Intelligent Near-RT RIC xAPP for 5G O-RAN Small Cells: Energy-Efficient PRB Optimization with E2 Control

Abstract

This paper presents the design, implementation, and experimental validation of an intelligent xApp for the Near-Real-Time RAN Intelligent Controller (Near-RT RIC), targeting 5G Open Radio Access Network (O-RAN) small-cell deployments. The system operates over three contiguous 100 MHz carriers in the 3.5-3.8 GHz band (3GPP n78) on a real srsRAN testbed. Built on the O-RAN Software Community (OSC) Near-RT RIC and using the E2SM-CCC service model, the proposed xApp dynamically adjusts the active Physical Resource Block (PRB) ratio and selectively deactivates underutilised carriers under low traffic. A four-step PRB sweep (100%, 70%, 50%, 30%) executed entirely through closed-loop RIC control achieved a 100% RIC_CONTROL_ACK rate across 18 dispatched commands; capping the PRB ratio at 30% concentrated transmit power on roughly one-third of the bandwidth, raising the peak SINR to 34.7 dB and the per-PRB spectral efficiency to 1.617 Mbps/PRB, a 4% gain over the fullallocation baseline. The corresponding transmitter energy saving was 54.9% at $P R B=30 \%$ and approximately 40% averaged across a 24-hour small-cell traffic profile, while downlink throughput at realistic smallcell load (below 10 Mbps) stayed within 2% of the static baseline. End-to-end Voice-over-NR (VoNR) calls were established between two physical handsets through a Kamailio/PyHSS IMS core, with Mean Opinion Score (MOS) improving from 3.4 to 4.5 under the proposed policy. The results show that intelligent PRB and carrier management through closed-loop RIC control is a practical, validated path to sustainable 5G small-cell operation on commodity SDR hardware.