Finite Impulse Response Adaptive Filters with Increased Convergence Speed

Abstract

Generally speaking, the operation of adaptive filtering can be divided in two phases: the phase of the parameter estimation convergence and the phase of the following of the variable parameters of the filter. In the convergence phase, the adaptive algorithm starts from some initial value of the vector of unknown filter parameters and tends to converge to optimal parameter values, in the sense of the adopted performance index (quality criterion) like e.g. the MSE algorithm.