Scientists at Reshetnev University in Krasnoyarsk have completed the first stage of creating a comprehensive system for modeling hybrid infocommunication networks, combining space, stratospheric, and ground segments. This was reported on March 17 by the university's press service, citing TASS. The development is capable of covering more than 99% of Russia's territory with communication services — including areas that ordinary cellular towers cannot reach. The project is being implemented as part of the National Technology Initiative program with the support of JSC "GLONASS VTP." Krasnoyarsk researchers have not only designed the network architecture but also created data flow management algorithms, solved the problem of frequency interference between satellites, and laid the mathematical foundation for designing next-generation telecommunication systems.
Greedy-Gradient Algorithms: Why Routing Speed is Crucial
The key technical problem of hybrid networks is the distribution of data flows between thousands of nodes in real-time. Standard library solvers found the optimal route in a network of a thousand elements in 42 seconds. The family of greedy-gradient algorithms (GGA) developed in Krasnoyarsk handles the same task in 13.8 seconds — almost three times faster, according to the university's press service.
For multi-satellite constellations, where satellites are constantly moving, and the network configuration changes every second, this difference is critical. Slow routing means packet loss and service degradation — especially for voice and video.
Frequency Conflicts: Solution via Genetic Algorithm
The second task is frequency interference. When hundreds of satellites and ground stations operate simultaneously, their signals overlap. Krasnoyarsk scientists have adapted a two-stage algorithm: first, frequency bands are distributed proportionally to the channel load, then the distribution is optimized via a genetic algorithm.
The university compares this mechanism to air traffic control separating aircraft by altitude. The methodology is especially important for scenarios where low-orbit satellites, stratospheric platforms, and ground towers operate simultaneously over the same territory.
System Specifications
- Coverage of Russia: more than 99%
- Routing speed (GGA): 13.8 seconds for a network of 1,000 elements
- Speed of standard solvers: up to 42 seconds for a similar task
- Network segments: ground, stratospheric, space
- Development stage: first stage completed
- Industrial partner: JSC «GLONASS VTP»
- Program: National Technology Initiative
What's Next
Project manager Konstantin Gaipov outlined the following steps: integrating the developed environment with existing simulators, researching neural network control of networks, and creating a hardware and software complex. The final product, according to him, will allow designing optimal networks for hard-to-reach regions — by accurately mathematical modeling, reducing the time and cost of developing new telecommunication services. The press service did not announce official deadlines for completing the next stages.
Developing a comprehensive environment for modeling hybrid networks is not just about conducting research, but also about contributing to the formation of the country's technological sovereignty in the field of telecommunications
For comparison: the largest commercial constellation of low-orbit satellites, Starlink, had more than 6,000 devices by 2025 — and did not solve the problem of integration with stratospheric platforms and ground infrastructure in a single control system. The Krasnoyarsk development was initially designed as an end-to-end environment covering all three segments.
For Russia, where about 30% of the territory is practically devoid of stable cellular coverage, hybrid networks are not an academic task. A system capable of designing optimal routes in seconds and automatically resolving frequency conflicts significantly reduces the cost of connecting Arctic regions, Siberia, and the Far East — regions critical for both logistics and state defense orders.