In the interdisciplinary research lab "Robotics & Perception Group", researchers from the University of Zurich and ETH Zurich are working together on autonomous drones and are asking themselves: How do drones find their way in an environment without external help (such as GPS)? And how can they fly independently?
There is room for progress in many areas. This project was developed in cooperation with Leica Geosystems, iniVation AG and Switzerland Innovation Park Zurich in Dübendorf. The researchers are working on advancing the flight speed of drones. Higher speeds allow drones to work more efficiently and quickly, but make it difficult for them to estimate how they are moving.
In Hangar 3 of the Zurich Innovation Park, the team has set up a "race track", which is flown by a human pilot with a drone equipped with various sensors. Among the sensors are also new so-called DVS cameras, which are capable of detecting far faster movements than conventional cameras.
The resulting datasets will be made publicly available. These will be the fastest data sets of this kind, and will enable researchers worldwide to develop new algorithms for faster autonomous drones and to compare them with each other.
Robotics & Perception Group, Universität und ETH Zürich:
Dr. Jeffrey Delmerico, Titus Cieslewski, Julien Kohler, Dr. Matthias Fässler, Prof. Dr. Davide Scaramuzza