Topics related to mapping, localization and learned driving
Organization
Autonomous Driving Lab
Abstract
Selection of topics related to mapping, localization and learned driving:
1. Mapping
1.1. Map real-life features necessary for autonomous driving (BSc, MSc)
1.2. Build a 3D point cloud map and use it for localization (MSc)
1.3. Convert real-world spatial data into a usable format (BSc, MSc)
2. Localization
2.1. Localization using SfM point cloud (MSc)
2.2. Better positioning using smartphones (MSc)
2.3. Fallback to lane following in case of GNSS failure (MSc)
2.4. Setting up RTK base station for accurate positioning (BSc)
3. Base autonomy
3.1. Get a selected autonomous driving software stack to work with actual car (MSc)
4. Learned driving
4.1. Adapt CARLA Leaderboard for Tartu simulation (BSc, MSc)
4.2. Adapt VISTA simulation to use depth (MSc)
4.3. Adapt VISTA simulation for human driving (BSc)
4.4. Use VISTA to learn driving policy for real world (MSc)
4.5. Scaling laws for end-to-end driving (MSc)
4.6. Train end-to-end model on all of the world’s data (MSc)
1. Mapping
1.1. Map real-life features necessary for autonomous driving (BSc, MSc)
1.2. Build a 3D point cloud map and use it for localization (MSc)
1.3. Convert real-world spatial data into a usable format (BSc, MSc)
2. Localization
2.1. Localization using SfM point cloud (MSc)
2.2. Better positioning using smartphones (MSc)
2.3. Fallback to lane following in case of GNSS failure (MSc)
2.4. Setting up RTK base station for accurate positioning (BSc)
3. Base autonomy
3.1. Get a selected autonomous driving software stack to work with actual car (MSc)
4. Learned driving
4.1. Adapt CARLA Leaderboard for Tartu simulation (BSc, MSc)
4.2. Adapt VISTA simulation to use depth (MSc)
4.3. Adapt VISTA simulation for human driving (BSc)
4.4. Use VISTA to learn driving policy for real world (MSc)
4.5. Scaling laws for end-to-end driving (MSc)
4.6. Train end-to-end model on all of the world’s data (MSc)
Graduation Theses defence year
2022-2023
Supervisor
Tambet Matiisen, Edgar Sepp
Spoken language (s)
Estonian, English
Requirements for candidates
Familiarity with ROS and neural networks is beneficial in many of the projects.
Level
Bachelor, Masters
Keywords
Application of contact
Name
Tambet Matiisen
Phone
5286457
E-mail