Sparse signal processing on estimation grid with constant information distance applied in radar

Radar obtains its parameters on a grid whose design supports resolution of underlying radar processing. Existing radar exploits a regular grid although the resolution changes with stronger echoes at shorter ranges. We compute the radar resolution from the intrinsic geometrical structure of data models that is characterized in terms of the Fisher information metric. Based on the information-based approach, we design an estimation grid whose cells have a constant Fisher information distance. In addition, we explore how this information-based grid can suit radar processing in practice and propose information-based processing on such an irregular estimation grid by applying the sparse signal processing from compressive sensing. Accordingly, the grid was adjusted to the sensing incoherence needed in sparse signal processing by setting a lower bound for the cell size. Our approach enables an adaptive estimation grid that can be adjusted with respect to the available resolution, the desired sensing incoherence, available computational power, and required operational priorities. The information-based design and processing are illustrated in a one-dimensional case of range estimation.