Page 378 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

2 Use Case Description

2�1 Description

The 5G Base stations or cell towers account for 60% to 80% of the total power consumption
in a telecom network. This is due to 5G network planning that relies on static configurations
and periodic manual updates, which are not suited for rapidly evolving urban environments.
Changes in urban infrastructure, buildings, vegetation, and terrain often lead to signal
obstructions, signal loss and retransmissions. This causes energy wastage in base stations.
These outdated methods not only degrade user experience but also contribute to increased
carbon emissions making them unsustainable for future 5G and 6G deployments.

The primary objective is to develop a self-adaptive, energy-efficient 5G/6G infrastructure
that minimizes carbon footprint while maintaining optimal performance. The use case aims
to empower telecom operators, smart city planners, and network engineers with an AI-driven
system that can autonomously adapt GIS real-time data to telco base stations for reducing
energy wastage.

This solution combines (i) GIS-based spatial intelligence, (ii) AI-powered Digital Twins that
emulates real-time 5G base stations based on GIS intelligence in 5G core for energy efficiency.

(i) GIS-based spatial intelligence:

The 5G Core communicates with ArcGIS using REST-based APIs (using JSON format).

It utilizes the Feature Service (https:// <arcgis-server>/arcgis/rest/services/<service-name>/
FeatureServer) to fetch data on buildings, terrain, and population density. This data is then used
to simulate and optimize base station configurations in real-time. Additionally, the Geometry
Service (https:// <arcgis-server>/arcgis/rest/services/Geometry/GeometryServer) is employed
to perform line-of-sight visibility analysis based on terrain and building data.

(ii) AI Powered Digital Twin

The 5G core emulates connected 5G base stations as digital twins for energy reduction in 5G
network. The twin helps in evaluating different base station parameters using ArcGIS, before
sending to physical OpenRAN 5G base station. To evaluate different parameters, the twin uses
Genetic Algorithms and Proximal Policy Optimization at 5G Core.

1) Genetic Algorithms (GA):

In a 5G core, GA is used to optimize network-wide parameters in base station twin for beamforming
and transmission power. ArcGIS has information on terrain, buildings, and population density.
The 5G Core using GA and ArcGIS to simulate various network configurations in 5G base station
twin. It then evolves to find the most efficient setup.

Optimizing Base Station Placement: GA evaluates the connected base station twin to maximize
coverage and minimize interference with energy efficiency as its goal. It ensures that each base
station is placed in an optimal position based on the surrounding terrain and building density.
Optimizing Transmission Power and Beamforming: The algorithm uses evolutionary principles
to adjust transmission power and beamforming techniques to ensure that signal strength is
sufficient in high-demand areas while reducing power usage in areas with low demand.

342