日本語 
SoftBank Corp. and Ericsson Japan K.K. have started deploying a new system that uses AI to automatically control and optimize Massive MIMO base station coverage. The system is already running at several large event venues across Japan. These include big arenas and dome style facilities in the Tokyo metropolitan area. The move follows a demonstration carried out at the Osaka Kansai Expo in 2025, where the technology was tested under real world conditions.
Large venues such as stadiums, arenas, and theme parks do not have stable traffic patterns. On non event days, usage is relatively low. On event days, traffic can spike sharply. On top of that, conditions can change suddenly. Weather issues, delayed start times, or restricted entry to certain areas can all shift where users are concentrated. Traditional network control methods rely on preset schedules. Those schedules often fail when real conditions do not match expectations. This is where SoftBank and Ericsson saw a clear gap.
The new system takes a different approach. It collects user distribution data from base stations every minute. AI analyzes that data and determines whether an event is happening. Based on that judgment, it automatically adjusts both horizontal and vertical coverage of Massive MIMO base stations. The adjustments happen in near real time and do not depend on predefined assumptions.
Also Read: Dialpad Introduces Real-Time AI in Japan
At the Osaka Kansai Expo, the system was applied to outdoor base stations. When traffic changed suddenly, improvements in communication quality were observed. Average download speeds increased by about 24 percent, going from 76.9 Mbps to 95.5 Mbps. The system also helped reduce congestion issues, including short data stalls during traffic spikes.
After confirming these results, SoftBank and Ericsson moved the system into live operation at selected large scale venues. The aim is to deliver more stable and consistent 5G performance in places where crowd behavior changes constantly.
English