Neuromorphic sensing and processing for autonomous vision-based landing

Prof. Guido de Croon and Jesse Hagenaars

TU Delft

Neuromorphic sensing and processing hold great promise for autonomously flying robots, be they small drones or spacecraft. Combining event-based cameras and spiking neural networks results in energy efficient and extremely fast vision that can deal with a wide range of illuminations. In my talk, I will present our work on optic-flow-based landing. I will discuss recent fundamental insights into autonomous optic flow landing, which have ramifications for robotics and biology alike. Furthermore, I will discuss how we are working towards a fully neuromorphic sensing and processing pipeline, with as goal extremely light-weight, energy-efficient, and high-performance autonomous landing.