An Analysis of a Near-End Crosstalk Cancelation System That Uses Adaptive Filters

An analysis of a near-end crosstalk (NEXT) cancelation system that uses adaptive digital filters is described. The analysis is based on two well-known models for the NEXT coupling factor, the Bradley and Lin models, and yields the minimum number of adaptive filters required to reduce the NEXT below a prescribed level to within a defined confidence factor. With the minimum number of adaptive filters known, the required computational resources for the application at hand can be estimated. The analysis is further extended to practical situations where the largest NEXT signals chosen for elimination are incorrectly detected, and estimates of the minimum and maximum increase in the uncanceled NEXT due to incorrect detection are then deduced. Simulations show that the estimated minimum number of adaptive filters required and the maximum and minimum increase in uncanceled NEXT due to incorrect detection are fairly close to corresponding estimates obtained on the basis of measurements for both the Bradley and the Lin models. Therefore, by using the proposed analysis the minimum number of adaptive filters can be deduced without the need for time-consuming and expensive experiments.