Ericsson and KDDI completed a large-scale field trial of AI-driven uplink optimization across KDDI’s commercial 4G and 5G network in Japan, demonstrating measurable gains in throughput, spectral efficiency, and signal quality under live traffic conditions. The trial used Ericsson’s Uplink Interference Optimizer (UIO) rApp running on the Ericsson Intelligent Automation Platform (EIAP), alongside a third-party rApp developed by Japan-based FYRA. Ericsson said the deployment marks the first live commercial implementation globally of its AI-driven UIO rApp and the first use of the EIAP platform in a live Japanese network environment.
The Q1 2026 trial covered approximately 1,500 5G cells and 1,300 4G cells across multiple frequency bands and radio technologies. Ericsson reported that the AI system selectively optimized the busiest cells in the network, producing a 9.6% average uplink throughput improvement in 4G and a 3.1% average improvement in 5G while network uplink traffic increased by roughly 10% during the test period. The companies also reported a 27% gain in 5G Signal-to-Interference-plus-Noise Ratio (SINR), along with higher uplink modulation and spectral efficiency across both generations of mobile technology.
The trial also validated interoperability between Ericsson and third-party rApps through the R1 interface standard, reinforcing broader industry efforts around open and interoperable O-RAN-aligned radio access networks. Ericsson said its Uplink Interference Optimizer achieved an Autonomous Networks Level 3.86 assessment using an adaptation of the TM Forum ANLET methodology, positioning the technology as a stepping stone toward Level 4 autonomous mobile network operations. Ericsson confirmed that the UIO rApp reached General Availability status during the project and said additional AI-driven rApps are planned as part of its expanding automation portfolio.
- Trial scale: ~1,500 5G cells and ~1,300 4G cells
- 4G uplink throughput improvement: 9.6%
- 5G uplink throughput improvement: 3.1%
- 5G SINR improvement: 27%
- Trial timing: Q1 2026
- Platform: Ericsson Intelligent Automation Platform (EIAP)
- Third-party rApp tested: FYRA Suite
- Standards alignment: O-RAN R1 interface compliance
- Autonomous maturity assessment: Level 3.86 using TM Forum ANLET adaptation
- Deployment status: Ericsson UIO rApp now generally available
Jean-Christophe Laneri, Head of Cognitive Network Solutions at Ericsson, said: “The successful field trial with KDDI confirmed the readiness of the UIO rApp, running on the EIAP, for a key role in achieving Autonomous Networks Level 4 live operation. This solution introduces advanced, robust automated optimization within mobile networks, greatly improving subscriber experience.”
🌐 Analysis: The Ericsson-KDDI field trial reflects a broader industry push toward AI-native radio access networks capable of self-optimization with minimal human intervention. Operators globally are exploring rApps, automation frameworks, and intent-based operations to reduce operational complexity while improving network efficiency as 5G traffic and AI-driven applications continue to scale. Uplink optimization is becoming increasingly important as XR, industrial automation, robotics, and AI inferencing workloads generate heavier upstream traffic patterns.
🌐 Analysis: Ericsson is competing aggressively in the emerging Autonomous Networks and Open RAN software stack against vendors including Nokia, Samsung Electronics, and Mavenir. The successful integration of a Japanese-developed third-party rApp also aligns with Japan’s national focus on open telecom ecosystems and AI-enabled digital infrastructure, particularly as the country accelerates investments in AI, semiconductors, and next-generation wireless platforms.
