SKT and Samsung use AI to optimise 5G base stations

AI-RAN Parameter Recommender – as it’s called – automatically recommends optimal parameters for each base station environment, we’re told in a Google translated press release.

The pitch is that mobile base stations are affected by different wireless environments depending on their geographical location and surroundings, and so there can be ‘significant differences’ in the quality of 5G mobile communication services in different areas using the same equipment. This technology ‘can be used to draw out the potential performance of 5G base stations and improve customer experience quality.’

They used deep learning to predict various wireless environments and service characteristics which then automatically derives optimal parameters for improving quality.

The firms also plan to set up systems that automatically adjust the output of base station radio waves or ‘reset the range of radio retransmission allowance’ when radio signals are weak or data transmission errors occur due to interference.

“This is a meaningful achievement that has confirmed that the potential performance of individual base stations can be maximized by incorporating AI,” said Ryu Tak-ki, head of SK Telecom’s infrastructure technology division. “We will accelerate the evolution into an AI-Native Network that provides differentiated customer experiences through the convergence of telecommunications and AI technologies.”

Choi Sung-hyun, head of the advanced development team at Samsung Electronics’ network business division added:  “AI is a key technology for innovation in various industrial fields, and it is also playing a decisive role in the evolution to next-generation networks. Samsung Electronics will continue to take the lead in developing intelligent and automated technologies for AI-based next-generation networks.”

Going forward, the firms plans to expand the scope of what can be optimised with AI, such as parameters related to future beamforming, and ‘developing real-time application functions.’