Lectures

Autonomous Driving

How a modern automated vehicle is built, from the sensor on the roof to the trajectory it drives. The course follows the functional chain the lab researches, and ends with a race in which every team’s own software has to get a small vehicle around the course.

What you learn

  • The setup of modern automated vehicles and the usual system architecture.
  • The strengths and weaknesses of camera, LiDAR, radar and GNSS/IMU, and why they are combined.
  • Perception: detecting and tracking the other road users.
  • Prediction: what those road users are going to do next.
  • Planning: turning that understanding into a trajectory that is safe and comfortable.
  • The open challenges — the ones the lab’s papers are about.

The practical part

Teams implement parts of the software stack of a small self-driving vehicle and run it in the lab. In the final session the vehicles drive a taped course with obstacles; the report of the last race is under Events.

Prerequisites

Machine Learning, and ideally Computer Vision. Python; ROS 2 can be learned during the course.