According to their hopes, this way they can remain functional even in the event of connection disruptions and can also operate in swarms.
Ukrainian startups are developing artificial intelligence systems that can control huge fleets of drones. With this, warfare enters hitherto unknown territory. In addition, it is hoped that the appearance of AI-controlled drones on the front lines will help to combat the increasing use of signal jamming by the Russians, as well as allow the operation of unmanned aerial vehicles (UAVs) in larger groups.
The development of artificial intelligence-powered drones in Ukraine is roughly divided into the following: visual systems that help identify targets and fly in drones, terrain mapping that helps with navigation, and more complex programs that can operate UAVs in connected “swarms”. One such company is Swarmer, which develops software that connects drones to networks. Decisions can be made immediately within the group, and a human intervenes only when an automated strike needs to be given the go-ahead. “We simply cannot increase the number with human pilots,” said Swarmer CEO Serhiy Kupriienko at the company’s office in Kyiv. “For a swarm of 10 or 20 drones, human control is practically impossible.”
Swarmer is one of more than 200 tech companies that have sprung up since Russia launched a war against Ukraine in 2022. In these companies, civilians with IT backgrounds develop drones and other tools to help their country take action against a much larger enemy. Kupriienko said that while human pilots have difficulty performing operations involving more than five drones, artificial intelligence can process hundreds of drones. The system, called Styx, controls a network of reconnaissance and strike drones in the air and on the ground, both small and large. Each drone can plan its own moves and predict the behavior of other drones in the swarm, he said. According to Kupriienko, in addition to enhancing operations, automation will also help protect the lives of drone pilots operating near the front line, as they are prime targets for enemy fire. The Swarmer’s technology is still under development and has only been experimentally tested on the battlefield, he added.
CEO of the Swarmer company, Serhii Kupriyenko
However, AI drone control systems will likely require a human to ensure that the system does not make mistakes when selecting targets. There are wide-ranging concerns about the ethics of weapons that exclude human judgment. The European Parliament warned in a 2020 research paper that such systems could breach international humanitarian law and lower the threshold for going to war.
AI is already being used in some of Ukraine’s long-range drone strikes, targeting military installations and oil refineries hundreds of kilometers inside Russia. These attacks sometimes involve swarms of around 20 drones. The central drones fly towards the target, but the others are tasked with knocking out or disrupting air defenses along the way. To do this, they use a form of artificial intelligence with human supervision to help spot targets or threats and plan possible routes.
The need for AI-enabled drones is becoming more urgent as both sides deploy Electronic Warfare systems that disrupt signals between pilots and drones. Especially small, cheap FPV (first person view) drones are sensitive to this. In 2023, these low-cost products became one of the main means of eliminating enemy vehicles for both sides, but their hit rate decreases as jamming increases. “We are already working with the concept that in the near future there will be no contact on the front line” between the pilot and the UAV, said Max Makarchuk, head of AI at Brave1, a defense technology center created by the Ukrainian government. rate is constantly decreasing. The hit rate of most FPV units is currently 30-50%, while for new pilots this rate may drop to 10%. According to his prediction, FPV drones will be able to achieve a hit rate of around 80%.
EW systems form an invisible jamming dome over the equipment and soldiers they protect. If the pilot loses contact with the drone, he can no longer control it and the vehicle will either fall to the ground or fly straight ahead. Automating the final stage of the drone’s flight to the target means that the drone no longer needs a pilot – thus neutralizing the disruptive effect of EW. To counter the EW threat, manufacturers including Swarmer have begun developing features that allow the drone to measure the target through its camera.
AI-enabled drones have been under development for years, but until now they were considered expensive and experimental. Even before the 2022 invasion, Russia was already developing AI-capable aerial and ground-based drones and had some successes. In Ukraine, the most important task for manufacturers is to produce a good AI targeting system for drones. This would allow them to be deployed en masse along the entire 1,000 km front line, because even now thousands of FPV drones are being used every week. Costs can be reduced by running AI programs on a Raspberry Pi. The small, inexpensive computer has become popular outside the educational sphere as well. Makarchuk said he estimates a simple targeting system that captures the shape seen by the drone’s camera would cost as little as $150 per drone.
Source: sg.hu
