Auto-linearization system of integrated communication and computation system – AICom

Adding self-automation capabilities is crucial for 6G networks, where the demand for intelligence in every facet of communication is higher than ever. A fully realized "Intelligence of Everything" (IoE) in 6G requires more than just communication services. To be able to offer solutions adapted to the needs of smart cities, autonomous vehicles or the metaverse, telecommunication systems must introduce self-automation and self-correction capabilities at all levels. The AICom project aims to integrate artificial intelligence (AI) techniques into telecommunication networks to enhance the linearization of the Radio-Frequency transmitters.
Archimedes famously said, "Give me a lever long enough and a fulcrum on which to place it, and I shall move the world." In the AICom project, over-the-air (OTA) techniques serve as the fulcrum, while an integrated communication and edge computing (ICEC) architecture is proposed to act as the lever to revolutionize 6G network linearization. Traditional server-based computations will be shifted to edge devices. Here, user equipment (edge devices) will train AI models using smaller, localized datasets, while DPD parameters are analogically aggregated via OTA channels and federated learning. This ICEC design enables auto-linearization, where DPD parameters are computed by the network itself, achieving low-latency and energy-efficient communication towards 6G Artificial Intelligence and Radio Access Network (AI-RAN).