KT Corp, Samsung Electronics validate AI-powered autonomous wireless network technology at baseball stadium

KT Corp and Samsung Electronics have validated technology that uses AI to automatically optimize wireless networks. A trial conducted at a real baseball stadium showed traffic imbalances between base stations improved by up to 34%, while the load concentrated on specific base stations fell by as much as 16%.

KT Corp announced Monday that the two companies had validated autonomous wireless network operation technology based on the digital twin platform "AI Cell Twin" at KT Wiz Park in Suwon.

AI Cell Twin is KT Corp's network digital twin platform that virtually replicates a live wireless network in a digital environment. It allows engineers to configure mobile usage patterns and traffic changes in a simulated communications environment and test various network operation scenarios.

Building on the platform, the two companies jointly developed autonomous operation technology in which AI derives optimal network management solutions, validates them in advance through the digital twin, and then applies them to the live wireless network.

Unlike conventional network automation technology, which operates based on preset criteria or individual metrics, the new technology comprehensively analyzes a wide range of data — including user distribution, usage patterns, network conditions and perceived service quality — to determine the best operational approach.

The two companies conducted the technical trial at KT Wiz Park in Suwon over roughly three months starting in July. They built the AI Cell Twin using wireless network data collected at the actual stadium and used it to train the AI powering the autonomous operation technology.

Baseball stadiums present a particularly demanding environment: large crowds pack into a confined space, and data consumption fluctuates sharply as fans stream in and out, watch videos and share content. When data usage concentrates in a specific zone, wireless service quality can degrade.

Tests run during actual KT Wiz Park home games confirmed that traffic concentrated on specific base stations was redistributed to neighboring ones. Traffic imbalances between base stations improved by up to 34%, and the load on heavily burdened base stations fell by up to 16%.

Predictions generated by AI Cell Twin also proved highly consistent with data observed on the live wireless network.

KT Corp and Samsung Electronics plan to continue validating the autonomous wireless network operation technology during this fall's baseball season. The companies intend to expand the technology's application beyond baseball stadiums to a wider range of communications environments.

Jeong Jin-guk, executive vice president and head of Samsung Electronics' Samsung Research Next-Generation Communications Research Center, said the two companies had gone beyond automating network optimization with AI to develop technology capable of applying optimization solutions to live networks through advance validation using a digital twin. "We will continue to lead network innovation with future communications technology that enhances both network operational efficiency and the quality of service users actually experience," Jeong said.

Lee Jong-sik, executive vice president and head of KT Corp's Future Network Lab in the Network Division, said the trial had confirmed the viability of deploying AI Cell Twin-based autonomous operation technology on commercial networks. "KT Corp will continue to advance AI Cell Twin technology to respond proactively to evolving network environments and, through that, build an autonomously operated network that delivers better communications quality to customers," Lee said.

Meanwhile, KT Corp unveiled AI-based network operation technology at "AX Tech Connect 2026" in September. At the event, the company showcased next-generation network technologies including "Autonomous Operation," which independently handles network management, and "Hyper AI Network," which provides the connectivity and computing infrastructure required for the AI era.