The birds represent a considerable danger to aircraft movements. But experts California Institute of technology (Caltech) decided to tackle this problem and developed a new algorithm that allows one drone to take out a whole flock of birds from the airspace of the airport. The algorithm presented in the study in the framework of the IEEE Transactions on Robotics.

The project was inspired by the “Miracle on the Hudson” in 2009 when the pilots of flight US Airways 1549 Sullenberger Chesley

(Chesley Sullenberger) and Jeffrey Scalzo (Jeffrey Skiles) had to make an emergency landing on water of the Hudson river in new York after takeoff the airliner Airbus A320-214 collided with a flock of canadian geese and both engines failed.

“Passengers of flight 1549 were saved only because the pilots were very skilled,” said associate Professor of aerospace and Department of development and applied Sciences Caltech soon-Jo Chung (Soon-Jo Chung), and principal investigator of the project of the use of drones in the management of flocks of birds. — It made me think that next time it might end tragically. So I started to look for ways to protect the air space from birds using their research experience in the field of autonomy and robotics”.

The existing strategy for the management of airspace include changes to the surrounding space to make it less attractive for birds, the use of trained falcons or even manned drones to scare off the pack. But Mr. Chang believes that these strategies can be costly or — in the case of manually operated drones are unreliable.

“When you are pushing back the birds of the air space, you must be very careful in positioning the UAV. If the drone is too far from the pack, it will not be her to drive away. If he is too close, the driver runs the risk to disperse the pack and make it completely unmanageable. This is a very difficult task for a manned flight,” said the head.

Technology researchers relies on the ability to control the swarm as a single organism, holding it together and changing the direction of movement. Each bird in the flock responds to the change in the behaviour of neighbouring individuals. Therefore, the drone moves to on the edge of flocks of birds changed course and gave the alarm next of kin, and the — next and so on until the entire flock does not change course. The positioning of the drone must be precise: if the external threat becomes too active and rushes to the flock, the birds will panic and act individually and not collectively.

Work in this direction, Mr. Chang began back in 2013 when I was going to create a homing flapping wings of the robot, whose mission is to mimic would be the Falcon. Although work in this direction allowed to create a completely new type of drone — a Bat-Bot, which the researcher introduced in 2017, he also found that the standard quadcopter can be just as effective to scare birds away.

Drone Bat Bot is really similar to a bat

Soon-Jo Chung with colleagues, including his former graduate student of Paranjape Aditya (Aditya Paranjape) from Imperial College London created a mathematical model of the dynamics of flocks of birds to describe how they are formed, maintained and respond to threats around the edges, as they convey information about it inside the pack. Their work is the development of algorithms that are intended for grazing sheep, with the only amendment that we are talking about three dimensions, not two.

Creating a good mathematical model of the behavior of flocks of birds, the researchers used it to understand reactions to the approach of external threats, and then applied this information to create a new algorithm for calculating the ideal trajectory of the drone to change the course of the pack in the right direction, not dissipating it. Tests in the field showed that a single drone is able to drive away flocks of a dozen birds. The efficiency of the algorithm is limited only by the size and number of birds. In the future, the researchers intend to create an algorithm that takes into account many drones and a few flocks of birds.


Leave comment

Your email address will not be published. Required fields are marked with *.