Chinese scientists have unveiled a new artificial intelligence system that could revolutionize how drone swarms identify and destroy enemy targets on the battlefield. Traditional drone operations face significant challenges because each drone in a swarm can monitor only a limited area, requiring constant data sharing to build a complete operational picture. This dependence becomes a vulnerability when enemy jamming disrupts the drone network, temporarily causing targets to disappear from operators’ view.
The new algorithm, called Heterogeneous Graph Spatio-Temporal Reasoning (HG-STR), enables drones to make informed decisions even with incomplete information while maintaining coordinated swarm activity. According to a study published in Acta Aeronautica et Astronautica Sinica, the system allows a swarm to explore large areas and neutralize every target, even when communications are unavailable.
HG-STR uses a heterogeneous graph approach, mapping each object in the operational area onto a network in which every connection carries specific, meaningful data. Each drone acts as a node, sharing its position, speed, ammunition status, and previous assignments. Enemy targets are also represented as nodes, with details about their locations and the remaining damage required to destroy them. Environmental data, such as unexplored terrain, also forms part of this interconnected information network.
To further improve operational efficiency, the researchers developed a compressed memory system that retains each drone’s past observations. This allows drones to continue making intelligent decisions during communication failures instead of starting from scratch. According to researchers at Northwestern Polytechnical University in Xi’an, tests indicate that HG-STR can achieve a 100% target-destruction rate.
In a simulated exercise, a dozen drones successfully detected and destroyed all targets across a 100 km × 100 km area while optimizing flight paths and reducing travel distance. The system can make decisions in 6.6 milliseconds, which is critical for high-stakes battlefield operations. Unlike existing autonomous algorithms that treat all information equally—friendly forces, enemies, and terrain—the HG-STR framework distinguishes between different data types, minimizing confusion and improving operational accuracy.
The researchers believe this breakthrough could pave the way for fully autonomous drone fleets capable of carrying out complex combat missions without human command, following a single instruction to find and destroy all enemy targets.