Description of the organization
The University of Zurich (UZH) is Switzerland’s largest university with more than 26 000 enrolled student, 640 professors and 3460 researchers. UZH is made up of seven faculties covering more than 100 different subject areas. As a member of the “League of European Research Universities”, UZH belongs to Europe’s most prestigious research institutions has received numerous distinctions for its contributions in the fields of medicine, economics and neuroscience.
UZH will be involved in the project through the Robotics and Perception Group (RPG). RPG consists of approximately 16 PhDs, postdocs and engineers that work on developing autonomous machines that can navigate themselves using only onboard cameras, without relying on external infrastructure such as GPS. RPG has developed numerous cognition algorithms ranging from model-based visual odometry and mapping algorithms to end-to-end learning using deep neural networks for autonomous navigation of aerial robots through streets. RPG has also pioneered the use of event cameras for robust and high speed navigation in challenging scenarios such as low light or high dynamic range scenes.
RPG will use its experience in aerial robotics, visual navigation and deep learning to contribute technical work to the design of a cognitive aerial systems architecture (WP2) and to the development of cognitive functionalities (WP3). RPG has a track record of developing vision-based cognition algorithms and applying them on real systems, which it has recently proven with its approach of combining a learning-based cognitive system with optimal control methods to win the 2018 IROS Autonomous Drone Race.
Involved Staffs
Davide Scaramuzza, Javier Hidalgo-Carrió
Role in the AERIAL-CORE project and main tasks
UZH will be the leader of WP3 providing strong expertise in cognitive functionalities for aerial robots and particularly for cluttered and complex environments. In addition, UZH will also be deeply involved in other work packages including WP2 for the design of the architecture and interfaces of the cognitive aerial system.