Aira Technologies, in Collaboration with Microsoft and Intel, Demonstrates an xApp to Dramatically Improve 5G MU-MIMO Efficiency and Throughput using AI

Aira Technologies, a pioneer in the application of Machine Learning (ML) to radically improve wireless telecommunications, announced that it will deliver a near-real time application (xApp) that will work in concert with the Azure Operator Distributed Services to deliver advanced 5G RAN capabilities that have long been requested by Mobile Network Operators (MNOs). Aira has used sophisticated machine learning (ML) techniques to develop one of the industry’s first xApp that enables better channel estimation and unprecedented channel prediction performance to help maximize downlink throughput and range. The Intel FlexRAN reference software, with a software based Layer 1 (L1), provides the flexibility to tap into the right part of the L1 to enable the xApp. Accurate channel prediction helps minimize inter-user interference in order to maximize multi-user MIMO performance. Disaggregated and programmable networks promise more innovation using software solutions.

“The application of ML to wireless baseband processing is an industry first. What we have been able to achieve here together with Microsoft and Intel showcases three very important things – the power of the Open RAN movement to accelerate innovation; the extraordinary potential for ML in cutting edge wireless applications; and, the enormous potential to improve real-time performance of the RAN through xApps,” said Anand Chandrasekher, Founder and CEO of Aira Technologies.

Recommended AI: QuickLogic Drives eFPGA Innovation with New Aurora Development Tool Suite

“As an industry leader in edge and cloud computing and artificial intelligence, we’re eager to work with Aira to showcase their AI capability at the 5G physical layer and its potential to significantly enhance RAN DU performance,” said Victor Bahl, Technical Fellow, and CTO of Azure for Operators at Microsoft. “We strongly believe in accelerating innovation into the RAN by ingesting data from all parts of the RAN and applying sophisticated state of the art AI techniques to it.”

“We at Aira think that wireless design should be data-driven, agile and adaptable. As algorithms adapt and optimize over time, the APIs that supply inputs into these applications should be equally adaptable and configurable,” said Ravikiran Gopalan, Founder and CTO of Aira Technologies. “Azure Operator Distributed Services makes Open RAN innovation much more developer friendly. Using Azure has significantly cut down the time and effort required for Aira to iteratively upgrade our xApp and modify the size and shape of the inputs required by the xApp.”

The Intel Xeon platform and the Intel FlexRAN reference platform, along with CapGemini’s L2 and L3 software platforms provided crucial building blocks, and facilitated easy incorporation of Aira’s cutting edge ML technology.

Recommended AI: Philips Speech and Sembly AI Launch SmartMeeting As Answer To New Meeting Culture

“What Aira is doing with machine learning in the RAN stack exemplifies the potential performance and energy efficiency improvements this technology can bring to radio access networks,” said Cristina Rodriguez, Vice President and General Manager, Wireless Access Networking Division at Intel. “By taking advantage of Intel Xeon processors and FlexRAN reference software for the DU, Layer 1 and above, Aira is not only demonstrating the power of the Open RAN community and potential benefits for bringing various partners into the vRAN ecosystem, they are also demonstrating the potential for improving RAN performance through innovation.”

“Capgemini is delighted to work with Aira to enable creation of their native AI-based 5G CU/DU solution with its feature rich RAN L2/L3 software including support for mature Open RAN interfaces, enabling integration with RIC and xApp environments,” said Shamik Mishra, CTO – Connectivity, Capgemini Engineering. “Aira’s innovation in applying sophisticated ML to the 5G physical layer enables significant performance improvements in the RAN. We are happy to contribute and partner with Aira in their innovation in the Open RAN space.”

The setup at the TIP community lab sponsored by Meta in Menlo Park, California, consists of a Viavi UE emulator and Viavi channel emulator that simulates varying channel conditions as seen by a Foxconn RU. The RU itself is connected in an end-to-end fashion to the Aira DU, a Capgemini CU, and a Capgemini Core. Aira is one of the first third party companies to leverage Microsoft’s programmable RAN platform to simplify the process of getting data from the DU and out over the E2 interface without having to define new service models. The Aira AI-based High Fidelity MIMO Channel Estimation and Prediction application (xApp) running on the Azure cloud exchanges data with the DU via the E2 interface. The xApp estimates and predicts the channel which is then compared to ground truth as established by the Viavi channel emulator. The output of the xApp is the prediction of channel evolution in-between S-slots, which results in improvements in RAN throughput and coverage.

Recommended AI: UTB Bot Unveils a New Way to Leverage Automation and Cryptocurrencies

[To share your insights with us, please write to sghosh@martechseries.com]