RF Holography and Perception in Static Environments: Dataset

Four distinct datasets are provided for microwave imaging purposes.

Ray-tracing_human_body.zip contains the ray-tracing simulation results for a male body model, while Ray-tracing_human_body_ground.zip includes the same simulation with an additional ground plane. The detailed data files and formats are as follows:

[Filename*].obj: Wavefront .obj file, 3D model used for ray tracing simulation.

[Filename*]_EM.mat: containing ray-tracing results, specifying the received electric fields at each receiver induced by transmitters at each frequency. It contains the following arrays

freq : simulation frequencies

results : a vector of structures, each containing the attributes {‘E’, ‘freq’}. The length of the vector corresponds to the length of ‘freq’, and each ‘E’ is a 3-dimensional array of size (num_of_Tx, num_of_Rx, 3). This array stores the 𝑥-, 𝑦-, and 𝑧-components of the electric fields at each receiver induced by transmitters.

tx : 𝑥-, 𝑦- and 𝑧-coordinates of each transmitter with a size of (3, num_of_Tx)

rx : 𝑥-, 𝑦- and 𝑧-coordinates of each receive with a size of (3, num_of_Rx)

aimg_[Filename*].mat: MATLAB imaging data generated by the back-projection algorithm, where in particular

aimg.X : 𝑥-positions of the imaging domain

aimg.Y : 𝑦-positions of the imaging domain

aimg.Z : 𝑧-positions of the imaging domain

aimg.fieldnames : names of the included different Cartesian vector components of the imaged sources, e.g., aimg.fieldnames = {‘𝑥’, ‘𝑦’, ‘𝑧’}

aimg.x : 𝑥-polarization of the polarimetric imaging results

aimg.y : 𝑦-polarization of the polarimetric imaging results

aimg.z : 𝑧-polarization of the polarimetric imaging results

FEKO_Simulations.zip includes two full-wave simulations conducted with FEKO. One simulation features three perfectly electrically conducting (PEC) spheres illuminated by a Hertzian dipole, while the other involves a PEC plate illuminated by a horn antenna.

[Filename*].efe: ASCII file containing simulated near-field data from the full wave simulation data FEKO, where the key parameters such as frequencies, measured positions, real and imaginary components of the electric fields are included.

aimg_[Filename*].mat: MATLAB imaging data generated by the algorithm introduced in [1], [2], and HOLDEN Deliverable D3.2, where in particular

aimg.X : 𝑥-positions of the imaging domain

aimg.Y : 𝑦-positions of the imaging domain

aimg.Z : 𝑧-positions of the imaging domain

aimg.fieldnames : names of the included different Cartesian vector components of the imaged sources, e.g., aimg.fieldnames = {‘𝑥’, ‘𝑦’, ‘𝑧’}

aimg.x : 𝑥-polarization of the polarimetric imaging results

aimg.y : 𝑦-polarization of the polarimetric imaging results

aimg.z : 𝑧-polarization of the polarimetric imaging results

Measurements_UAV_AnechoicChamber.zip provides measurements of a UAV collected at the Chair of High-Frequency Engineering, Technical University of Munich.

[Filename*].nsiascii : containing near-field measurement data, typically S12 or S21 parameters in dual-polarization from a vector network analyzer (VNA). The frequency information is embedded in the filename, e.g., 6000000000.nsiascii indicates a frequency of 6 GHz. Other parameters, such as probe positions, VNA setups, as well as the amplitude and phase of the measured signals, are also included in this file.

aimg_[Filename*].mat: MATLAB imaging data generated by the algorithm introduced in [1], [2], and HOLDEN Deliverable D3.2, where in particular

aimg.X : 𝑥-positions of the imaging domain

aimg.Y : 𝑦-positions of the imaging domain

aimg.Z : 𝑧-positions of the imaging domain

aimg.fieldnames : names of the included different Cartesian vector components of the imaged sources, e.g., aimg.fieldnames = {‘𝑥’, ‘𝑦’, ‘𝑧’}

aimg.x : 𝑥-polarization of the polarimetric imaging results

aimg.y : 𝑦-polarization of the polarimetric imaging results

aimg.z : 𝑧-polarization of the polarimetric imaging results

TUM_HFT_WiFi_Mannequins_DATASET.zip contains three measurements from a planar office scanner system installed on a side wall of a typical office room, also at the Chair of High-Frequency Engineering, Technical University of Munich. One video file and one PowerPoint presentation illustrating the measurement campaign are provided. The three different measurements belong to three different arrangements of plastic/wooden mannequins covered with zinc-aluminum spray, where either none one or two mannequins are present within the office.

mannequin_[Filename*].mat: Continuous wave measurement data from a planar scanner. The data is stored in MATLAB data format. The transmission and reflection coefficients between a fixed transmit antenna and a moving (on the planar scanner) antenna, both vertically polarized, are measured for various arrangements of mannequins within a typical office environment. 

freqs [21x1]: frequencies of measurements

S11 [162x80x21]: complex-valued reflection coefficient for each measurement location and frequency

S21 [162x80x21]: complex-valued transmission value for each measurement location and frequency

X/Y [162x80]: 𝑥- and 𝑦-location of the measurement samples

More detailed descriptions of the dataset can be found in HOLDEN Deliverable D3.3. The ray-tracing simulation framework and the holography imaging method are introduced in Deliverables D3.1 and D3.2, respectively.

Part of the data and results are found in:

[1] Q. Wang and T. F. Eibert, “Microwave imaging of electromagnetic wave absorbers in an antenna measurement chamber,” in IEEE Int. Symp. Antennas Propag., Jul. 2024, pp. 1139–1140.[2] Q. Wang, M. M. Saurer, and T. F. Eibert, “3D near-field passive radar imaging utilizing phase corrected frequency domain inverse source reconstructions,” IEEE Trans. Antennas Propag., submitted, also to TechRxiv.