Top Headlines:
- Chinese researchers have developed a vision-based UAV swarm perception framework that operates without GNSS or inertial navigation data.
- China reportedly directed Taiwanese opposition politicians to cut NT$470 billion from a NT$1.25 trillion special defense budget built around drones.
- Researchers at a China Electronics Technology Group subsidiary have identified 4 structural bottlenecks threatening Beijing's 2026-2030 military modernization plan.
- China is piloting a new WeChat mini-program in Shanghai and Sichuan that streamlines flight clearance for low-risk drone operations.
- Shanghai grounded all drone flights citywide for 6 days ahead of President Xi Jinping's first appearance at the World Artificial Intelligence Conference.
- More than a dozen APEC economies signed a cooperative action initiative this week to align technical standards and certification frameworks for the low-altitude economy.
- DJI Agriculture's global drone fleet passed 700,000 aircraft across more than 100 countries.
PLA-Linked Study Frees Drone Swarms From GNSS Reliance
Chinese researchers have published a new vision-based cooperative perception framework, which enables UAV swarms to map disaster sites without GNSS or IMU data while maintaining localization error to roughly 1.06m even amid heading variations of up to 270 degrees.
The system works on an anchor-wingman architecture: each wingman drone runs detection onboard and transmits only compressed feature tensors, 128 KB per aircraft, rather than raw video. Then, a Proj-IoU matching mechanism eliminates more than 50% of overlapping detections between wingmen. The whole pipeline runs at 6.7 frames per second on edge hardware light enough for a small UAV to carry.
