Automated Test Bed for Passive Millimeter Wave Imaging

• Nathaniel Riehl, Computer Engineering, University of Delaware
• Benjamin Mirotznik, Electrical Engineering, University of Delaware

• Vishal Saxena, Electrical and Computer Engineering, University of Delaware
• Thomas Dillon, Electrical and Computer Engineering, University of Delaware
• Mark Mirotznik, Electrical and Computer Engineering, University of Delaware

Passive millimeter wave imaging (PmmWI) has emerged as a promising technology for non-invasive sensing and security applications due to its unique ability to detect thermal radiation from objects and individuals quickly, with high resolution, and without the need for active illumination. In this study, we present the development and implementation of a comprehensive test bed dedicated to evaluating system architectures for W-band (75-110GHz) PmmWI, aiming to optimize image resolution and minimize distortions.

The test bed’s design incorporates three linear motion stages for collecting axial, planar, and cubic scans. It also includes millimeter wave sources, detectors, calibration equipment, radar-absorbent material, additively fabricated metaoptics, power supplies, and custom, 3D printed adjustable mounting. By utilizing this test bed, we ensure consistent measurements. Calibration procedures are employed to eliminate systematic errors and guarantee precise and reliable results.

Control algorithms were developed using MATLAB and executed through a custom application, allowing users to characterize lenses with various adjustable settings for precise measurements. The ability to precisely control linear movements allows the characterization of lens behavior across a range of focal distances, providing valuable insights into lens aberrations and their impact on image quality.

Scans of different lenses enable the measurement of their focal lengths by determining the relative intensity of received millimeter waves. This is accomplished by analyzing radiometer data recorded through raster scans, a scanning pattern that records data line by line. To recover small signals buried in significant noise, the radiometer data is processed through a software lock-in amplifier implemented in MATLAB. This lock-in amplifier is crucial for effectively eliminating noise, as millimeter waves radiated from human bodies are extremely small in magnitude.

Over the course of this study, we designed, constructed, and programmed a test bed for PmmWI. This test bed is capable of automatically recording millimeter waves and processing the data to generate images. The results obtained from this test bed contribute to the advancement of lightweight, high-quality lenses for enhancing the performance and portability of compact passive millimeter wave imaging systems.