Autonomous driving systems and advanced driver assistance systems require a description of their surrounding environment. Part of this description is facilitated by fusing object information from different sensors. Historically, object fusion algorithms relied on classical methods. With emergence of advanced machine learning algorithms, new methods have been developed in recent years that enable sensor fusion to be carried out by deep neural networks.
Camera and Radar are two sensor modalities commonly used in many autonomous drive applications. These two sensors complement each other since they have different failure modes. Using deep neural networks to detect objects in camera images is a relatively well explored problem. In comparison, fusing radar and camera using a deep learning based technique is a recent emergent and less explored problem. The purpose of this thesis is to explore different network architectures that facilitate radar-camera fusion.
In this master thesis project, you will focus on:
- Perform a literature study on different existing network architectures for fusing camera and radar.
- Implement, train and evaluate a deep neural net that fuses radar and camera.
- Document the results and lessons learned.
We are looking for 2 students with an interest in deep learning for autonomous driving. The following skills would be highly valuable:
- Python programming
- Machine learning
- Reading scientific papers
- Handling large datasets
Having had deep learning related courses and some hands-on experience with these methods is a plus.
Please send in individual applications with CV, motivational letter and grade transcripts.
Planned start: January 2022, with some flexibility.
Final application date: 30th of November 2021
Duration: 30 ECTS
For questions regarding the project, please contact: firstname.lastname@example.org, email@example.com