Scientists Train Drones to Search for Lost Hikers

AP Photo
The Associated Press

ZURICH, Switzerland, Feb. 10 (UPI) — Every year in Switzerland, emergency centers respond to more than 1,000 calls concerning injured and lost hikers. In larger countries, where more people venture into larger swaths of wilderness, the problem is even greater.

That’s why robotics engineers and computer scientists in Switzerland are teaching drones to recognize and navigate forest trails. Drones are inexpensive and can more quickly cover ground, maximizing the efficiency of search and rescue missions.

The quicker a lost or injured hiker is found, the sooner medical assistance can be delivered.

But drones aren’t yet adept and navigating through a dense array of obstacles. Toteach drones how to fly through a forest, scientists at the Dalle Molle Institute for Artificial Intelligence and the University of Zurich have developed new nagivation software.

The software’s algorithms allow the drone to process forest imagery and recognize man-made trails.

“Interpreting an image taken in a complex environment such as a forest is incredibly difficult for a computer,” Alessandro Giusti, a researcher at the Dalle Molle Institute for Artificial Intelligence, said in a news release. “Sometimes even humans struggle to find the trail!”

Giusti and his colleagues trained the drone by uploading 20,000 images of hiking trails, collected from a helmet cam taken on several treks through the Alps. Armed with the new visual data and its deep learning software, the drone was able predict the correct path of a trail 85 percent of the time — compared to a human success rate of 82 percent.

The drone’s software is constantly re-evaluating its surroundings, making decisions about the path of the trail ahead of it by analyzing its vast memory of similar trails.

“Our lab has been working on deep learning in neural networks since the early 1990s,” added Juergen Schmidhuber, science director at the Dalle Molle Institute for Artificial Intelligence. “Today I am happy to find our lab’s methods not only in numerous real-world applications such as speech recognition on smartphones, but also in lightweight robots such as drones. Robotics will see an explosion of applications of deep neural networks in coming years.”