These drones are able to hunt in swarms

Evolve swarms of drone to create light animation at night is one thing, managing to associate them so that they can carry out tasks together and independently is another. During an animation, the drone are all managed by a central computer which gives them instructions to control their trajectory and the area is devoid of obstacles.

If it is a question of giving autonomy to each drone so that it can manage its task by collaborating with other drones, this requires a completely different intelligence and this must be on board. Faced with this problem, Chinese researchers from the University of Zheijang in Hangzhou have conducted impressive experiments, published in Science Robotics. They managed to make a swarm of ten drones evolve in a fairly dense forest and for which they had not programmed any data beforehand. Each drone had to move towards a common arrival point while avoiding its fellow drones and obstacles in the environment.

Drones map the environment around them and share this data in real time with each other. © Zheijang University

As impressive as it is disturbing

Those drone the size of the palm of the hand are lined with sensors and can individually process and share the data they collect to other drones in real time. They know how to maintain their safety distances to avoid collision in flight or with trees or surrounding obstacles. Even better, with all of this data collected in real time, they devise strategies for sneaking around in order to optimize the flight time between two points. So, obviously, if we combine this individual and collective intelligence with systems for tracking a target, these swarms of drones can just as easily be used to carry out rescue operations, or attacks.