Higher 5G throughput with Mavenir and Aira’s AI innovation

Mavenir, the cloud-native network infrastructure provider, and Aira Technologies, a provider in AI-driven network automation and intent-based intelligence for telecom operators worldwide, today announced a successful demonstration of an AI-powered high-fidelity MIMO channel estimation and prediction solution.

This integration leverages Mavenir’s Open RAN-based O-DU to enhance network capacity, enabling up to 35% more data transmission over the same spectrum, based on Aira’s simulation data. This results in faster speeds and an improved 5G experience for more users simultaneously, potentially setting a new benchmark for Multi-User MIMO (MU-MIMO) performance.

By utilising advanced proprietary machine learning algorithms, Aira's AI-based spectrum efficiency solution enhances spectral efficiency, increases throughput, and significantly reduces the total cost of ownership for mobile network operators. The solution integrates seamlessly with Mavenir’s commercial O-DU, making use of existing baseband processing hardware without the need for additional AI-specific hardware such as GPUs or accelerators. This innovative approach enables operators to achieve enhanced performance improvements with minimal operational complexity and cost.

The Open RAN framework provided by Mavenir is claimed to facilitate seamless integration of Aira's AI software, meeting the stringent timing demands of baseband processing. Aira’s Insight Engine employs AI to deliver more accurate estimations of signal-to-noise ratio (SNR), user mobility, and RF environment metrics. This data feeds into Aira’s AI-powered channel prediction model, resulting in improved MU-MIMO efficiency and network throughput.

"The application of machine learning to wireless baseband processing at this level is an industry first,” said RaviKiran Gopalan, CTO and Co-Founder of Aira Technologies. “Together with Mavenir, we are demonstrating three pivotal advances: the power of the Open RAN ecosystem to accelerate innovation, the untapped potential of ML in next-generation wireless applications, and the transformative capability of AI to redefine the RAN,” he added.

“Our partnership with Aira in applying AI inline to time-sensitive channel estimation illustrates the potential of Open RAN to integrate third-party innovations into the RAN ecosystem at a rapid pace,” noted Sachin Karkala, SVP and GM of RAN at Mavenir. “The outcomes of the ongoing outdoor trial are very encouraging and indicate an increase in downlink throughputs and spectral efficiency for TDD MU-MIMO systems. Open RAN will accelerate the adoption of AI in the RAN ecosystem, delivering significant benefits to operators,” Kalra further added.