DroNet Under Development In Zurich

A research project is underway in Zurich, Switzerland that is intended to teach drones how to navigate city streets in much the same way that birds do. The software, called DroNet, is a convolutional neural network that learns to fly and navigate by flying and navigating.

In an abstract for an article posted in IEEE Robotics and Automation Letters, the authors ... Antonio Loquercio;  Ana Isabel Maqueda;  Carlos R. Del Blanco, and;  Davide Scaramuzza ... say that DroNet "can safely drive a drone through the streets of a city. Designed as a fast 8-layers residual network, DroNet produces, for each single input image, two outputs: a steering angle, to keep the drone navigating while avoiding obstacles, and a collision probability, to let the UAV recognize dangerous situations and promptly react to them.

"But how to collect enough data in an unstructured outdoor environment, such as a city. Clearly, having an expert pilot providing training trajectories is not an option given the large amount of data required and, above all, the risk that it involves for others vehicles or pedestrians moving in the streets.

"Therefore, we propose to train a UAV from data collected by cars and bicycles, which, already integrated into urban environments, would expose other cars and pedestrians to no danger. Although trained on city streets, from the viewpoint of urban vehicles, the navigation policy learned by DroNet is highly generalizable.

"Indeed, it allows a UAV to successfully fly at relative high altitudes, and even in indoor environments, such as parking lots and corridors."

Engadget reports that the training was accomplished bo attaching GoPro cameras to cars and bicycles, as well as using video publicly available on Github.

In the early stages of the experiment, the aircraft have learned to avoid oncoming traffic, obstacles like traffic pylons and pedestrians.

(Image from YouTube video)

FMI: Original Report, ieeexplore.ieee.org/document/8264734/?reload=true