12px13px15px17px
Date:12/02/16

Drones learn to search forest trails for lost people

According to the University of Zurich, every year in rescue services Switzerland receives about 1,000 calls about lost in the woods. Joint efforts of researchers from this University and also experts from the Italian Institute for the study of artificial intelligence IDSIA was created by the drone, which is able to navigate in the conditions of the forest. Quadrocopter scans the surrounding area with cameras looking for people well-trodden paths and moves strictly along them. Using such drones in the future will significantly speed up the search of a lost people.

The creation of this drone was made possible thanks to special software, which are artificial intelligence based on neural network for processing arrays of data. With two built-in cameras, the quality is no superior to the ones that are installed in modern smartphones, quadrocopter quickly focused in the forest and deftly navigates the paths while avoiding collisions with obstacles.

Armed with the latest developments in the field of artificial intelligence, a team of scientists went to a Swiss forest, where they have made tens of thousands of photos to teach future drone-rescue. Being admitted through the neural network, these photos have taught artificial intelligence how does the forest look like, and what are the paths which prefer to move people. As a result, the quadrocopter in 85% of cases successfully relied on rough terrain. Given that even the people who took part in the experiment, the trails found only in 82% of cases we can say that the machine has confidently won, and the project is extremely successful.

Before rescue services Switzerland will have at its disposal a fleet of quadrocopters-the search engines will go a long time, because scientists have work to do. But it should be recognized that this area of science is growing at a solid pace.




Views: 480

©ictnews.az. All rights reserved.

Facebook Google Favorites.Live BobrDobr Delicious Twitter Propeller Diigo Yahoo Memori MoeMesto






24 November 2024

23 11 2024