Introduction


One of today's most prominent technology advancements is artificial intelligence. Artificial intelligence is a subset of machine learning, the process of creating software that can learn by itself and provide users with more efficient digital services.
Today, 5G has become AI's greatest ally and is truly changing the landscape of enterprise connectivity.
Let's look at how they are related and how you can use this relationship to your business advantage.

Artificial intelligence and machine learning in the 5g network


Artificial intelligence is used in a wide range of fields and industries. It aids companies in better understanding their clients, automating procedures, and even enhancing the systems that support a company, including its supply chain.
Today, 5G is significantly more powerful than 4G. Given this, it is not surprise that 5G is currently the go-to networking technology because it can serve the metaverse and connect millions of IoT devices. Real-time users of virtual reality

5G can be a critical factor in supporting the integration of ML and AI into the edge network. The figure below shows how 5G enables the simultaneous connection of multiple IoT devices generating vast amounts of data. Wireless operators can now provide the following thanks to the combination of ML and AI with 5G multi-access edge computing (MEC):
  • High degree of automation made possible by a distributed ML and AI architecture at the network's edge
  • Application-based traffic control and aggregation across heterogeneous access networks
  • Dynamic network partitioning to address different use cases with additional QoS requirements
  • An ML/AI-as-a-service offering for end users

How 5G and AI work together


The power of 5G comes at the cost of complexity, complexity best handled by artificial intelligence:
  • AI organizes 5G servers in the operator's network,
  • Artificial intelligence can optimally allocate spectrum and optimize it for the latency, bandwidth, and reliability of each device;
  • AI can detect network intrusion.
  • Even more interesting is how AI is reaching new levels in the context of Industry 4.0, as 5G enables it to connect every tool, person, and machine. For example, with a wireless 5G camera, you'll be able to control a robot arm, pilot robots in warehouses, or even measure the performance of every single hand-carried tool in an automotive chain.

Machine learning/AI in 5G


Every research community has attempted to assess how machine learning would affect 5G in their field of study in light of the growing benefits and advancements of ML in wireless communication. We have a number of publications on the implications of ML on the physical layer, security issues, managing radio resource, etc. as a result. This makes it very challenging to provide a concise summary of the application of AI/ML and its influence in 5G. As a result, we can categorize the works that apply ML to 5G into two categories: "generic ML classification," where we take into account all potential ML approaches from the literature, and "deep learning-based categorization," which only considers deep learning. Several leading publications think deep learning to be the most promising ML approach for the high complexity of 5G.

AI/ML is revolutionizing the 5G planning and optimization process.


The initial planning process is one of the most critical processes for determining the ultimate performance of a mobile network and its success in both technical and financial aspects. A poorly dimensioned network will always demand more involvement from the control and control teams to get the network performance to an acceptable level because the initial planning will also determine how the operation, operation, control, and control processes operate. Infrastructure (node placement), spectrum, configuration parameters and procedures, power consumption, network capacities to handle worst-case scenarios (peak traffic or busy hours), development of bandwidth demand over time, etc. must all be decided throughout this phase. In addition, planning and deploying a next-generation mobile network is not a greenfield task, i.e., starting from scratch. In fact, this planning task should consider already existing legacy systems and assets such as point of presence, pre-existing base station, optical fiber to connect core network elements, data center, etc.
In this complex optimization problem, i.e., network planning, where several input parameters are uncertain random variables or distributions, AI can play a vital role in handling the high complexity and providing highly efficient solutions. The AI integration module consists of three parts:

1. Data acquisition and preprocessing


Mobile network operators (MNOs) generally work with complex, disparate data sets, with helpful information stored in multiple systems, such as operations and management systems (OMS), billing systems, inventory, network elements, customer relationship systems for management (CRM), etc. However, for MNOs to achieve high performance in the future, they must adopt efficient big data tools to bring together all the necessary and profitable datasets. An AI-based planning system should be able to intelligently analyze and correlate all these different data sources. Intelligent and effective data management must include all the necessary functions of data collection, cleaning, filtering, correlation from multiple sources, and relevant data retrieval.


2. Knowledge discovery


this is the learning phase where we try to learn and understand the traffic and congestion pattern, user behavior, resource utilization, QoS, future location, faulty/problem elements/areas, and their impacts on network efficiency. This phase should help understand network performance and QoE, identify network anomalies, perform optimization, predict performance, disruptions, and requirements, and automate management and operations, such as self-configuration, self-optimization, and self-healing.


3. Knowledge utilization


this phase will use the acquired knowledge (e.g., prediction of time/frequency/spectrum traffic pattern analysis, the behavior of identified users, etc.) to decide on actions to be applied to network elements. /config. Such actions can either fix some system bugs or improve performance to adapt to current/upcoming situations and scenarios in the operational environment. In other words, this section provides options and/or planning for sharding, virtualization, edge computing and the impact of each decision and/or planning opportunity, network expansion plan or resource utilization plan decisions, corrective action suggestions.

Deployment of ML and AI in 5G


There are three ways wireless carriers can implement AI:
  • ML and AI algorithms are being integrated into individual edge devices for low processing power and rapid decision-making.
  • creating lightweight machine learning and artificial intelligence (ML and AI) engines for multi-access edge computing (MEC) for real-time computation and dynamic decision
  • creating low-latency IoT services that are appropriate for a variety of usage scenarios.
  • An AI and machine learning platform integrated into a system orchestrator for centralized deployment that can handle intensive computation and storage for projections and historical analysis

5G advances using machine learning


At this year's Mobile World Congress Barcelona, we showcased some of our latest 5G technology innovations across various disciplines. One common theme in many of our demos is the use of machine learning techniques to improve the overall 5G system. Our demos show how AI can benefit from the power and efficiency of 5G operating in various use cases in the sub-7GHz and mmWave bands. Check out our demos that use machine learning to improve:
  • Massive MIMO channel state feedback for increased user throughput and system capacity
  • Mobile mmWave beam prediction for higher power and longer battery life
  • Positioning accuracy thanks to a combination of 5G measurements, GNSS, multipath profiles, sensor inputs
  • Network planning for efficient mmWave coverage expansion with different types of nodes
  • AI-powered RF sensing for indoor positioning today with Wi-Fi, 5G in the future

Conclusion


5G and AI/ML are the two core components driving future innovation and are inherently synergistic  AI enhancements can help improve 5G system performance and efficiency, while the proliferation of 5G-connected devices can support distributed intelligence with continuous improvement in AI learning and inference. . As the role of intelligence on the device becomes increasingly important, the transformation of the connected, intelligent edge has begun, which is key to realizing the full potential of our 5G future.