How Autonomous Driving Networks Could Redefine Future Telecom Infrastructure

1. From a Macro Perspective on Autonomous Driving Networks

The term "autonomous driving network" borrows from autonomous vehicles and is used in the industry to emphasize a network’s self‑decision capability, though true autonomy is questionable because networks ultimately serve human intent. The name functions like a brand, conveying a value proposition rather than a literal description.

Instead of focusing on implementation details, we should first consider the network’s value proposition: what it aims to achieve. Imagining an unrestricted vision of such a network helps guide future work.

As AI rapidly advances, the future intelligent society will rely on a transformed network infrastructure that acts as an enabler. According to senior leadership, when societal complexity grows, providing ever‑simpler network services becomes a competitive advantage.

The vision of an autonomous driving network is to disrupt current networking paradigms, driving industry and societal progress, and therefore cannot be understood solely from a traditional network viewpoint.

2. Core Capabilities of Future Autonomous Driving Networks

Key capabilities from a user perspective include continuous availability, fault perception, resource pooling for unified scheduling, and intelligent rerouting. These functions enable automatic fault recovery with minimal human intervention and must be tightly integrated with availability management processes.

Additional essential functions are open interfaces for diverse industry applications, intelligent analytics for network management, and support for multi‑scenario adaptability.

3. Staged Development Path

Inspired by the five levels of autonomous vehicles, the network evolution is divided into five stages. Currently, networks are at the "recommendation" level, focusing on perception capabilities such as automatic fault detection and cause recommendation. As capabilities mature, they will progress to a "control" level with quality perception and rerouting, and finally to a fully autonomous level supporting diverse environments and services.

4. SMART² Architecture – A Simplified 6S Network View

The proposed architecture follows the SMART² principles: Service‑oriented, Multi‑Access, AI‑driven, Robust, Trust, Tailored, and Safety/Security. Here, “Smart” denotes intelligence, while “Simplified” emphasizes delivering network functions with minimal complexity for users.

In English naming, the network could be called a "6S Network" to highlight these six key attributes.