Chinese scientists have managed to accelerate satellite internet speeds up to 1 Gbps, making it 5 times faster than Elon Musk’s Starlink internet from SpaceX. This was done through a 2-watt laser through space, from a satellite parked in a stationary orbit more than 60 times higher than Starlink rivals.
Satellite laser downlinks offer high-speed internet, but their transmission is challenged by atmospheric turbulence. This natural interference scatters the laser beams, turning them into faint, blurry patches that can span several hundred meters by the time they reach the ground.
In response to this problem, a Chinese research team led by Professor Wu Jian of the Peking University of Posts and Telecommunications and Liu Chao of the Chinese Academy of Sciences introduced a solution called AO-MDR synergy. This method is designed to counteract the signal disruption caused by turbulence.
“This method effectively prevents communication quality degradation caused by extremely low signal power,” the researchers wrote in a peer-reviewed paper published on June 3 in the Chinese-language journal Acta Optica Sinica.
At an observatory in Lijiang, southwest China, researchers put their system to the test using a 1.8-metre (5.9-foot) telescope aimed at an unnamed satellite orbiting 36,705 kilometres above Earth. Inside the telescope, 357 micro-mirrors were used to reshape distorted laser light, sharply reducing wavefront distortion caused by the atmosphere.
To further enhance the signal, a device known as a multi-plane light converter (MPLC) was used to divide the incoming light, which travels through a multi-mode fibre, into eight base-mode channels. From these, the three strongest signals were identified and combined in real time using a custom-built algorithm known as “path-picking,” powered by specialized chips.
The team recorded a noticeable boost in signal strength from this method. Compared to adaptive optics (AO) alone, the combined AO+MDR approach delivered a substantial performance gain, especially at critical levels where maintaining signal reliability is most challenging.
This method not only brings performance improvements but also reduces errors, with chances of usable signals increasing from 72% to 91.1%, which is significant when it comes to high-value data transmission.