A Chinese research team has developed an AI algorithm designed to autonomously coordinate drone swarms that can continue hunting targets even under communications jamming and severely degraded visibility.

The system, called Heterogeneous Graph Spatio-Temporal Reasoning (HG-STR), was developed by a team led by Zhang Dong of the School of Astronautics at Northwestern Polytechnical University in Xi’an. 

Current drone swarm systems typically rely on stable communications links and relatively simple coordination protocols. If radio links are disrupted or sensors are degraded, their performance can rapidly deteriorate. HG-STR is reportedly designed to address that vulnerability.

China Reports Simulation of Autonomous Drone Swarm Algorithm Capable of Independent Target Selection

Researchers in China have presented a simulated system, reportedly called HG-STR, that enables coordinated drone swarms to operate with minimal or no human intervention in… pic.twitter.com/SsxxQxeN6O

— Washington Eye (@washington_EY) May 30, 2026

The system reportedly operates through a “heterogeneous graph,” a structured network in which each element is assigned a distinct role and meaning. In this framework, friendly drones are treated as one type of node, search areas as another, and potential targets as a separate category, allowing the model to better distinguish and interpret complex operational relationships.

Traditional swarm algorithms often process all incoming information, friendly units, enemy forces, and terrain, as largely uniform data. According to the research team, this can reduce clarity in complex environments and limit performance under dynamic battlefield conditions.

According to the researchers, the system is capable of achieving a “100 percent kill rate” in simulations while operating fast enough to keep pace with modern combat conditions.

“This technology suggests a future where swarms of drones could be sent into a high-risk, jammed environment, cut off from human command with a single final order: find and kill them all,” a Beijing-based defense expert told South China Morning Post. 

The researchers said they are now looking to move the algorithm beyond simulation and into real-world testing. They also aim to further strengthen its resilience, enabling it to cope not only with lost communications, but also delayed signals and corrupted data.

“Future work will focus on exploring the lightweight deployment and real-flight verification of this algorithm on embedded airborne platforms with limited computing power,” they wrote.