Coherent MIMO radar imaging with model-aware block sparse recovery

We propose a framework for coherent MIMO radar imaging that allows to combine range imaging using iterative block sparse recovery with conventional, grid-less azimuth imaging. In each iteration of a block-sparse range recovery algorithm, we incorporate a projection step using an arbitrary azimuth imaging algorithm. The proposed imaging framework is hence not limited to azimuth imaging on a grid. This is in contrast to imaging approaches using sparse recovery that are available in the literature and that usually require rather coarse angular grids. Our simulation results show that this model-aware imaging framework achieves improved range and azimuth imaging performance in comparison to separate azimuth estimation after range imaging with block sparse recovery or the conventionally used matched filtering approach.