Research on URBan road traffic In challenging environments – With the advancement of autonomous driving technology, the integration of autonomous vehicles into everyday life is becoming increasingly prevalent. To ensure the proper functioning of autonomous vehicles, it is necessary to equip them with various sensors that can perceive the surrounding environment and transmit the gathered information to the vehicle’s computer system. However, in challenging environmental conditions, the image data captured by these sensors may contain significant amounts of noise, which can potentially compromise the safety of the vehicle’s operation.

This research project aims to employ computer vision techniques to decouple images into domain related components and domain independent components. By utilizing domain-independent components, which are not influenced by environmental factors, for target recognition and localization, the impact of environmental conditions can be minimized. Ultimately, the proposed system will employ multi-task learning and effectively utilize the diverse range of collected information to guide the autonomous driving of the vehicle.

The project fosters collaboration between TU Wien and Fudan University (School of Computer Science), China.

Funding

This project receives funding through Austria’s Agency for Education and Internationalisation (OAED) as part of its Eurasia-Pacific Uninet program.