The relatively cheap drones can be deployed in large numbers to aid rescue workers

AI drones track forest paths to help locate missing people

Drones that can recognise and follow man-made forest paths using an artificial-intelligence (AI) algorithm could be used to find missing people in secluded areas.

The drones, which were developed by researchers from three universities including the University of Zurich, are equipped with a pair of small cameras, similar to those used in smartphones.

They observe the environment through the tiny lenses and use powerful AI algorithms to interpret the images and determine man-made trails from naturally occurring ones. The software then steers the device in the corresponding direction of visible trails.

Every year, thousands of people lose their way in forests and mountain areas. In Switzerland alone, emergency centres respond to around 1,000 calls annually from injured and lost hikers.

It is hoped the new technology will help to complement and aid the work of rescue service teams. The researchers believe that because they are relatively inexpensive, they can be rapidly deployed in large numbers to cover wide geographical areas.

The drones could help to substantially reduce the response time and the risk of injury to missing persons and rescue teams alike.

“While drones flying at high altitudes are already being used commercially, drones cannot yet fly autonomously in complex environments, such as dense forests,” said Professor Davide Scaramuzza from the University of Zurich. “In these environments, any little error may result in a crash and robots need a powerful brain in order to make sense of the complex world around them.”

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

The Swiss team solved the problem using a deep neural network, a computer algorithm that learns to solve complex tasks from a set of "training examples" in similar fashion to how the human brain learns from experience.

In order to gather enough data to "train" their algorithms, the team hiked several hours along different trails in the Swiss Alps and took more than 20,000 images of trails using cameras attached to a helmet.

When tested on a new, previously unseen trail, the deep neural network was able to find the correct direction in 85 per cent of cases. By comparison, humans faced with the same task guessed correctly 82 per cent of the time.

Professor Juergen Schmidhuber, who also works at the Dalle Molle Institute, believes that with increasing research into robotics, the number of applications using deep neural networks will "explode".

The research team has warned that much work is still needed before a fully autonomous fleet will be able to swarm forests in search of missing people. The next step will be to teach the drones to recognise humans.

Drone manufacturer Yuneec recently revealed a new model in its ‘Typhoon’ series that uses Intel's RealSense technology to automatically avoid obstacles and buildings when navigating around open terrain.

Recent articles

Info Message

Our sites use cookies to support some functionality, and to collect anonymous user data.

Learn more about IET cookies and how to control them