The Physical Layer in 5G: How It Works Today
The physical layer is the heart of any wireless network: it is responsible for converting our digital data into radio waves that travel through the air. In current 5G networks, it performs several critical tasks that determine the quality of our communications. When you send a message, the network encodes that information to protect it from errors, converts it into signals using advanced modulation techniques, and transmits it through multiple antennas simultaneously to increase speed [ref1]. Everything is dynamically adjusted according to real-time signal quality to balance throughput and error rate.
The key problem is that current methods must deal with important limitations. These decisions rely on mathematical models that assume ideal conditions, but reality is often more complex: rapid environmental changes and a large number of users can lead to a severe and non-acceptable performance loss [ref2].
The Transition to 6G: Artificial Intelligence as an Optimization Engine
Future 6G networks are revolutionizing this approach by incorporating AI directly into the physical layer. While current systems operate based on predefined mathematical formulas that model the channel, 6G enables learning directly from real-world data, allowing continuous adaptation and optimization [ref3].
- Intelligent Channel Prediction
AI enables the prediction of how the channel will behave, anticipating signal degradation and adjusting parameters before issues arise. Deep learning models have shown improved communication quality and stability, especially in dynamic environments [ref4]. - Self-Learning Systems
Researchers are developing systems in which the entire transmission chain (from encoding to reception) is designed using neural networks, learning optimal strategies for each network and channel condition [ref5]. - Dynamic Resource Optimization
AI also makes real-time decisions on how to efficiently allocate frequencies, power, and antennas among users, using techniques that test and learn to improve the network in new situations [ref3]. A clear example is the European 6G-LEADER project, which has already demonstrated significant reductions in overhead and latency by applying these methods in real environments [ref6]. - Intelligent Beamforming
At the high frequencies used in 6G, where signals are highly directional, AI helps point beams in the best direction without wasting time testing all possible options [ref7].
Challenges and Future Outlook
Despite its advantages, the application of AI to the physical layer still faces challenges: high computational demands, the need for large amounts of data to train models, and the lack of standardization in international organizations such as 3GPP [ref3] are key barriers that must be addressed by the scientific community.
Conclusion
We are moving from 5G networks that optimize the physical layer purely with mathematical models, toward 6G networks that continuously learn and predict the most efficient way to communicate. Thanks to projects like 6G-LEADER and advancements in AI, communication will be faster, more secure, and more adaptive than ever before.
References
[ref1] “Understanding the 5G NR Physical Layer.” Keysight Technologies, 2021.
[ref2] Gangfada, I. A., et al. “An Evaluation of the Physical Layer Characterization of 5G Networks.” IEEE ICOEI, 2020.
[ref3] Watson, C., Woods, K., Shyy, D. “6G and Artificial Intelligence & Machine Learning.” MITRE, 2021.
[ref4] Navabi, S., et al. “Predicting Wireless Channel Features Using Neural Networks.” IEEE ICC, 2018.
[ref5] Huang, H., et al. “Deep Learning for Physical-Layer 5G Wireless Techniques.” arXiv:1904.09673, 2019.
[ref6] “6G-LEADER Project” European Commission SNS JU, 2024.
[ref7] “AI in Massive MIMO and Smart Beamforming for 5G and 6G Networks.” Presidency University, 2024.