Radar beam sharpening using an optimum FIR filter

Imagery that is obtained by noncoherent radar sensors suffers from a multitude of quality degradations. Among these, antenna-induced poor angular resolution is one of the most troublesome. We present a new approach to beam sharpening that improves the apparent resolution of the radar sensor several times depending on the signal-to-noise ratio. The technique consists of a compensating filter whose coefficients are obtained by minimizing the expected quadratic estimation error. This filter is then applied to sets of real 35 GHZ noncoherent radar imagery, and its performance is demonstrated. Results of the implementation of the technique on real-time digital signal processing hardware are also provided.