How to control a swarm of hundreds of combat robots?

How can a swarm of hundreds of combat robots be controlled?

Image: Darpa

Pentagon research agency wants to be able to monitor and control hundreds of unmanned systems simultaneously in real time

The dream of military strategists is not only the deployment of autonomous systems that can autonomously execute commands and thus replace soldiers. Of course, this is only attractive as long as the opponent does not also have autonomous systems, so that not only robots fight among themselves, but they also kill people on both sides or destroy targets or paralyze them through cyberwarfare. It has been enough for adversaries to sneak armed drones into U.S. or German airspace in swarms, overwhelming air defenses by sheer volume.

But this could still last a little, before unmanned and remote-controlled combat robots come in the air, on the ground, in the water, in the universe and, naturally, also virtually in the Cyberspace. Not only are such robots being feverishly developed, but also the possibility of using them in swarms, for example, to overcome an opponent by numbers alone or to better accomplish a task together. The problem is how to control such a swarm of robots so that they can act differently to deal with a problem more effectively. Humans were concerned that if the swarm did not act automatically or in a different way, the problem would be solved. coordinated by AI programs, mentally overwhelmed, because this would require either an extreme multitasking ability or an extremely fast, averbal coordination of several pilots.

Darpa, the Pentagon’s research agency, is of course aware of this problem, which arises with the development of robot swarms. Whoever has a quicker solution here will first restart the arms race, as has already been the case with the development of drones. Now Darpa has launched the OFFSET (OFFensive Swarm-Enabled Tactics) program to enable soldiers to control robotic swarms.

The target seems to be urban combat, which in the last 20 years has become one of the core areas for military intervention in asymmetric wars. Just now it concerns also again the conquest of coarser cities such as Mosul and Raqqa, military strategists want to prepare also for employments in incalculable megacities, in which new conflict areas are amed, which require new strategies, if one does not want to describe them simply with artillery and bombard from airplanes and thereby, see Mosul, Fallujah, Ramadi etc.., hits the civilian population and destroys the city and its infrastructure. Megacities can’t really be cordoned off, either; units have to penetrate them and find and fight the enemy in the obscure jungle of enclosed space, which also extends into the depths, without becoming a target themselves.

How can a swarm of hundreds of combat robots be controlled?

Picture: Darpa

Darpa also writes that "urban canyons, with their tall vertical structures, confined spaces, and limited visibility, constrained military communications, mobility, and tactics, even in the best of times". This was going to be even more difficult if U.S. forces did not control the areas and were not "on supply chains, infrastructure, and prior knowledge of local conditions and potential hazards" can leave. While drones and unmanned vehicles had proven helpful in such difficult urban environments, they were just more effective when they could be deployed in swarms. The robots themselves are not the problem, but the ability of soldiers to control these swarms in urban areas is.

OFFSET is intended to stimulate the development of swarming techniques that can control 100 or more robots "increasingly autonomous robots" can be controlled. With that "bring new scalable, dynamic capabilities such as distributed perception, robust and resilient communications, distributed computing and analytics, and adaptive collective behavioral structures to the battlefield", according to Timothy Chung of Darpa.

The important thing is immersive and interactive human-machine interfaces, which allow users to "Be able to monitor and control hundreds of unmanned systems simultaneously in real time". Desired "Grammar of swarm interaction", such as already exists in soccer computer games, where pre-programmed plays are combined with free moves to allow dynamic actions and reactions in real time. Simulations are desired to first see what trains could work best in the real world. In addition, apps should be created that can be optimized by a community of developers. Overall, the aim is to have swarm systems with which "urban opponents" outsmart and outperform the market in terms of technology and performance.

Leave a Reply

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