mmWave and midband 5G experimental data in an industrial setting

This dataset is part of the Industrial Network Repository (IN-Rep). IN-Rep Metadata: Ref. Application Domain Use case Technology Location Data N. Bouzar, L. De Nardis, G. Caso, M. Neri, F. Elbahhar and M.-G. Di Benedetto, "Range-free positioning for Industrial Internet of Things in a mixed public-private midband and mmWave 5G deployment", IEEE Conference on Standards for Communications and Networking, Workshop on Communication Networks for Next-Generation Industrial Internet of Things, September 15 - 17, 2025, Bologna, Italy. DOI: 10.1109/CSCN67557.2025.11230588 2,3,4 AGV 5G,UWB* BI-REX 5G: RF data, ToA;UWB*: Distance,RSSI, CIR *The UWB data are under preparation for release in an updated version of the dataset. The dataset includes 5G data collected in the industrial pilot line of the Big Data Innovation & Research Excellence (bi-rex) center, located in Bologna, Italy. The pilot line hosts production machines, robots and industrial demos in an area of approximately 300 m2, covered by a private 5G network operating at both 3.7 GHz and 26 GHz, as well as by public 5G networks at 3.7 GHz.Data were collected at 18 different locations in the area hosting the pilot line using the Rohde & Schwarz TSMA6 Mobile Network Scanner, part of the R&S Mobile network testing products, including a GPS receiver for data geo-mapping and accurate synchronization, and an up/down converter enabling the TSMA6 to collect simultaneously 4G data and 5G data both in in the 698-3800 MHz frequency range and at mmWave (26000 MHz), using two separate antennas. Collected data include RF parameters (RSSI, RSRP, SINR, RSRQ) and Time of Arrival information.Data are suitable for range-free positioning using machine learning algorithms such as the Weighted k-Nearest Neighbors algorithm.In the repository, you will find the 5G raw data, the script to process them, as well as the implementation of the WKNN positioning algorithm. ├── 5G_MNC_Processing.py├── 5G Raw Data ├── l10_5G.xlsx ├── l11_5G.xlsx ├── l12_5G.xlsx ├── l13_5G.xlsx ├── l14_5G.xlsx ├── l15_5G.xlsx ├── l1_5G.xlsx ├── l16_5G.xlsx ├── l17_5G.xlsx ├── l18_5G.xlsx ├── l2_5G.xlsx ├── l3_5G.xlsx ├── l4_5G.xlsx ├── l5_5G.xlsx ├── l6_5G.xlsx ├── l7_5G.xlsx ├── l8_5G.xlsx └── l9_5G.xlsx└── WKNN ├── apply_commonality_weighting_loo.m ├── Bi-Rex_Dataset.mat ├── compute_multi_param_distances_loo.m ├── compute_single_param_distances_loo.m ├── create_fingerprint_matrices_loo.m ├── create_global_unique_pcis.m ├── display_loo_results.m ├── main.m ├── perform_wknn_positioning_loo.m ├── plot_loo_results.m └── wknn_leave_one_out_positioning.m Please for more Information refer to the companion paper:N. Bouzar, L. De Nardis, G. Caso, M. Neri, F. Elbahhar and M.-G. Di Benedetto, "Range-free positioning for Industrial Internet of Things in a mixed public-private midband and mmWave 5G deployment", accepted for the IEEE Conference on Standards for Communications and Networking, Workshop on Communication Networks for Next-Generation Industrial Internet of Things, September 15 - 17, 2025, Bologna, Italy.本数据集隶属于工业网络数据集仓库（Industrial Network Repository, IN-Rep）。 IN-Rep 元数据： 参考文献 应用领域 用例 技术 采集地点 数据内容 N. Bouzar、L. De Nardis、G. Caso、M. Neri、F. Elbahhar 与 M.-G. Di Benedetto，《面向混合公私制中频段与毫米波5G部署场景下工业物联网的免测距定位技术》，发表于2025年9月15日至17日于意大利博洛尼亚举办的IEEE通信网络标准会议（IEEE Conference on Standards for Communications and Networking）下一代工业物联网通信网络专题研讨会，DOI：10.1109/CSCN67557.2025.11230588。 2,3,4 自动导引车（Automated Guided Vehicle, AGV） 5G、超宽带（Ultra Wide Band, UWB*） BI-REX 5G：射频数据、到达时间（Time of Arrival, ToA）；UWB*：距离、接收信号强度指示（Received Signal Strength Indicator, RSSI）、信道冲激响应（Channel Impulse Response, CIR） *注：UWB相关数据正处于更新版本数据集的发布准备阶段。 本数据集包含采集自意大利博洛尼亚大数据创新与研究卓越（Big Data Innovation & Research Excellence, bi-rex）中心工业试验线的5G数据。该试验线占地面积约300平方米，部署有生产设备、机器人与工业演示装置，覆盖了工作于3.7 GHz与26 GHz频段的私有5G网络，以及3.7 GHz频段的公共5G网络。 研究团队使用罗德与施瓦茨（Rohde & Schwarz）TSMA6移动网络扫描仪（隶属于R&S移动网络测试产品线），在试验线区域内的18个不同采样点位完成数据采集。该扫描仪配备用于数据地理映射与精准同步的GPS接收机，以及一套上下变频器，可通过两路独立天线同时采集698-3800 MHz频段与毫米波（26000 MHz）频段的4G与5G数据。 采集得到的数据包含射频参数（RSSI、参考信号接收功率（Reference Signal Received Power, RSRP）、信干噪比（Signal to Interference plus Noise Ratio, SINR）、参考信号接收质量（Reference Signal Received Quality, RSRQ））以及到达时间信息。本数据集适用于基于加权k近邻（Weighted k-Nearest Neighbors, WKNN）等机器学习算法的免测距定位任务。 在本仓库中，您可获取5G原始数据、数据处理脚本，以及WKNN定位算法的实现代码： ├── 5G_MNC_Processing.py ├── 5G 原始数据 │ ├── l10_5G.xlsx │ ├── l11_5G.xlsx │ ├── l12_5G.xlsx │ ├── l13_5G.xlsx │ ├── l14_5G.xlsx │ ├── l15_5G.xlsx │ ├── l1_5G.xlsx │ ├── l16_5G.xlsx │ ├── l17_5G.xlsx │ ├── l18_5G.xlsx │ ├── l2_5G.xlsx │ ├── l3_5G.xlsx │ ├── l4_5G.xlsx │ ├── l5_5G.xlsx │ ├── l6_5G.xlsx │ ├── l7_5G.xlsx │ ├── l8_5G.xlsx │ └── l9_5G.xlsx └── WKNN 模块 ├── apply_commonality_weighting_loo.m ├── Bi-Rex_Dataset.mat ├── compute_multi_param_distances_loo.m ├── compute_single_param_distances_loo.m ├── create_fingerprint_matrices_loo.m ├── create_global_unique_pcis.m ├── display_loo_results.m ├── main.m ├── perform_wknn_positioning_loo.m ├── plot_loo_results.m └── wknn_leave_one_out_positioning.m 如需获取更多信息，请参阅配套论文： N. Bouzar、L. De Nardis、G. Caso、M. Neri、F. Elbahhar 与 M.-G. Di Benedetto，《面向混合公私制中频段与毫米波5G部署场景下工业物联网的免测距定位技术》，已被2025年9月15日至17日于意大利博洛尼亚举办的IEEE通信网络标准会议下一代工业物联网通信网络专题研讨会收录。