Zurich Team Uses Advanced Technology to Win the First Formula Student Driverless Race
Do you remember the projects you made in college? They certainly weren't as cool as the self-driving car, built by engineering students at the Swiss Federal Institute of Technology in Zurich.
The first annual Formula Student Competition for driverless cars was recently held in Hockenheimring, Germany. Over a dozen teams competed, but it was the students from Zurich that had the technology needed to win.
Flüela, the team's car, is named after a Swiss mountain pass. The vehicle has seen competition before, when it entered the 2015 Formula Student Electric race.
According to IEEE Spectrum, Flüela is equipped with both lidar and optical sensors, as well as several other ways of measuring its ground Velocity. For navigation, the autonomous hot rod can either follow a pre-loaded map or build a map of its own after a few laps.
Challenges faced on the track
To build its own map, the car uses lidar to identify obstacles such as traffic cones. Unfortunately, lidar doesn't recognize colors, which is how the cones on the left and right sides of the track were distinguished.
The manual-map making still wasn't working as well as the team had hoped, so they activated a "discovery mode." Discovery mode involved an optical camera system capable of detecting cone colors. This allowed the car to map cones itself while driving. The team says the algorithm was sourced from an unrelated drone.
Just as important, was helping the car find its position relative to the map it was building. To accomplish this, the vehicle had to be able to accurately measure speed. The team addressed this issue with wheel-speed sensors, along with an inertial measurement unit. This setup allowed vehicle speed data to be sent to the lidar, helping the vehicle to stay on track.
Speed didn't win the race
Wheel speed sensors are not the fastest measurement devices. This meant that Zurich team wasn't the swiftest on the track. They still won though, since they were the only team to complete all the laps. The second-place team froze twice after it detected cones that weren't there.
Building an autonomous race car requires a lot of trial and error. Many of the teams will return next year with new knowledge gathered during the race.
The whole thing makes you wish you were still in college, doesn't it?
- Mitsubishi Considers Vehicle-Deployed Drones
- Bird Announces Upgraded E-Scooter and New Delivery Services
- Cummins Debuts Class 6 Heavy-Duty Hybrid Truck
- AutoMobility LA: the Car Show for Autonomous and Electric Vehicles
- CityFreighter Showcases “Last Mile” Electric Delivery Truck Concept
- Mobile Device Alliance Creates Standard For Autonomous Driving Systems
- 8,000 Reservations Placed for This Tiny Electric Bubble Car
- Mercedes Concerned About Warranty Costs for Ultra-Advanced EQC SUV