Fast-Adapting Matrix Filtering

Implementation and convergence of adaptive processors are of continuing interest. To this end, a fast-adapting, linear, general recursive matrix-filtering algorithm is described. Two complex xoefficients are optimized at each recursion. A signal and interference estimation procedure is included. Results for some low-resolution, frequency domain beamformer trials with two targets, and using a single sample set, are presented. Considerable adaptive improvement occurs, along with target resolution, compared with a correlation beamformer. The algorithm shows rapid and direct convergence.