A simple solution for frequency sharing

As the online population increases, there is likely to be a rise in the demand for high-speed data, accompanied by increasing competition for limited frequency bands. This could risk the performance of both. Now, a practical framework has been developed by an international team led by Mohamed-Slim Alouini that shows how to improve terrestrial network performance when satellite networks interfere with the data rates.^[1]

The 28-GHz band is crucial for the 5G network, the fastest currently available, but it is also licensed to Earth stations that communicate with satellites. The result is spectrum co-existence, in which terrestrial and satellite networks interfere with each other. The interference compromises data transmission by cellular users in the 28-GHz band within a specific area of the Earth stations, an area referred to as the exclusion zone.

“Our approach aimed to find the optimal size of the exclusion zones that maximizes the speed of terrestrial networks for all users,” says Aniq Ur Rahman, a Ph.D. student at the University of Oxford who earned his master’s degree at KAUST while working on this project. “This is a computationally exhausting problem that will only become more complicated as more users come online.”

To simplify the problem, Rahman, Alouini and Mustafa Kishk, who worked on the project as a postdoc at KAUST and is now an assistant professor at Maynooth University in Ireland, applied stochastic geometry to define data rates in the 28-GHz band as functions of the sizes of the exclusion zones of Earth stations and of backhaul points. Backhaul points serve as routing points in terrestrial networks.

This approach transforms the discrete problem of frequency allocation into a continuous one, changing the problem from nearly impossible computationally to merely expensive computationally.

To simplify the problem further, they approximated the continuous expressions using the sigmoid function, a mathematical expression that high school students study. This innovation streamlines the optimization process, making it faster and significantly less complex when compared to traditional combinatorial techniques.

With this practically real-time solution, cellphone users can experience improved data speeds of up to 30 percent.

“When we realized the computational efficiency of our method, we were surprised that this had not been presented before,” says Kishk, “It simplifies a problem that will persist in the industry as we deal with more users.”