Design of radar grid cells is based on the resolution of radar processing. In typical radar, the cell size is constant, i.e. the estimation grid is regular. However, the resolution is not constant, as it increases at closer ranges due to stronger radar echoes. Radar resolution is computed from the intrinsic geometrical structure of data models that is conveniently characterized in terms of the Fisher information metric. Based on these information-geometry calculations, an estimation grid is designed whose grid cells have a constant Fisher information distance. In addition, high-resolution radar processing on such an irregular estimation grid is proposed by applying the sparse-signal processing from compressive sensing. This novel information-based approach in radar design and processing is illustrated in a range-only case.