In this repository you can find python scripts for microwave imaging using a vector network analyzer (VNA) and a frequency modulated continuous wave (FMCW) radar system.
The FMCW radar system is realized with commercially available components. More details can be found in my publications (https://ieeexplore.ieee.org/document/10305102) and (https://ieeexplore.ieee.org/document/10590586).
As a VNA I used the E8361C from Keysight.
For the imaging with the vector network analyzer the delay-and-sum (DAS) beamforming algorithm is implemented. For the FMCW radar system the standard synthetic aperture radar (SAR) approach is used. However, both algorithms are ultimatively just the application of a matched filter.
To compute an image based on a set of s-parameter measurements, it is necessary that you have a folder with the corresponding .s2p touchstone files. A separate touchstone file is required for every antenna position. The name of the file indicates the measurement number, e.g. 0.s2p, 1.s2p, 2.s2p and so on. To generate the image, it is necessary to run the script Imaging_VNA.py. In the variable 'path' you just have to specfiy the folder with your touchstone files. The final image is saved as a pickle file in the folder 'Pickle Files' in the directory. If this folder does not already exists, it will be created automatically.
For the image generation with the FMCW radar system you have to run the file Prepare_Radar_Data.py.
First the phase compensated range FFT of every measurement of each measurement is computed and combined in a matrix. This matrix and the corresponding distance vector is saved in two pickle files spectrum.pkl and distance_corrected.pkl. For this, it is necessary that there is a folder named 'Pickle Files' in the directory. If this folder does not already exists it will be created automatically. Then the image is calculated based on spectrum.pkl and distance_corrected.pkl. The input data for Prepare_Radar_Data.py are .csv files that contain the output data of the radar system that was recorded using an oscilloscope. This includes the time values, the I-part of the IF signal and the Q-part of the IF signal. In the variable 'path' you just have to specfiy the folder with your .csv files. The name of the file indicates the measurement number, e.g. 0.csv, 1.csv, 2.csv and so on.
The actual preparation of the radar data is done in Radar_Evaluation_Modul.py. This file contains the class 'radar_measurement_evaluation'. The imaging steps are implemented in Radar_Imaging_Modul.py. This file contains the class 'radar_imaging'.
The final image is saved as a pickle file in the folder 'Pickle Files' in the directory.
All the signal processing steps that are implemented can be found in the .pdf file 'Flow_Chart_Comparison'.
The repository contains two scripts (Ideal_VNA_Data_Generator.py and Ideal_Radar_Data_Generator.py) for the generation of ideal VNA and radar data. The ideal data is saved in the folders 'Ideal Data VNA' and 'Ideal Data Radar'. If these folders do not already exist they will be created automatically. The default settings are chosen so that you simply have to download the entire repository. Then you have to execute Ideal_VNA_Data_Generator.py and Imaging_VNA.py to get a VNA image. To generate a radar image the correct order is Ideal_Radar_Data_Generator.py and Prepare_Radar_Data.py.
For the generation of the ideal data the user has to specify an array of (x,y)-coordinates. To compute the ideal FMCW radar signals the corresponding IF frequency and phase shift based on the ideal signal model is calculated. For the ideal VNA data a set of microstrip lines with the corresponding lengths is simulated using the scikit-rf package.
If you have any comments feel free to write me an e-mail to [email protected].
Exemplary measurement data is unfortunately too large for GitHub. If you are interested you can write me an e-mail and I will provide a Google Drive link where you can find measurement data.
I hope you find my scripts helpful!