非平稳环境下数据与模型协同驱动的网络传输容量优化测试数据集

数据集内容描述：该数据集是课题二核心指标（无线传输容量）进行第三方测试时产生的实验数据。基于Matlab仿真软件模拟生成无人机轨迹数据集、数模协同驱动上行吞吐量数据集和传统波束追踪算法的上行吞吐量数据集。针对5G毫米波场景验证基于数据与模型协同驱动的无人机无线容量优化方案对5G上行业务传输性能优化效果，主要性能指标为：提升上行无线传输容量30%。软件仿真模型描述的非平稳传输环境为以下条件：1.多架无人机与单台地面基站之间点对点传输；2.无人机以随机轨迹飞行，地面基站位置固定；3.双方之间为视距信道，无遮挡物和反射路径。通过基于高斯过程机器学习方法，基站端灵活调整波束成形方案并优化波束跟踪，达到提高上行信噪比的目的，从而提升无线网络传输容量，实现无人机与基站的高速数据传输。其中001为轨迹数据，大小为(100,3，10)数组，其中100为时隙数，3为x,y,z坐标，10为无人机数目；002为测试结果数据，包含10组测试上行容量数据。003为所搭建的无线传输容量硬件测试平台在高动态非平稳环境下5G上行业务的无线传输容量的指标组织进行三方测试，测试时所得的电磁数据。
数据来源：该数据集是课题二核心指标（无线传输容量）进行第三方测试时产生的实验数据，来源于仿真软件测试以及搭建平台纪录的实测数据。
采集方式：基于Matlab仿真软件模拟生成无人机轨迹数据集、数模协同驱动上行吞吐量数据集和传统波束追踪算法的上行吞吐量数据集。5G终端周围电磁环境数据使用信号分析仪远程控制采集分析，指标项由雾计算节点监测采集。在无线传输容量硬件测试平台上进行，通过信号分析仪实时采集分析5G终端当前周围电磁环境数据，并将得到的空闲频谱推荐列表发送至雾计算节点，雾计算节点根据数据模型协同驱动算法对基站的频点波束进行调整，并实时监测调整前后的5G上行业务传输吞吐量。本数据集包含5G终端周围电磁环境原始数据。
加工方式：基于Matlab仿真软件模拟生成无人机轨迹数据集、数模协同驱动上行吞吐量数据集和传统波束追踪算法的上行吞吐量数据集，平台原始数据利用numpy加工处理，以.mat以及.npz格式导出
设备情况：本次测试仪器包含5G终端、无人机、信号分析仪、天线、雾计算节点、信号源、信道仿真仪、基站速率监测电脑以及5G核心网基站设备（包含RRU、BBU、核心网和接收服务器等部分）、英国Spirent公司的SpirentTestCenterN4U仪表方案。其中STC-N4U提供多接口高流量监控模型，能够对物理网络的吞吐容量执行压力测试。提供了GUI界面和Tcl远程自动化脚本接口两种配置方式，可以与软件仿真系统进行对接。本次测试部署了一台STC-N4U设备，使用两个接口，用作双向的网络容量吞吐打流，可生成打流所需的吞吐数据包，并测试接收到的网络吞吐的容量、时延等参数。 数据集包含一个文件夹、一个数据集说明文件.docx、一个论文和一个三方测试报告。文件夹名称为数据集实体文件，包含2个data文件、1个mat文件，1个DS-Store文件和2个xlsx文件，数据内容总量约为831MB。

Dataset Content Description: This dataset is the experimental data generated during third-party testing of the core indicator (wireless transmission capacity) of Project 2. Matlab simulation software is used to generate three types of datasets: UAV trajectory dataset, uplink throughput dataset driven by data-model collaboration, and uplink throughput dataset of traditional beam tracking algorithm. It is used to verify the optimization effect of the data-and-model collaborative driven UAV wireless capacity optimization scheme on 5G uplink service transmission performance in 5G millimeter-wave scenarios, with the main performance target of increasing uplink wireless transmission capacity by 30%. The non-stationary transmission environment modeled by the software simulation meets the following conditions: 1. Point-to-point transmission between multiple UAVs and a single ground base station; 2. UAVs fly along random trajectories while the ground base station remains fixed; 3. There is a line-of-sight (LoS) channel between the two sides, with no obstacles or reflection paths. By adopting Gaussian Process Machine Learning method, the base station can flexibly adjust the beamforming scheme and optimize beam tracking, so as to improve the uplink signal-to-noise ratio (SNR), thereby enhancing wireless network transmission capacity and realizing high-speed data transmission between UAVs and the base station. Among them, 001 is trajectory data, which is an array of size (100, 3, 10), where 100 represents the number of time slots, 3 represents x, y, z coordinates, and 10 represents the number of UAVs; 002 is test result data, including 10 sets of uplink capacity test data; 003 is the electromagnetic data obtained during the third-party test of 5G uplink service wireless transmission capacity indicators of the built wireless transmission capacity hardware test platform in high-dynamic non-stationary environments.
Data Source: This dataset is the experimental data generated during third-party testing of the core indicator (wireless transmission capacity) of Project 2, which comes from simulation software test data and measured data recorded by the built test platform.
Collection Method: The UAV trajectory dataset, uplink throughput dataset driven by data-model collaboration, and uplink throughput dataset of traditional beam tracking algorithm are simulated and generated based on Matlab simulation software. The electromagnetic environment data around 5G terminals is collected and analyzed via remote control of a signal analyzer, and the indicator items are monitored and collected by fog computing nodes. The test is carried out on the wireless transmission capacity hardware test platform: the signal analyzer collects and analyzes the electromagnetic environment data around the 5G terminal in real time, and sends the obtained free spectrum recommendation list to the fog computing node. The fog computing node adjusts the frequency point and beam of the base station according to the data-model collaborative driven algorithm, and monitors the 5G uplink service transmission throughput before and after the adjustment in real time. This dataset contains the original electromagnetic environment data around the 5G terminal.
Processing Method: The UAV trajectory dataset, uplink throughput dataset driven by data-model collaboration, and uplink throughput dataset of traditional beam tracking algorithm are simulated and generated based on Matlab simulation software. The original platform data is processed using numpy, and exported in .mat and .npz formats.
Equipment Setup: The test instruments and equipment used this time include: 5G terminal, UAVs, signal analyzer, antenna, fog computing node, signal source, channel simulator, base station rate monitoring computer, and 5G core network base station equipment (including RRU, BBU, core network, receiving server and other components), as well as the SpirentTestCenterN4U instrumentation solution from Spirent Communications (UK). STC-N4U provides a multi-interface high-traffic monitoring model, which can perform stress testing on the throughput capacity of the physical network. It supports two configuration methods: GUI interface and Tcl remote automation script interface, and can be docked with the software simulation system. One STC-N4U device was deployed in this test, using two interfaces for bidirectional network capacity throughput testing, which can generate throughput data packets required for traffic testing, and test parameters such as received network throughput capacity and latency.
This dataset includes one folder, one dataset description document.docx, one paper, and one third-party test report. The folder named Dataset Entity Files contains 2 data files, 1 mat file, 1 DS_Store file, and 2 xlsx files, with a total data content of approximately 831 MB.

搜集汇总

背景与挑战

背景概述

该数据集是针对5G毫米波场景下无人机无线容量优化的测试数据，包含仿真生成的无人机轨迹数据集、上行吞吐量数据集及硬件测试平台实测数据，旨在验证数据与模型协同驱动方案对提升上行无线传输容量的效果。数据集总大小约831MB，包含多种格式的文件，适用于相关领域的研究和开发。