Japan’s KDDI boosts mobile network with Ericsson AI

Photo by Launde Morel

KDDI has partnered with Ericsson  to implement an AI-based solution to optimise its LTE radio access network and improve end-user experience.

The solution, which is part of Ericsson’s managed services offering, automatically finds optimal radio network parameters in the network resulting in increased spectrum efficiency between bands and improved throughput.

The collaboration is part of KDDI’s AI and automation program that seeks to find innovative ways to modernise optimisation and provide better performance to the operator’s customers.

In a series of trials conducted in dense urban areas, the Ericsson AI-based solution improved network throughput by nearly 10%. Normally, KDDI would have needed thousands of hours to complete a network-wide analysis to come up with optimised parameters but with Ericsson’s solution, the process now takes less than 30 minutes to complete.

With the outcome exceeding expectations, KDDI and Ericsson are now expanding this approach several additional geographical areas in Japan.

“We have identified network performance improvement as a critical factor for delivering a great network experience to our customers,” said Yasuo Ohkouchi, general manager of KDDI’s mobile network technical development division.

“However, today we’re spending a huge amount of time to do this,” Ohkouchi said. “Using the machine learning-based methodology provided by Ericsson, we see that network optimisation, that takes each area’s specific attributes into consideration, can be done in a much shorter about of time.”

Chris Houghton, head of Northeast Asia market at Ericsson, said their AI-based network optimisation is a technology designed to improve network performance with completely new levels of speed, scalability and effective results.

In 2018, Ericsson established Global Artificial Intelligence Accelerator labs to develop concrete use cases and test them in live network environments. Ericsson is further enhancing its AI capabilities to support customers’ automation efforts in the 5G era.