AI-enabled Massive MIMO
Project Status: set-up
Start Date: March 2020
End Date: February 2022
Budget (total): 3787 K€
Effort: 34.8 PY
Name: Arman Shojaeifard
E-mail: arman.shojaeifard(Replace this parenthesis with the @ sign)bt.com
BT, United Kingdom
Vilicom, United Kingdom
University of Bristol, United Kingdom
Loughborough University, United Kingdom
Nokia Bell Labs France, France
Nokia Bell Labs Stuttgart, Germany
Universität Stuttgart, Germany
IMST GmbH, Germany
The fifth-generation of mobile network (5G) aims at transforming the role of wireless technology in society through providing ubiquitous connectivity between all people and all devices. In the initial phase of 5G, coverage will only be available in certain areas in a number of cities worldwide. Increasing the 5G footprint requires additional investments from the operators to deploy new antennas, add spectrum, and upgrade their existing infrastructure. Telcos hence need to look for efficiency and monetization opportunities, beyond simply providing improved connectivity. 5G through its inherent attributes is anticipated to create the potential for a host of new services, use-cases, business and revenue models. To achieve full monetization, 5G networks must embed new emerging technologies including edge caching, cloud computing, and artificial intelligence (AI) – with the goal of generating real insight and optimizing key performance indicators (KPIs). Creating strong partnerships between telcos, vertical industries, and public agencies is the way forward. The CELTIC-NEXT platform provides this exact ecosystem for forging such partnerships towards joint innovation. AIMM is a consortium of such structure that will incorporate AI capabilities in 5G and beyond, with a particular emphasis on the radio access network (RAN) and its key technology: massive multiple-input multiple-output (MIMO). The main foci of AIMM include:
- Defining KPIs, use-cases and the business case for AI-embedded 5G and beyond.
- Exploring novel antenna configurations including collocated and distributed structures, and the use of intelligent metasurfaces for practical deployment.
- Building of comprehensive and sophisticated AI algorithms for 5G network operation and optimisation of the air-interface.
- Implementing practical AIMM testbeds based on off-the-shelf software radio technology for both centralised and cell-less antenna systems.
The AIMM project targets radical performance improvements in existing 5G RAN and beyond through extremely-large massive MIMO setups powered through and managed by the latest advancements in AI.