Next-generation networks $(5 \mathrm{G} / 6 \mathrm{G})$ provide capabilities such as network slicing, enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communications (URLLC), and edge computing. However, configuring the related radio, core, slicing, and edge resources requires significant operational expertise. This work proposes an Intent-Based Networking (IBN) framework that combines Natural Language Processing, Large Language Models (LLMs), TMF921-compliant intent representation, and CAMARA APIs to simplify the definition and activation of network-sensitive vehicular services. In the proposed model, a Service Provider establishes a business agreement with a Network Operator, while the operator administrator defines service requirements through a conversational interface. These requirements are translated into machine-readable intents and mapped into network and edge orchestration actions. The framework is evaluated through a Teleoperated Driving (ToD) use case for autonomous vehicle repositioning. Results show that the intent translation and management pipeline introduces limited and repeatable overhead, while orchestration time is mainly affected by the underlying MANO/IaaS platform. The results indicate that combining IBN and CAMARA APIs can support flexible service preparation by operators and dynamic service consumption by applications.
Andrea Speranza, P. Giardina, Giacomo Bernini et al.· International Conference on...· 0 citations
Configuring 5G networks through standardized exposure interfaces—the Network Exposure Function (NEF) and the Common API Framework (CAPIF)—remains operationally burdensome, requiring manual navigation of 3 GPP parameter schemas, OAuth2 authentication, and live network state. This paper presents a modular, intent-driven architecture that translates natural language operator goals into schema-compliant NEF API configurations using a compact, locally deployed Large Language Model (LLM), Qwen3-4B (4 billion parameters). A deterministic pre-classification layer reserves LLM inference for semantically complex requests, while a six-stage pipeline produces near-deterministic, conflict-aware JSON output. Two complementary validation paths are provided: a Standard Mode with closed-loop Quality of Service (QoS) feedback via a NEF emulator, and a Free5GC Mode that validates the same translation logic against a Free5GC-based experimental environment with real user-plane traffic. Demonstration across five vertical scenarios confirms end-to-end correctness in both the emulated NEF environment and the Free5GC-based experimental environment.
Hao You, Chathura Galkandage, Naércio Magaia et al.· IEEE Conference on Network S...· 0 citations