Processing of multiple-receiver spaceborne arrays for wide-area SAR

Abstract

The instantaneous area illuminated by a single-aperture synthetic aperture radar (SAR) is fundamentally limited by the minimum SAR antenna area constraint. This limitation is due to the fact that the number of illuminated resolution cells cannot exceed the number of collected data samples. However, if spatial sampling is added through the use of multiple-receiver arrays, then the maximum unambiguous illumination area is increased because multiple beams can be formed to reject range-Doppler ambiguities. Furthermore, the maximum unambiguous illumination area increases with the number of receivers in the array. One spaceborne implementation of multiple-aperture SAR that has been proposed is a constellation of formation-flying satellites. In this implementation, several satellites fly in a cluster and work together as a single coherent system. There are many advantages to the constellation implementation including cost benefits, graceful performance degradation, and the possibility of performing in multiple modes. The disadvantage is that the spatial samples provided by such a constellation will be sparse and irregularly spaced; consequently, traditional matched filtering produces unsatisfactory results. We investigate SAR performance and processing of sparse, multiple-aperture arrays. Three filters are evaluated: the matched filter, maximum-likelihood filter, and minimum mean-squared error filter. It is shown that the maximum-likelihood and minimum mean-squared error filters can provide quality SAR images when operating on data obtained from sparse satellite constellations. We also investigate the performance of the three filters versus system parameters such as SNR, the number of receivers in the constellation, and satellite positioning error.