpectedly – start talking about the subject of self-driving cars.

“We haven’t talked that much about self-driving cars at school. But I think it’s interesting, says Elin who already has a great interest in cars. After the summer, she will start studying at a technical high school. 
“I really like cars! So, this is fun,” she says, reaching for the iPad to give it another try.

Classmate Sanna doesn’t really share Elin's passion for cars and technology, but still finds the workshop somewhat useful. In the autumn, she will go on to studying hairdressing, if everything goes according to plan.

“I'm usually not that interested in things like this. But I think it's interesting to see how the systems work. It’s most fascinating to see how cars can drive themselves,” she says and delves into a possible future scenario:

“Imagine if you’re a truck driver and the truck is self-driving. Then you can continue driving while being asleep,”
Sanna says and makes the whole group laugh.

Suddenly all eyes are turned to the cut path on the floor. One of the other groups has made the car work.

“This looks great!” exclaims Ola.

The small Kiwi car finds its way between the cones at a steady speed, completely by itself. And even though it looks promising, it soon gets into problem as it drives straight into a yellow cone. The group has no choice but to return to the drawing board. Ola tries to explain what went wrong. 

“Do you see that the image is fuzzy? It’s not completely clear. This means that the car will beware of everything,” he explains.

Can you beat the record?

In the other room, the groups are battling another task. The focus here is not on getting the car to drive by itself. It’s about getting the Kiwi car around the track with the help of hand control with human help. 

“The previous groups’ record was 18 seconds! Which is really good. Can you beat it?” Millie asks.

The teams immediately accept the challenge and throw themselves over the cars on the floor. Here, too, cones are lined up in a formation that forms a track for the cars. One in each group times with a timer clock while someone else in the group tries to steer the Kiwi car correctly, without hitting any cone. With mixed success. Cones are slightly overturned here and there, and the timer is consequently zeroed. But no one wants to be a quitter. After a few attempts and with a lot of focus - and quite a lot of laughter - some groups manage to get the car around the track in just over a minute.

And pretty soon it's time to gather all the groups and finish the workshop.
Once gathered in the classroom, Ola demonstrates what an optimal color setting looks like for the Kiwi car to perceive its surroundings in the best way. They try out together using the iPad controls while the camera view of the Kiwi car is projected on a screen. 

“So, what seems to be the problem now?” The question is directed to the class.
“The blue one!” the group agrees.
“Yes exactly! The blue color needs to be adjusted a bit. I can tell that you’re getting the core of the principle, and that’s the most important thing,” says Ola.
 
He continues to show the students pictures of real self-driving trucks and cars that, just like the Kiwi car, have been developed at Chalmers. Soon they will be tested on a nearby track. The students listen attentively, as if they’re really taking in what it means that the systems that they’ve just tested can be used on real roads. 

"It’s the future"

The workshop seems to have left an impression on the students. Even on those who may not be planning for a career in technology.

“I probably won’t work with things like this, but it's very cool. It's the future,” Noa states and leaves the Science Festival this time around.

At the same time, Ola, Millie and Liv are getting ready to receive the next group of school students.

“We believe that it can be difficult to deal with these technically complex subjects in school. Our intention is to peel off the technically difficult but still give a good insight into how the technology works. It’s important to make these subjects easily accessible, partly to educate the public, but also to attract interest for technology among students in these age groups. And that we managed to show that what the students did is relevant to real vehicles, was really good,” says Ola.

“I thought it became clear that the students got curious about how self-driving cars work. And when they got to try it out themselves and drive the Kiwi car, they became very engaged and interested,” Millie concludes.

Text: Lovisa Håkansson

Page manager Published: Thu 23 Jun 2022.