With each new technological generation, wireless networks produce enormous amounts of data and become more sophisticated. They are the best candidates for AI optimization because of this. As deep neural networks and machine learning become widely used, people can directly witness their effectiveness. A transcription program called Descript Artificial Intelligence and a huge language model called ChatGPT serve as examples of how Artificial Intelligence uses data-driven optimization to address challenging issues like human speech. AI is excellent at maximizing difficult situations, which makes it a great choice for improving wireless network performance.
What role does AI play in 5G networks?
As 5G advances, the 3GPP is researching AI and ML for cellular standards. AI applications for energy savings, load balancing, and mobility optimization focus on the air interface. Channel status information, beam control, and positioning will be examined in 3GPP Release 18. 3GPP establishes frameworks and evaluation methodologies, not AI/ML models. O-RAN ALLIANCE studies AI/network ML’s orchestration role beyond 3GPP. The RAN Intelligent Controller (RIC) supports AI/ML applications. This RIC supports near-real-time and non-real-time xApps. Existing xApps and rApps use AI for network orchestration and spectral efficiency. As the O-RAN ecosystem evolves, more applications will appear.
How will AI be used by 6G?
6G already showcases a significant dependency on AI and machine intelligence in shaping future wireless networks. Notably, the industry often uses the term “AI native,” even without a formal definition. To explore these AI-native networks, one can build upon the existing model, factoring in current trends in virtualization and RAN disaggregation. A variety of vendor-specific AI/ML models are likely to appear in each network block. These networks could also go by the term “AI-native networks.”
While traditional 5G networks utilize multiple human-engineered processing units in the air interface, 5G Advanced plans to use ML for specific optimizations. Furthermore, 6G is likely to employ deep neural networks to form the complete air interface.
What role will AI play in optimization?
Building on the idea that AI and ML can improve network management, 6G aims to leverage these technologies for optimization. For example, AI could regulate power usage by dynamically adjusting component activity based on operational needs. Currently, xApps and rApps control power-hungry components like amplifiers at the base station level.
However, AI’s ability to quickly solve complex problems and analyze large data sets extends optimization potential to city-wide or even national networks. Consequently, cells could adjust in real-time for energy-efficient resource use, and entire base stations could deactivate during low-usage periods. At present, such wide-scale reconfiguration takes days or weeks, limiting its feasibility. Yet, advancements in AI algorithms remain promising and attract the attention of infrastructure providers.
Importantly, the integration of AI into wireless networks isn’t restricted to 6G. Ongoing research across the ecosystem aims to develop new, reliable models for current and future wireless systems. Upon deploying this technology, it’s essential to train AI models on diverse data sets, evaluate their performance against existing methods, and establish new testing protocols for AI-enabled modules.
Undoubtedly, AI will revolutionize wireless communications in the coming decade. As AI models and testing methodologies mature, expect more innovative applications in both 5G and 6G networks.
See also : AI & Wireless 2024
