How Edge AI Is Driving The 5G Shift Better

Authored by Mr. Somenath Nag, Vice President of Marketing & Corporate Strategy, Calsoft Inc

In 2022, high-speed 5G networks, Artificial Intelligence (AI), and Edge computing are the key pioneering and disruptive technologies individually revolutionizing multiple industry verticals. They are helping us realize innovative business models with the 5G rollouts, which are further enabling the Internet of Things (IoT) technologies for connected devices and smart sensors. These IoT devices can efficiently support ubiquitous connectivity, higher data rate, and ultra-low latency solutions. Now, with Edge Computing, advanced computational capabilities are coming closer to the users, at the network edges, thereby reducing end-to-end latency.

Considering these developments, Communication Service Providers (CSPs) are proactively adopting AI to reap exciting and reasonable opportunities in telecom. The main challenge, however, is integrating these technologies as they can be highly transformative for the real-time processing of massive data volumes and for lowering the latency, as well as for enhancing the performance of multiple use cases. 

In this technology-led market, Edge Computing is advancing through the execution of Edge AI in multiple edge devices such as smartphones, drones, and Automatic Guided Vehicles (AGVs), and more. This application of AI to the Edge devices will lead to more efficient wireless communications, longer battery life, and enhanced user experiences. The low latency and high capacity of 5G will also allow distributed Edge AI processing for devices, Edge Cloud, and the central cloud. It will enable flexible system solutions for a variety of use cases such as smart manufacturing, intelligent retail, boundless XR and smart healthcare.

The Role of Edge AI in 5G World

Edge AI can ensure efficient services by processing massive amounts of data to provide real-time analytics. This paradigm shift moves Machine Learning (ML) to where the data originates and processes the data at the source, in real-time. Edge AI offers autonomous application of advanced ML and Deep Learning methods on IoT devices to compute locally, which eliminates the need for additional cloud services. On-device Edge AI further enhances the overall end-to-end 5G system, reducing operating costs.

Key purposes of Edge AI in a 5G ecosystem are:

• Improve the end-user experience: More intelligent beamforming and power management ensure better data rate and Quality of Service (QoS). ML or AI builds contextual awareness on devices to optimize performance and power consumption.
• Enhance system performance: On-device inferences ensure better spectrum utilization and mobility, reducing network data traffic.
• Better radio awareness and enhanced radio security: Helps in detecting any malicious activities and securing the Radio Access Network (RAN). 

Edge AI together with 5G can pave the way for more promising industrial applications such as:

• Manufacturing and Industrial IoT: Enables factory automation and predictive maintenance with on-premises Edge Analytics.
• Healthcare: Helps in real-time patient monitoring and provides automated records.
• Intelligent Retail: Creates a next-level retail experience through electronic price tags, pick-up and return kiosks, speedy drone delivery, and shoppers’ analytics.
• Smart Homes: Ensures that Edge AI is positioned within the home for real-time response and enhances privacy by maintaining relevant third-party data.
• Metaverse: Paves the way for immersive applications such as XR or Augmented Reality/Virtual Reality (AR/VR), thereby producing seamless digital experiences.

How Edge AI Improves 5G Adoption

For the telecoms to drive, optimize, enhance, and accelerate into this Edge AI generation, the CSPs or mobile operators will need to consider certain advancements.

• Eliminate vendor-lock in and proprietary hardware or software resources and step into open-source solutions.
• Enable faster Time-to-Market (TTM) by containerizing the Virtual Network Functions (VNFs). 
  • Software deployment and management can now be automated with open-source container orchestration systems such as Kubernetes. Cloud-native applications can also be automated and orchestrated using private or public cloud platforms.
• Realize Self-Organizing Networks (SONs) from Network Slicing with AI to enable fast onboarding of VNFs and to underpin diverse services.

In this new Edge AI model, the positioning of Edge and AI can vary depending on the use cases:

• On-device Edge: Devices such as smartphones will host AI capabilities.
• On-premises Edge: Edge capabilities can be placed within any enterprise with AI capabilities.
• Mobile Operators Network Edge: Edge AI is placed anywhere on the operator's network, Edge data centers, base stations, and gateways, wherein it can utilize the Multi-access Edge Computing (MEC) framework.

Afterword

All things considered, 5G with Edge AI enables innovative market platforms which leads to newer revenue models and higher RoI. Edge AI also becomes critical for 5G to design and orchestrate the entire network fabric, which inherently improves the network performance while keeping the OpEx substantially low. As a result, Edge AI in the 5G ecosystem can empower smart enterprise applications and Private Network deployments.