Kalman estimation of frequency-domain interference reduction coefficients for angle-modulated signals

The problem addressed is that of realizing efficient and effective interference reduction filters used to enhance angle-modulated, keyed signals of interest embedded in very severe narrowband interference environments. Effectiveness is measured in terms of the ability of the filters to attenuate interferers in an optimal (in the maximum likelihood sense) manner, and to converge rapidly. The first measure of effectiveness is achieved with very long (on the order of 8000 points) impulse response filters; the latter is achieved with a Kalman-based approach to filter coefficient estimation. Efficiency, measured in terms of multiply-adds, is achieved by processing entirely within the frequency domain, where the efficiency of the FFT algorithm can be exploited.