Random sampling strategies in multistatic SAR

On the design of vehicle trajectories for high quality multistatic SAR images

“…The standard bistatic trajectory performed extremely poorly, and did not approach the G M IN bound because of many eigenvalues very near to zero, which was previously observed [8]. This behavior can be attributed to the reasons set forth in guideline 4.…”

“…Finally, the random-sampling strategy considered in [8] is compared to the deterministic trajectories of this work. In this case, no special strategy or thinking is applied to the problem; the measurement ellipses are generated by a large swarm of SAR vehicles situated randomly above the image at random heights from h = 1 to h = 2 image widths.…”

“…The purpose of this work is to study optimal image-reconstruction techniques under these constraints. In [8] we considered the performance of the specific case of uncoordinated (or random) sampling in such a system. In this work, we develop design guidelines for general multistatic trajectories, and the application of these guidelines in coordinated UAV trajectory design.…”

“…The parameter G is a noise figure, giving a noise power gain of the reconstruction process. In [8], we developed asymptotic performance bounds for G when A is a large random matrix representing the measurements from a large swarm of randomly-placed SAR vehicles.…”

“…In [8] we introduced a performance metric for the image reconstruction process given a measurement noise variance σ 2 I :…”