Spectrum and channel occupancy evaluations of a hospital’s private 5G test network

Abstract

Spectrum scarcity and coexistence challenges pose critical obstacles to deploying reliable wireless services in hospital environments, where uninterrupted, low-latency communications are essential for patient care and medical procedures. This thesis presents the first spectrum occupancy and channel occupancy measurement campaign spanning 0.4 GHz – 6.1 GHz, conducted in two operational hospitals in Oulu: a newly constructed facility equipped with a private fifth generation of mobile network technology (5G) test network and an older hospital lacking 5G infrastructure. Using a high-performance radio frequency (RF) front-end (Agilent E4446A spectrum analyzer, wideband antenna, low-noise amplification) and MATLAB-driven measurement scripts, we conducted 24-hour continuous frequency sweeps in clinical operating rooms under both idle and active (surgical) conditions, executed before and after the 5G deployment, and similar measurements were conducted in the old hospital’s operation room where no 5G network is deployed.

Initially, we performed all the analysis across the entire spectrum, i.e., 0.4 – 6.1 GHz, and presented the results with detailed discussion. Later, for clarity, we segmented the measured spectrum into eight technology specific bands (GSM at 900 MHz and 1800 MHz, LTE Bands 1 (2100 MHz), 3 (1800 MHz), and 7 (2600 MHz), 5G test network at 3.9 – 4.1 GHz, ISM bands for Wi-Fi/Bluetooth at 2.4 – 2.4835 GHz, and unlicensed 5 GHz Wi-Fi at 5.150 – 5.350 GHz, 5.470 – 5.725 GHz, and 5.725 – 5.850 GHz) and computed the channel occupancy (CO) metric. Our key findings demonstrate that the private 5G test network (operating in 3.9 – 4.1 GHz spectrum) remains spectrally isolated: there is zero measurable leakage below 3.9 GHz or above 4.1 GHz into adjacent long-term evolution (LTE) or WiFi bands, confirming compliance with 3GPP/ETSI out-of-band emission masks. The maximum 5G-band received power (-16 dBm) is orders of magnitude below International Commission on Non-Ionizing Radiation Protection (ICNIRP) and Institute of Electrical and Electronics Engineers (IEEE) safety limits, ensuring negligible RF exposure for patients and staff. Channel occupancy peaks reveal high utilization in legacy cellular bands (up to 99 % at 742 MHz) and moderate loading in industrial, scientific, and medical bands (approximately 6%), whereas the 3.996 GHz 5G channel exhibits only approximately 4% occupancy, confirming its suitability for ultra-reliable low-latency communications.

Building on these measurements, we propose spectrum-level guidelines for hospital 5G networks, including dynamic noise-floor thresholding, planned guard-band allocation, occupancy-aware frequency assignment, and avoidance of congested ISM channels for safety-critical services. Finally, we identify the 3.9 – 4.1 GHz mid-band as the optimal sub-6 GHz allocation for private 5G ultra-reliable low-latency communication (URLLC) in healthcare, with 2.6 GHz and 3.5 GHz mid-bands recommended for massive machine type communications (mMTC). Our results provide an empirical foundation for spectrum planning, interference mitigation, and safe 5G deployment in mission-critical hospital settings.