A General Technique for Radar Resolution Analysis Based on a Quadratic Approximation

We develop a general technique to assess the resolution of radar systems. The technique approximates the matched filter response (MFR) with a quadratic function and solves the half-power beamwidth (HPBW) as a quadratic form. The approximation is shown to be accurate and consistent with several classical results in the radar literature. It is also shown to be directly related with the Cramér-Rao lower bound (CRLB) at a fixed signal to noise ratio (SNR) for the general class of complex exponential signals. This generalizes an old and widely used formula and allows the accuracy to be assessed together with the resolution. The results are demonstrated in several example scenarios, including direction of arrival (DOA) estimation with non-uniform multiple-input multiple-output (MIMO) arrays, near-field bistatic MIMO radars, forward looking MIMO synthetic aperture radar (SAR) systems and non-coherent radar networks.