Chinese research team publishes drone swarm algorithm claiming 100 percent kill rate in tests
Chinese researchers published a peer-reviewed drone swarm algorithm, HG-STR, in Acta Aeronautica et Astronautica Sinica on May 19. The paper claims a 100 percent kill rate in tests, with fixed-wing drones autonomously coordinating under jammed communications and blocked vision.
A research team from northwestern China has published a drone swarm algorithm called HG-STR (Heterogeneous Graph Spatio-Temporal Reasoning) in the peer-reviewed journal Acta Aeronautica et Astronautica Sinica on May 19. The paper claims the algorithm achieves a 100 percent kill rate in tests, allowing fixed-wing drones to autonomously locate and engage targets even when their communications are jammed and their visual sensors are blocked. The journal is China's top aviation publication. The paper describes HG-STR as the first known approach able to maintain that level of effectiveness while operating at speeds the authors describe as compatible with modern combat tempo. The research team is based in northwestern China. Context: Chinese military-academic output on autonomous swarm warfare has been accelerating through peer-reviewed channels, not only state-media demonstrations. Earlier in 2026, Chinese state television showcased the Atlas swarm system, with a single command vehicle handling coordination for a fleet of fixed-wing drones launched from one platform. What remains open: the 100 percent figure refers to results reported in the published research, not to operational deployment against an adversary equipped with jamming countermeasures, decoys, layered radars, and active anti-drone defenses. The peer-reviewed claim and any future combat-condition validation are separate questions. Source: South China Morning Post reporting on the paper in Acta Aeronautica et Astronautica Sinica.