Fast initialization of data-driven Nyquist in-band echo cancellers

A echo canceller fast training scheme for data-driven Nyquist in-band echo cancellers is presented. This scheme simultaneously estimates the desired near and far echo canceller coefficients by sending a special periodic training sequence and correlating a segment of the sequence with the real echo samples. The requirements and the generation of the training sequence are discussed. It is shown that the fast training method can also provide the parameters for the fast initialization of the phase roll compensator in the echo canceller, and thus complete the fast training of the entire echo canceller. Compared to the conventional LMS training and other training methods, this scheme provides accurate estimates of the echo canceller coefficients in a significantly reduced training time.< >