基于NLOS/LOS信道稀疏度特征的38.901仿真信道数据集

系统级仿真中信道生成逻辑完全遵循3GPP协议38.901。对于每一个用户到每一个基站之间的链路，先生成大尺度衰落信息，例如pathloss、shadowfading，再进行多径构建。多径分为cluster、ray，依次计算每一条径的时延、功率、角度，然后计算极化系数、初始相位，最后进行多径合成生成每个通道对之间的时域信道。每一个cluster在时域上表现为一个冲激，根据频率采样点数对时域响应进行DFT，得到每个子载波上的频域信道。根据上下行信道互易性，将下行信道矩阵进行共轭转置即得到上行信道矩阵，同理按顺序输出即可。之后将初始采集数据中的一个信道样本点在空域做FFT变换，在频域做IFFT变换，形成较为稀疏的角度-时延域，将角度-时延域上的数据进行排序，并计算能量的累计，从而得到处理后数据。数据集内容：本数据集为仿真模拟数据，包括NLOS和LOS原始信道和处理后的信道。
数据类型与格式：数据类型为表格文件，数据格式为txt，mat，pdf。
采集时间与地点：2022年11月，北京邮电大学。
数据产生及加工处理方式：数据集采集通过远程使用移动研究院InspurNF8480M5服务器，处理器Intel(R)Xeon(R)Gold6148CPU@2.40GHz，ubuntu20.04系统仿真信道，windowsserver2016系统处理数据。使用netbeans软件编译c++代码，htcondor软件搭建集群，matlab2022b软件处理数据。分析数据通过matlab，版本为matlab2022b,具体采用了二维离散傅立叶变换等来加工和处理。

The channel generation logic in the system-level simulation fully complies with 3GPP Protocol 38.901. For each user-to-base station link, large-scale fading information such as path loss and shadow fading is first generated, followed by multipath construction. Multipath components are divided into clusters and rays. The delay, power, and angular parameters of each path are calculated sequentially, followed by the polarization coefficients and initial phase. Finally, multipath combining is performed to generate the time-domain channel between each pair of antennas. Each cluster exhibits an impulse in the time domain. The discrete Fourier transform (DFT) is applied to the time-domain response based on the number of frequency sampling points to obtain the frequency-domain channel for each subcarrier. According to the uplink-downlink channel reciprocity, the uplink channel matrix can be obtained by taking the conjugate transpose of the downlink channel matrix, and the output can be generated in the same sequential order. Subsequently, one channel sample point from the initially collected data is subjected to FFT in the spatial domain and IFFT in the frequency domain, forming a relatively sparse angle-delay domain. The data in the angle-delay domain is sorted, and the cumulative energy is calculated to obtain the processed data.

Dataset Content: This dataset is simulation data, including both NLOS and LOS raw channels and processed channels.

Data Type and Format: The data is in tabular file format, with supported formats including TXT, MAT, and PDF.

Collection Time and Location: November 2022, Beijing University of Posts and Telecommunications.

Data Generation and Processing Methods: The dataset was collected by remotely using the Inspur NF8480M5 server from China Mobile Research Institute, which is equipped with an Intel(R) Xeon(R) Gold 6148 CPU @ 2.40GHz. The Ubuntu 20.04 operating system was used for channel simulation, while the Windows Server 2016 operating system was used for data processing. C++ code was compiled using NetBeans software, a cluster computing environment was deployed using HTCondor software, and data processing was carried out with MATLAB 2022b. Data analysis was conducted using MATLAB 2022b, with two-dimensional discrete Fourier transforms and other methods adopted for data processing and refinement.

搜集汇总

背景与挑战

背景概述

该数据集是一个仿真模拟数据，包含NLOS和LOS原始信道及处理后的信道，遵循3GPP协议38.901进行系统级仿真。数据格式包括txt、mat、pdf等，采集于2022年11月北京邮电大学，通过特定服务器和软件（如matlab2022b）进行仿真和处理，适用于通信技术领域的研究。