Ericsson and KDDI Successfully Field-Trial AI Uplink Optimization, Advancing Towards Autonomous Network Level 4
Ericsson and KDDI have successfully completed a large-scale field trial of Ericsson's AI-driven UIO rApp on KDDI's commercial network in Japan. This trial demonstrated significant improvements in 4G and 5G uplink performance, even under increased traffic, proving the robustness of the AI model and the potential for rApps to contribute to higher levels of network autonomy. The UIO rApp is now globally available.
📋 Article Processing Timeline
- 📰 Published: May 11, 2026 at 23:00
- 🔍 Collected: May 11, 2026 at 14:31
- 🤖 AI Analyzed: May 11, 2026 at 15:00 (28 min after Collected)
Leveraging Ericsson's rApp and domestic third-party FYRA's rApp on KDDI's commercial network, a field trial of AI-driven uplink optimization in 4G and 5G has succeeded.
Improvements in uplink performance were confirmed even under an increase of over 10% in uplink traffic, demonstrating the robustness of the AI model under sustained load.
It has been proven that rApps can contribute to higher levels of autonomous networks.
Ericsson (NASDAQ: ERIC) and KDDI have successfully completed a large-scale field trial (hereinafter, "this trial") of Ericsson's AI-driven UIO (Uplink Interference Optimizer) rApp. The UIO rApp, operating on the EIAP (Ericsson Intelligent Automation Platform), demonstrated predictable uplink performance improvements across 4G and 5G live networks, proving its readiness for greater network autonomy.
This trial was conducted in Q1 2026, targeting a multi-technology, multi-band cluster consisting of approximately 1,500 5G cells and 1,300 4G cells, to verify whether AI-based uplink interference management could bring scalable improvements under real traffic conditions. The UIO rApp selectively optimized cells carrying the majority of network traffic, achieving substantial uplink performance improvements without compromising network stability.
Key performance evaluation results in the selected cells are as follows:
- Throughput improved in both 4G (average 9.6% improvement) and 5G (average 3.1% improvement), leading to enhanced end-user experience.
- Consequently, 5G SINR (Signal-to-Interference-plus-Noise Ratio) improved by 27%.
- Uplink modulation efficiency and spectral efficiency improved for both technologies.
This trial with KDDI is the world's first instance of an AI-driven UIO rApp being deployed in a live network environment, and also the first time the EIAP platform has been used in a live environment in Japan. The UIO rApp was evaluated as achieving level 3.86 autonomy in uplink optimization, applying the TM Forum's ANLET methodology. These results clearly demonstrate the role this application, and the valuable and achievable use case of "AI-powered uplink optimization," can play in telecommunications service providers' journey towards network autonomy level 4 and beyond.
Jean-Christophe Laneri, Head of Cognitive Network Solutions at Ericsson, commented:
"The success of this field trial with KDDI confirms that the UIO rApp, operating on EIAP, can play a significant role in realizing live operations at Autonomous Network Level 4 (AN L4). This solution introduces advanced and robust automated optimization to mobile networks, significantly enhancing subscriber experience. The UIO rApp is now generally available for the global market, and as part of Ericsson's expanding rApp portfolio, we plan to deploy further AI-driven rApps in the future."
Hiroaki Ando, Representative Director and President of Ericsson Japan K.K., stated:
"High-reliability and low-latency uplink performance is extremely important as a foundation to support physical AI use cases, which Japan positions as a priority area in its semiconductor and digital industry strategy. This trial is a major step towards realizing that. Ericsson will continue its close collaboration with KDDI and support the development of Japanese society and industry."
A key outcome of this trial was also the successful integration of a third-party rApp, demonstrating that EIAP can support both Ericsson and non-Ericsson rApps in compliance with the R1 standard. In addition to Ericsson's UIO, KDDI also tested FYRA Suite rApp, a third-party rApp developed in Japan, in this trial. The combination of Ericsson's rApp and a third-party rApp shows that an open rApp platform and the EIAP developer portal can enable an innovation ecosystem where Japanese companies can develop domestically and deploy globally. This further strengthens KDDI's commitment to an open and interoperable RAN ecosystem compliant with O-RAN.
Improvements in uplink performance were confirmed even under an increase of over 10% in uplink traffic, demonstrating the robustness of the AI model under sustained load.
It has been proven that rApps can contribute to higher levels of autonomous networks.
Ericsson (NASDAQ: ERIC) and KDDI have successfully completed a large-scale field trial (hereinafter, "this trial") of Ericsson's AI-driven UIO (Uplink Interference Optimizer) rApp. The UIO rApp, operating on the EIAP (Ericsson Intelligent Automation Platform), demonstrated predictable uplink performance improvements across 4G and 5G live networks, proving its readiness for greater network autonomy.
This trial was conducted in Q1 2026, targeting a multi-technology, multi-band cluster consisting of approximately 1,500 5G cells and 1,300 4G cells, to verify whether AI-based uplink interference management could bring scalable improvements under real traffic conditions. The UIO rApp selectively optimized cells carrying the majority of network traffic, achieving substantial uplink performance improvements without compromising network stability.
Key performance evaluation results in the selected cells are as follows:
- Throughput improved in both 4G (average 9.6% improvement) and 5G (average 3.1% improvement), leading to enhanced end-user experience.
- Consequently, 5G SINR (Signal-to-Interference-plus-Noise Ratio) improved by 27%.
- Uplink modulation efficiency and spectral efficiency improved for both technologies.
This trial with KDDI is the world's first instance of an AI-driven UIO rApp being deployed in a live network environment, and also the first time the EIAP platform has been used in a live environment in Japan. The UIO rApp was evaluated as achieving level 3.86 autonomy in uplink optimization, applying the TM Forum's ANLET methodology. These results clearly demonstrate the role this application, and the valuable and achievable use case of "AI-powered uplink optimization," can play in telecommunications service providers' journey towards network autonomy level 4 and beyond.
Jean-Christophe Laneri, Head of Cognitive Network Solutions at Ericsson, commented:
"The success of this field trial with KDDI confirms that the UIO rApp, operating on EIAP, can play a significant role in realizing live operations at Autonomous Network Level 4 (AN L4). This solution introduces advanced and robust automated optimization to mobile networks, significantly enhancing subscriber experience. The UIO rApp is now generally available for the global market, and as part of Ericsson's expanding rApp portfolio, we plan to deploy further AI-driven rApps in the future."
Hiroaki Ando, Representative Director and President of Ericsson Japan K.K., stated:
"High-reliability and low-latency uplink performance is extremely important as a foundation to support physical AI use cases, which Japan positions as a priority area in its semiconductor and digital industry strategy. This trial is a major step towards realizing that. Ericsson will continue its close collaboration with KDDI and support the development of Japanese society and industry."
A key outcome of this trial was also the successful integration of a third-party rApp, demonstrating that EIAP can support both Ericsson and non-Ericsson rApps in compliance with the R1 standard. In addition to Ericsson's UIO, KDDI also tested FYRA Suite rApp, a third-party rApp developed in Japan, in this trial. The combination of Ericsson's rApp and a third-party rApp shows that an open rApp platform and the EIAP developer portal can enable an innovation ecosystem where Japanese companies can develop domestically and deploy globally. This further strengthens KDDI's commitment to an open and interoperable RAN ecosystem compliant with O-RAN.