Computing the feedback filter of the decision feedback equalizer at the FFT speed

This paper is devoted to the efficient computation of the feedback and feedforward filters of the decision feedback equalizer. The key operation in the computation of the filters is the spectral factorization. The paper proposes to adopt the efficient method for spectral factorization due to Kolmogoroff, which runs at the speed of FFT. Since the duration in the time domain of the impulse response obtained from the spectral factorization is not under the control of the designer, truncation of that impulse response becomes necessary. We propose a method that mitigates the performance loss due to the truncation, and validate the proposal by computer simulation on a variety of channels.