Coverage@CS-KAU: A Dataset of Coverage Measurements Across Indoor Environments, Propagation Scenarios, and Radio Access Technologies

Le résumé fourni par la source

We present Coverage@CS-KAU, a dataset containing data collected for multiple Radio Access Technologies (RATs), designed to support the analysis of indoor wireless coverage and propagation in realistic environments. While prior datasets have typically been limited to individual RATs, specific indoor settings, or narrowly defined propagation scenarios, Coverage@CS-KAU brings these dimensions together within a unified measurement campaign. Collected at the Department of Mathematics and Computer Science at Karlstad University, Sweden, the dataset spans five RATs — LTE-Advanced (LTE-A), Narrow Band Internet of Things (NB-IoT), 5G New Radio (NR), WiFi, and Ultra-Wideband (UWB) — and includes measurements from offices, a laboratory, and an open space. It captures both indoor-to-indoor and outdoor-to-indoor conditions and offers fine spatial granularity. In addition to mobile system channel indicators and WiFi received signal strength measurements, the dataset provides UWB channel impulse responses alongside received signal strength indicators and ranging information. Coverage@CS-KAU is intended as a resource for multi-faceted analyses in channel modeling, coverage evaluation, positioning, and communication system optimization via machine learning. Measurement Environment The entire campaign was conducted on the 4th floor (Plan 4) of the CS@KAU building. The floor map presented in the auxiliary file "Measurement_Map.pdf" marks every measurement point across the three areas, the UWB responder locations, and the LTE-A / 5G NR radio-cell positions of the CARL-W indoor private deployment available on campus; in addition, the auxiliary file "WiFi_AP_Locations.pdf" shows the positions of the Access Points located at the 3rd, 4th and 5th floor of the CS@KAU building. Measurement Areas Area 1 - Offices Multiple offices separated by walls and partitions along two corridors of Plan 4. Most rooms are 2.5 × 3 m and similarly furnished; larger rooms host multi-point measurement grids. LTE-A / NB-IoT / NR — 104 points in 69 offices WiFi — 40 points in 38 officesScenarios — I-to-I & O-to-I Area 2 (21D404) - CARL-W Laboratory Approximately rectangular lab of ~45 m2 hosting one of the four LTE-A / 5G NR radio dot pairs of the CARL-W indoor private deployment, enabling I-to-I measurements on a structured grid. LTE-A / NB-IoT / NR — 16 pointsWiFi — 10 pointsUWB — 10 points, 5 responders each Area 3 (21A442/21A443) - Open Space ~12 × 16 m open study area, lightly furnished, separated by a permanent glass wall into two halves. Also hosts a CARL-W LTE-A / 5G NR radio dot pair. LTE-A / NB-IoT / NR — 43 points WiFi — 10 points UWB — 10 points, 5 responders each Measurement Setup RAT Instrument Key Parameters LTE-A / NB-IoT / 5G NR Rohde & Schwarz TSMA6 scanner with omnidirectional RF antenna (698 – 3800 MHz) Bands: n28, 20, 32, 3, 1, 7, 42, n78 ~5 min per measurement point Channel indicators measured on the appropriate broadcast reference, namely: RS — Reference Signal (LTE-A) NRS — Narrowband Reference Signal (NB-IoT, one set per antenna port Tx0/Tx1) SSB — Synchronization Signal Block, comprising PBCH, DM-RS, PSS, and SSS (5G NR, used because NR PCIs do not transmit a cell-specific RS) WiFi Apple MacBook Pro 13 running the ThingsLocate scanning routine 2.4 GHz and 5 GHz bands Per-AP BSSID and beacon RSS UWB Qorvo DWM3001CDK boards (DW3110 transceiver), IEEE 802.15.4-2015 / FiRa compliant Channel 9 (centre frequency 8 GHz) 64-symbol preamble, 6.81 Mb/s Dual-Sided Two-Way Ranging One initiator and five responders Dataset Contents The dataset is organised by measurement area and, within each area, by location (room or grid point) and RAT. Point-level data are provided in .csv (3GPP), .txt (WiFi), and per-responder .txt (UWB) files; for measurement areas where UWB measurements were collected, a text file providing the UWB responder coordinates and a readme file reporting the value of the antenna bias and explaining the procedure to apply it to the raw distance estimates are also provided; finally, an annotated floor map showing all measurement points and information on the position of LTE-A/5G NR dots and UWB responders is also included. A brief summary of the features is provided below, while full details are provided in the "Supplementary Material" pdf file: 3GPP (LTE-A, NB-IoT, 5G NR) Per-sample channel indicators (RSSI, RSRP, RSRQ, SINR) in CSV; for NB-IoT, separate traces per antenna port (Tx0, Tx1). WiFi Per-AP RSS traces (BSSID + RSS samples) from every access point detected at a measurement point. UWB Raw I/Q samples per responder (five files per point) together with metadata extracted from the transceiver registers — path amplitudes, diagnostic parameters, raw estimated distance, and RSSI. Beyond the present document and its support files, the dataset contents are as follows: Coverage@CS-KAU/├── Measurements.zip Archive containing the measurement data, organised by measurement area as described above └── Auxiliary_files.zip ├── Measurement_Map.pdf Annotated 2D floor map of Plan 4 (.pdf) and per-area UWB responder coordinates ├── WiFi_AP_Locations.pdf Annotated 2D floor map of Plans 3, 4 and 5 (.pdf) showing the position of WiFi APs └── visualization_tool.py Python visualisation tool Visualization Tool A standalone Python visualization tool accompanies the dataset and allows reviewers to interactively explore every measurement file across the five RATs. The tool is built on Tkinter and Matplotlib and exposes a common selection workflow — technology, measurement area, location, file — with plots tailored to each RAT. 3GPP. Select a PCI (or NPCI, for NB-IoT) from those detected in the chosen file, ranked by descending sample count; the tool renders a 4 × 2 panel showing the time series and the empirical distribution of the four channel indicators (RSSI, RSRP, RSRQ, SINR), with sample mean and standard deviation overlaid. For NB-IoT, a control lets the user toggle between the two antenna ports (Tx0 and Tx1) transmitting the NRS. WiFi. Aggregates the RSS samples of every AP detected at the chosen point into a single empirical distribution annotated with the sample mean, standard deviation, and observed range. UWB. Reconstructs the Channel Impulse Response (CIR) magnitude from the raw I/Q samples, overlays the current CIR on the mean magnitude envelope with a one-standard-deviation shaded band, and provides a zoomed tail view for multipath inspection. Previous / Next controls step through consecutive CIRs, with the transceiver-level metadata (estimated distance, after applying the proper antenna delay compensation factor, and RSSI) displayed alongside. Running the tool python3 visualization_tool.py Requires numpy, pandas, matplotlib, and tkinter (bundled with most Python distributions). Click the Visualization Tool button above to view, copy, or download the code; the script must be saved inside the unzipped Measurements folder (alongside the three Measurement Area sub-folders) before it is run.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.