Lost and injured hikers are a particularly big problem for Switzerland, whose emergency responders receive more than a thousand calls for help a year, but a group of researchers from the Dalle Molle Institute for Artificial Intelligence and the University of Zurich has developed a method for common drones to navigate deeply forested trails in search of missing persons.
It’s easier said than done, as the complicated backdrop of vegetation combined with fluctuating light and shadows can throw sensors for a loop.
“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, said in a news release. “Sometimes even humans struggle to find the trail.”
The scientists strapped three GoPros to the heads of hikers to capture more than 20,000 multi-angle shots, then programmed their drones using a deep neural network designed for pattern recognition and learning by training, similar to a brain. It was by any measure a success: The drone figured out the direction of the trail 85 percent of the time, while humans could only manage 82 percent.
Professor Luca Maria Gambardella, director of the Dalle Molle Institute, said, “Many technological issues must be overcome before the most ambitious applications can become a reality. But small flying robots are incredibly versatile, and the field is advancing at an unseen pace. One day robots will work side by side with human rescuers to make our lives safer.”
His colleague from the University of Zurich, Davide Scaramuzza, added, “Now that our drones have learned to recognize and follow forest trails, we must teach them to recognize humans.”