Project NEXCOM

Aperçu

NEXCOM aims to redefine future 6G Radio Access Networks by transforming them from communication-centric infrastructures into integrated communication and computing platforms. By jointly optimizing wireless communications, distributed computing, and AI-native intelligence, the project introduces novel paradigms such as distributed In-RAN computing, over-the-air computing, generative semantic communications, and computing-oriented security to enable scalable, energy-efficient, and intelligent wireless services.

📅 1er october 2026

• Rafik Zayani, IETR : coordinator

The NEXCOM project proposes a fundamental transformation of the Radio Access Network (RAN), evolving from traditional communication-centric infrastructures toward integrated communication and computing platforms. Rather than simply transporting raw data, NEXCOM introduces a new paradigm in which wireless transmissions are designed to directly support function computation and task execution. Extending the concepts of Cloud-RAN (C-RAN) and Multi-access Edge Computing (MEC), NEXCOM redefines the architecture of future wireless networks by embedding distributed computing capabilities within the RAN itself, enabling the next generation of intelligent, ultra-reliable, low-latency, and energy-efficient AI-native services.

To realize this vision, NEXCOM is structured around four complementary scientific pillars:

1. Distributed In-RAN Computing: Transforming the RAN into a distributed computing platform where communication and computation are jointly orchestrated across RAN nodes, enabling efficient execution of AI/ML tasks while optimizing communication, computation, and energy resources.
2. Distributed Over-the-Air Computing (AirComp): Exploiting the physical properties of electromagnetic waves and wireless channels to perform in-network information processing and goal-oriented computation, rather than conventional bit-wise data reconstruction, while remaining compatible with existing wireless standards and Open RAN architectures.
3. Generative RAN Communications: Introducing a novel communication paradigm in which RAN nodes exchange goal-oriented semantic prompts instead of raw data streams. These compact representations are interpreted by AI models to reconstruct only the information required by the target task, dramatically reducing communication overhead, latency, and energy consumption.
4. Computing-Oriented Security: Developing lightweight, AI-driven PHY- and MAC-layer security mechanisms specifically designed for communication-computation systems, providing confidentiality, integrity, authentication, and resilience against eavesdropping, spoofing, jamming, and data pollution attacks.

NEXCOM adopts a holistic methodology that tightly integrates information-theoretic modeling, wireless PHY/MAC design, AI-native orchestration, semantic and wave-domain information processing, and joint communication-computation architecture co-design.

The proposed concepts will be experimentally validated through proof-of-concepts based on a distributed multi-processing platform interconnected by parametrizable radio links, enabling joint communication-computation co-design and orchestration.