DOA Estimation Dataset
收藏ieee-dataport.org2025-03-24 收录
下载链接:
https://ieee-dataport.org/documents/doa-estimation-dataset
下载链接
链接失效反馈官方服务:
资源简介:
The experimental measurement model includes the transmitter, receiver and signal processing parts:The transmitter part includes 1 antenna, capable of transmitting signals at a frequency of 1090 MHz.The signal supplied to the transmitting antenna is generated by the FMCOMMS5 transceiver circuit.To generate the transmit signal, a Matlab program is implemented on the computer to control the FMCOMMS5 circuit through the ZC706 FPGA board.The control process is transmitted via Ethernet communication using a LAN (Local Area Network) cable.Specifically, the signal is generated in the Matlab program at the baseband of 100 kHz, the carrier frequency is set to 1090 MHz.The process of raising the frequency from 100 kHz to 1090 MHz is done inside the AD9361 chip.The transmission of the 1090 MHz frequency signal by the transmitting antenna is carried out continuously.The transmitting antenna is placed at a distance of 3.0 m from the receiving antenna.This distance ensures the transmitter source is in the far field with respect to the four-element receive antenna array.The receiver part consists of four Vivaldi antennas, connected to the four receiver channels of the FMCOMMS5 circuit by four high-frequency coaxial cables with SMA connectors.Vivaldi antennas are placed at a distance of λ/2, which is 13.76 cm with an error of about 0.2 cm.As shown in Figure 3.15, receiver channels Rx_1, Rx_2 are connected to AD9361_1, receiver channels Rx_3, Rx_4 are connected to AD9361_2.These four receiving channels have the ability to synchronize phase with each other.The FMCOMMS5 circuit connects to the FPGA ZC706 control circuit through the FMC interface.Signal acquisition is also controlled by a Matlab program on the computer. The Matlab program will set parameters with a sampling rate of 20 MHz, the received signal will have a baseband of 100 kHz and the number of samples in one recording is 2000. Due to IQ, the received signalobtained from four channels will have size 4×2000 in complex number form. The measurement process to create the data set is performed iteratively. Specifically, instead of having to move the transmitting antenna to different angular positions, this change will be done by rotating the array of four receiving elements around a symmetrical axis, with a protractor to determine the rotation angle. The value shown on the protractor is determined to be the correct angle value of the transmitting antenna. Note that the value read from the protractor may also contain errors, but they are very small. The measurement process is performed sequentially for angles from -60 to 60 degrees with a step of 1 degree.Each corner performs 1000 signal receptions, so the total number of received signals for all corners is 121000 signals, size 4×2000 in complex number form.Each received signal is saved to a file in *.mat format to serve further processing as well as verifying algorithms.
该实验测量模型包括发射器、接收器和信号处理部分:发射器部分包含1个天线,能够以1090 MHz的频率传输信号。供应给发射天线的信号由FMCOMMS5收发机电路生成。为生成传输信号,计算机上实现了Matlab程序以通过ZC706 FPGA板控制FMCOMMS5电路。控制过程通过局域网(LAN)电缆使用以太网通信进行传输。具体而言,信号在Matlab程序中以100 kHz的基带频率生成,载波频率设置为1090 MHz。将频率从100 kHz提升至1090 MHz的过程在AD9361芯片内部完成。1090 MHz频率信号的传输由发射天线持续进行。发射天线放置在接收天线3.0米远处,此距离确保发射源相对于四元素接收天线阵列处于远场。接收器部分由四个Vivaldi天线组成,通过四根带有SMA连接器的高频同轴电缆连接到FMCOMMS5电路的四个接收通道。Vivaldi天线放置在λ/2的位置,即13.76 cm,误差约为0.2 cm。如图3.15所示,接收通道Rx_1、Rx_2连接至AD9361_1,接收通道Rx_3、Rx_4连接至AD9361_2。这四个接收通道具有相互同步相位的能力。FMCOMMS5电路通过FMC接口连接至FPGA ZC706控制电路。信号采集也由计算机上的Matlab程序控制。Matlab程序将设置采样率为20 MHz的参数,接收信号将具有100 kHz的基带,单次记录中的样本数为2000。由于IQ,从四个通道获得的接收信号将以4×2000的复数形式呈现。创建数据集的测量过程是迭代进行的。具体来说,无需移动发射天线到不同的角度位置,这一变化将通过围绕对称轴旋转四个接收元素阵列来完成,使用量角器确定旋转角度。量角器上显示的值被确定为发射天线的正确角度值。请注意,量角器上读取的值也可能包含误差,但它们非常小。测量过程以1度的步长从-60度到60度依次进行。每个角落执行1000次信号接收,因此所有角落接收到的信号总数为121000个,以4×2000的复数形式呈现。每个接收信号都保存为*.mat格式的文件,以供进一步处理以及验证算法。
提供机构:
ieee-dataport.org
搜集汇总
数据集介绍

背景与挑战
背景概述
DOA Estimation Dataset是一个用于方向估计的实验测量数据集,包含121000个复数形式的信号数据,采集自四元素接收天线阵列在不同角度下的信号接收。数据集适用于信号处理和深度神经网络的研究。
以上内容由遇见数据集搜集并总结生成



