Currently commercially available drones can be extremely useful tools, when it comes to agricultural monitoring or even rescue missions. However, in the future drones could be used on rescue missions inside buildings, caves and other objects but for that they must be able to orient themselves very quickly even in the dark. Now scientists from the University of Zurich have taught drones to fly quickly in a very low-light environment.
Currently available drones can be very quick, but in terms of autonomous flying they must be in a nice, light-rich environment. Even when drones are flying with pilot’s guidance, light is always a concern. Their cameras are usually not meant to be used in darkness, which reduces drone’s speed tremendously.
Professional drones use special scanners, but they are very expensive. Scientists from Switzerland thought of a solution – a drone camera, which is inspired by a human eye and can thus be used at high speed at low-light environment.
These cameras can also allow the drones to infer their position and orientation in space. Now GPS is usually used for that, but it doesn’t work so well indoors. The new camera, inspired by a human retina, doesn’t need a lot of light to see the full picture – they monitor changes in brightness for each pixel.
This allows them to create a sharp image of their environment even at low-light and high-speed situations. New software was also developed to make sense of these inputs.
These drones will be usable in a variety of applications, but for now, scientists are thinking about search and rescue missions, when conventional drones cannot be used. Such conditions include missions at dusk or dawn as well as situations requiring high-speed flying.
For now scientists have tested only the early prototype of the camera. It is working well, but it is working indoors only and a lot has to be done before such drone could be deployed to an actual mission. Camera itself has to be improved and software needs to be tested outdoors. Davide Scaramuzza, leader of the research team, said: “We think this is achievable, however, and our recent work has already demonstrated that combining a standard camera with an event-based camera improves the accuracy and reliability of the system”.
Autonomous drones would be extremely helpful. They could search for lost people on their own as well as monitor situation after natural disasters. However, before they can be deployed some questions, including flight at dark as well as high-speed operations, have to be solved.
Source: University of Zurich