Vehicular MIMO SAR Imaging in Multipath Environments

Abstract

Current synthetic aperture radars (SAR) are most effective in simple, open terrains where direct-path propagation can be assumed. For ground-vehicle based SAR from a moving platform, however, strong multipath scattering off terrain features with the same direction of arrival and delays as direct path returns, results in serious imaging artifacts. Moreover, the dilemma between spatial coverage and azimuth resolutions and the along track sampling constraints are limiting factors which have thus far been precluded vehicular SAR in urban areas. In this thesis, multi-input multi-output (MIMO) forward looking synthetic aperture radar is developed for imaging from a moving ground vehicle in urban multipath environments. MIMO methods are utilized to improve SAR images by suppressing directions of departure which would otherwise be multipath scattered and added to direct path returns by applying a three dimensional non-causal spatial filter in the direction-of-departure (DOD), direction-of-arrival (DOA), and Doppler-frequency domains which also enables the image with wide-swath and high resolution simultaneously. Both conventional and adaptive MIMO SAR methods are presented and compared in a multipath imaging simulation. The results suggest MIMO SAR offers substantial gains versus conventional SIMO imaging in presence of multipath.