A Millimeter-Wave Dynamic Antenna Array for Classifying Objects Via Sparse Fourier Domain Sampling

Abstract

We present a dynamic antenna array operating at 38 GHz for detecting strong Fourier domain (spatial frequency) features of a scene to demonstrate the potential for object classification. Sharp edges on objects manifest as broad-spectrum spatial frequency signals in the Fourier domain at angles orthogonal to the edge direction. By sampling only a sparse ring of spatial frequency information, these broad spatial frequency responses can be identified, and objects can be classified without full image reconstruction. The technique is based on active interferometric imaging, where noise signals are transmitted to mimic thermal radiation, thereby allowing the use of the Van Cittert-Zernike theorem to capture scene intensity information in the Fourier domain. Sparse spatial frequency information is captured by a dynamic 38 GHz two-element interferometric antenna array that rotates in the spatial domain to synthesize a ring filter in the spatial frequency domain. We experimentally demonstrate that edge-induced broad-spectrum responses can be identified for the classification of simple objects using this method.