China Telecom Lab Transmits HD Drone Video via Narrowband Satellite

The Artificial Intelligence Research Institute of China Telecom (TeleAI) has demonstrated real-time 1080p high-definition video transmission from a lightweight drone directly over a narrowband satellite link operating at under 100 kilobits per second. The system integrates an onboard satellite terminal with TeleAI's proprietary generative transmission architecture, known as GenTrans, under its wider AI Flow framework. By transitioning data transmission from raw bitstreams to semantic token streams, the system bypasses the long-standing operational trade-off between drone payload weight, flight endurance, and visual transmission quality in offline and disaster-affected environments.

Small and mid-sized unmanned aerial vehicles (UAVs) operating in deep sea, border, or disaster relief zones routinely lose terrestrial cellular coverage and lack the payload capacity required for bulky high-throughput satellite dishes. Delivering stable 1080p video through sub-100 kbps narrowband links enables standard lightweight multi-rotor drones to execute reconnaissance without sacrificing flight endurance or requiring costly airborne relay stations.

The Artificial Intelligence Research Institute of China Telecom, known as TeleAI, has achieved direct 1080p high-definition video transmission from a lightweight unmanned aerial vehicle over a narrowband satellite connection, according to a report by People's Daily. The deployment utilizes TeleAI's proprietary "AI Flow" intelligent network framework alongside an onboard generative intelligent transmission system dubbed "GenTrans." According to the institute, the technology addresses a long-standing compromise in low-altitude aerial sensing: lightweight drones operating beyond cellular and fiber coverage previously had to either carry heavy, power-hungry satellite equipment that shortened flight duration, or settle for degraded, low-resolution video feeds. According to Li Xuelong, Chief Technology Officer and Chief Scientist of China Telecom as well as Dean of TeleAI, GenTrans marks a paradigm shift from traditional "bitstream" transmission to "token stream" transmission. By completing video encoding directly onboard the drone and transmitting semantic tokens via an uplink operating at under 100 kilobits per second (kbps), the framework reconstitutes fluent 1080p video at the receiving station. Li noted that this unified transmission framework coordinates computing, communication, and storage resources, enabling task-oriented and semantic communication paradigms alongside legacy networks. To translate the concept into a deployable platform, the TeleAI research team resolved three primary engineering bottlenecks on a customized six-rotor drone. First, researchers developed an "attitude-vibration collaborative control" mechanism that dynamically aligns the aircraft's physical posture with its target orientation, stabilizing the camera payload's field of view and preserving satellite link alignment while maneuvering. Second, the team implemented cavity-level electromagnetic shielding, optimized antenna layouts, and electromagnetic seals across control links. These measures mitigate multi-radio-frequency interference between the satellite transmitter, onboard computing hardware, and standard flight control and GPS positioning receivers. Third, the hexacopter adopted a layered lightweight structural design to balance the weight of the integrated satellite payload with operational flight endurance. TeleAI stated that the successful test represents a critical deployment of intelligent network technology in the low-altitude aerial domain, freeing autonomous sensing platforms from reliance on ground-based telecommunication infrastructure. The research institute indicated plans to expand scale verification across air, space, ground, and maritime scenarios to support China's low-altitude economy initiatives.