Lucas Ferreira, Doktorand
My project aims at developing low-power technologies that would enable insect-sized autonomous drones. The frontiers of current technology lie at the power-intensive processing devices to perform autonomous navigation. Research has shown that a mosquito-sized drone (100 mg) could achieve stable flight, and perceive its environment with 200 mW. On the other hand, current processing devices that are able to perform state-of-the-art algorithms for autonomous flight demand 10-100W, which require bulky batteries and rendering those drones currently inviable, creating research opportunities.
One example illustrating the Simultaneous Localization and Mapping (SLAM) system, which is the application that needs to be supported by my low-power processor, can be seem in the image below from Xiang et al (DOI 10.1109/ICRA48506.2021.9560900).
Based both on the drone’s camera and movement, a real-time sparse map of the environment can be synthesized, while the drone’s pose (position and orientation) is obtained. Both map and 3D-poses are crucial for traversing between map points, while avoiding obstacles, and navigating autonomously.
A video of the SLAM algorithm that will be supported by my low-power processor can be seem in the video below: