SYSTEM, DEVICE, AND METHOD FOR TIME-DOMAIN EQUALIZER TRAINING USING A TWO-PASS AUTO-REGRESSIVE MOVING AVERAGE MODEL

A system, device, and method for time-domain equalizer training uses a two-pass auto-regressive moving average model. A communication channel is first modeled using p1 poles and q1 zeros to form a first shortened channel impulse response having a first approximation H1(z)=B1(z)/1+A1(z), wherein q1, is greater than a predetermined cyclic prefix length. A time-mirrored image of the first shortened channel impulse response is then formed, and the resulting time-mirrored image of the first shortened channel impulse response is modeled using p2 poles and q2 zeros to form a second shortened channel impulse response having a second approximation H2(z)=B2(z)/1+A2(z), wherein q2 is less than or equal to the predetermined cyclic prefix length. The time-domain equalizer coefficients are determined by combining A1(z) and A2(1/z) with an appropriate amount of delay.