Comparative study of the generalized adaptive neural filter with other nonlinear filters

The generalized adaptive neural filter (GANF) is a new type of adaptable filter. The GANF relies upon neural functions to set up a filtering operation. The authors study a few of the possible neural networks which can be used in a GANF. The capabilities of the neural nets are examined and the filtering abilities of the GANF are obtained through simulation. While the GANF structure used is somewhat simplified, the filter is also compared with other nonadaptive filters. These filters provide a reference so that relative performance can be more realistically judged.<>