The Autonomous Driving Lab (ADL) – Tetiana (Ukraine)
This is a post that we are doing as a part of a series covering different work and research opportunities in Tartu. Stay tuned for more!
Are you interested in Autonomous Driving and all news about Tesla or Waymo makes you very excited? If you want to learn more about that, then this blog post is for you! You will read more about what you can do in UT to get involved more in the field of self-driving.
Autonomous Driving Lab (ADL) was established in 2019 in collaboration with Bolt. The aim of ADL is to conduct experimental research using existing technological solutions in autonomous driving and develop new adjusted technology on that basis. To pique your interest, check out the video below! Isn’t that cool? The team managed to get these results in one year!
Now, let’s talk more about teams. Six research groups from the Institute of Computer Science and Institute of Technology of the University of Tartu work on this project. Each team focuses on different aspects of autonomous driving such as mapping and localization, end-to-end driving, security, human-vehicle interaction and more. In turn, I am working on Safety in Autonomous Driving. I ended up in the project as an Industrial Master Student. By the way, I already talked about that before in this blog post.
In the Safety team, we assess risks related to autonomous driving. One of the most important sensors a car needs to perceive the world is a camera. Cameras work as the eyes of the car. Camera images are passed to different neural networks (object detectors) to get information on what is in the image and where it is located. Let’s say there is something on the road. Object detector looks at the image and says that it’s 83% sure that this is some plastic bag and 17% sure it is a rock . To play safe, the right thing to do is to decrease speed and try to avoid this mysterious object. We try out different object detectors and analyze the cases where they fail or produce some unexpected results. Based on this, we are creating a new evaluation metric for object detectors.
There are several ways to get involved in the project. Let’s list them down!
- You can write a thesis with someone from the project. For this, you can open the thesis database and just find theses related to ADL (look at the column “Organization”).
- You can directly contact people from the project, express your interest, say how you can contribute to the project. List of contact can be found on ADL’s website.
When I started working on the project, I had no experience at all in autonomous driving. All I knew is that self-driving cars drive without a human driver! However, I was so excited about the idea of autonomous driving which pushed me to deepen into the topic. So, if you are in doubt about your skills, don’t be. What you need is passion about the topic and some wit!
But what if you are not completely sure that you want to work on the project but still want to try your hand in autonomous driving? For this, you can take one of the following courses:
- Introduction to Autonomous Driving (LTAT.06.011). Here you will be introduced to the basics of autonomous driving. This course covers all necessary topics such as sensing and perception, sensor fusion and interaction and behavior modelling. You can find more about course here.
- Autonomous Vehicles Project (LTAT.06.012). This course is more practical. You can choose a track you are interested in: research, DeltaX or ADL. More about the course can be found here.
Pictures from the DeltaX competition 2021 (Photo credits: Thamara Luup)
Additionally, on 21 April, UT and Bolt signed a significant five-year agreement to expand collaboration in the development of autonomous vehicle technology. It means that the research group will become bigger and more opportunities will be created for students!
If after reading this post, you are ready to dive into the field of autonomous driving, then don’t waste your time, get on board and make your academic year more fun and remarkable!
More information about ADL can be found here.