Chinese scientists have developed a new artificial intelligence system that could revolutionise drone warfare by enabling swarms to achieve a 100 per cent kill rate, even when communications are jammed. The algorithm, named Heterogeneous Graph Spatio-Temporal Reasoning (HG-STR), was detailed in a study published in the journal Acta Aeronautica et Astronautica Sinica.
Under normal conditions, drones in a swarm can only view a small part of the battlefield, relying on communication to build a complete picture. Enemy jamming systems can disrupt this, causing targets to appear and disappear. The new AI system allows drones to make effective decisions using incomplete information while maintaining coordination.
The HG-STR algorithm uses a 'heterogeneous graph' system where each object is mapped onto a web, with threads conveying its true meaning. Each drone becomes a node carrying data on position, speed, remaining ammunition, and previous tasks. Enemy targets and surrounding areas are also represented as nodes, with information on location, damage required, and unsearched areas.
A compressed memory system enables drones to remember past observations, so they can continue reasoning even when communication fails. In tests, a swarm of ten drones covering a 100 km by 100 km area destroyed all targets more efficiently, with decision-making taking just 6.6 milliseconds. Researchers from Northwestern Polytechnical University in Xi'an say HG-STR is the first AI algorithm with the potential to achieve a 100 per cent kill rate.
Existing autonomous drone algorithms treat all data as the same type, leading to confusion. The new system segregates information into categories, such as high-priority threats or opportunities for collaboration. This allows drones to instantly understand whom to help and whom to hunt, potentially enabling fleets to operate in environments cut off from human command.
