Near-Field Synthetic Aperture Radar Imaging Based on a MIMO FMCW Radar

Synthetic aperture radar (SAR) imaging has traditionally relied on the assumption that the target lies in the far field of the radiating antenna, where the impinging wavefronts can be approximated as locally planar. However, in a growing number of short-range applications, including security screening, non-destructive testing, and compact biomedical imaging, targets are located within the near field of the antenna. In this case, the pronounced spherical curvature of the wavefronts invalidates the conventional far-field approximation and introduces a coupling between the range and cross-range dimensions of the imaging problem. This Thesis reviews the fundamental principles of far-field SAR imaging and extends the theoretical framework to the near-field regime. A near-field signal model is derived, together with the corresponding resolution limits, and focusing strategies based on the angular spectrum method and time-domain back projection are investigated. The resulting near-field imaging framework is experimentally implemented using a 60 GHz frequency-modulated continuous-wave (FMCW) multiple-input multiple-output (MIMO) radar combined with a mechanical scanning system. Image reconstruction is performed using the back projection algorithm and further refined through a minimum entropy autofocus stage aimed at compensating for residual phase errors. The proposed system is validated on several experimentally acquired scenes containing both metallic and non-metallic objects. Image quality and geometric accuracy are assessed under both single-input single-output (SISO) and MIMO configurations. The experimental results demonstrate that near-field back projection, combined with entropy-based autofocus, enables accurate and high-resolution radar imaging at short ranges.