This work presents the experimental characterization, propagation modelling, and emulation of an indoor 5G Standalone private network. A complete 5G Standalone testbed was deployed using OpenAirInterface (OAI), a USRP B210, a 5G Core Network, and a commercial Quectel UE. Controlled Line-of-Sight measurements were used to characterize radio- and traffic-level performance and to establish a repeatable physical baseline. The measured link was then compared with the Close-In analytical path-loss model and with site-specific propagation simulations performed using Sionna RT. The Sionna-generated channel was subsequently converted into a discrete-time FIR representation and integrated into OAI RFSimulator while preserving its absolute complex coefficients. This integration highlighted the need for a consistent mapping between physical RF power and the digital IQ scale used by the simulator. An RF-to-digital calibration methodology was therefore developed from experimental characterization of the USRP B210 transmit and receive chains and extended to receiver gain and noise, providing a hardware-referenced interpretation of simulated RSRP and SNR. The calibrated Sionna-RFSimulator configuration successfully supported synchronization, network registration, and bidirectional user-plane traffic without normalizing away the physical channel gain. Overall, the developed framework establishes a consistent relationship between the propagation conditions of the physical 5G link and their representation in full-stack digital emulation. By combining a characterized physical testbed, site-specific propagation modelling, and RF-to-digital calibration, the work provides a foundation for future physical-to-digital studies and for the development of more complete Digital Twin representations of indoor 5G environments.
Questo lavoro si concentra sulla caratterizzazione sperimentale, sulla modellazione e sull’emulazione preliminare di una rete privata 5G Standalone in ambiente indoor. A tale scopo, è stato realizzato un testbed basato su OpenAirInterface, una USRP B210, una rete 5G Core e un modem 5G commerciale Quectel impiegato come User Equipment. Il setup è stato validato mediante l’analisi del piano di controllo, l’ottimizzazione dei parametri hardware e l’esecuzione di misurazioni controllate del traffico in downlink e uplink. In condizioni indoor controllate e in presenza di visibilità diretta (Line of Sight), sono stati raccolti diversi indicatori a livello radio e di traffico, tra cui SS-RSRP, SNR, BLER, MCS, allocazione dei PRB, throughput, jitter e perdita di pacchetti. È stata quindi definita una metodologia di calibrazione basata sui segnali di riferimento, finalizzata alla stima del path loss a partire dalle misure SS/PBCH. I risultati sperimentali sono stati successivamente confrontati con le previsioni del modello analitico Close-In e con le simulazioni di ray tracing effettuate mediante Sionna RT. I risultati ottenuti forniscono un riferimento calibrato per lo scenario indoor considerato e mettono in evidenza le differenze di comportamento tra il traffico in downlink e quello in uplink con la configurazione OAI adottata. Infine, le Channel Impulse Responses (CIR) generate da Sionna sono state esportate e integrate in OAI RFSimulator mediante un modello di canale a Finite Impulse Response (FIR). Questa integrazione preliminare dimostra la fattibilità dell’impiego, all’interno dello stack software di OAI, di canali derivati da simulazioni di ray tracing. Al tempo stesso, evidenzia la necessità di un’ulteriore fase di calibrazione, finalizzata a stabilire una corrispondenza coerente tra le ampiezze digitali utilizzate da RFSimulator, i valori di RSRP e SNR riportati da OAI e il comportamento della catena RF reale di ricezione.
MODELLING AND EXPERIMENTAL VALIDATION OF INDOOR 5G PRIVATE NETWORKS
MARCHESE, MARIAROSA
2025/2026
Abstract
This work presents the experimental characterization, propagation modelling, and emulation of an indoor 5G Standalone private network. A complete 5G Standalone testbed was deployed using OpenAirInterface (OAI), a USRP B210, a 5G Core Network, and a commercial Quectel UE. Controlled Line-of-Sight measurements were used to characterize radio- and traffic-level performance and to establish a repeatable physical baseline. The measured link was then compared with the Close-In analytical path-loss model and with site-specific propagation simulations performed using Sionna RT. The Sionna-generated channel was subsequently converted into a discrete-time FIR representation and integrated into OAI RFSimulator while preserving its absolute complex coefficients. This integration highlighted the need for a consistent mapping between physical RF power and the digital IQ scale used by the simulator. An RF-to-digital calibration methodology was therefore developed from experimental characterization of the USRP B210 transmit and receive chains and extended to receiver gain and noise, providing a hardware-referenced interpretation of simulated RSRP and SNR. The calibrated Sionna-RFSimulator configuration successfully supported synchronization, network registration, and bidirectional user-plane traffic without normalizing away the physical channel gain. Overall, the developed framework establishes a consistent relationship between the propagation conditions of the physical 5G link and their representation in full-stack digital emulation. By combining a characterized physical testbed, site-specific propagation modelling, and RF-to-digital calibration, the work provides a foundation for future physical-to-digital studies and for the development of more complete Digital Twin representations of indoor 5G environments.| File | Dimensione | Formato | |
|---|---|---|---|
|
tesi_Marchese.pdf
File consultabile da postazioni biblioteca
Descrizione: Elaborato di tesi magistrale, focalizzato sulla modellazione e integrazione in OAI RFSimulator di una rete privata 5G indoor
Dimensione
4.26 MB
Formato
Adobe PDF
|
4.26 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14253/110598