ABSTRACT
Adaptive filters can be employed in numerous other applications, such as system identification, automatic regulation, linear prediction, etc. The need for more-efficient and robust algorithms is, therefore, continuously increasing. This chapter presents an overview of some of the known algorithms and their basic properties. Adaptive infinite-duration impulse response digital filters are still being investigated as possible alternatives to finite impulse response (FIR) adaptive filters in applications where a high filter order is required. Transform-domain FIR adaptive filters are filters using gradient-based algorithms operating on a transformed input signal. An algorithm that employs such a technique is called least-mean-squares-Newton algorithm. The LS method was probably first used by Gauss in the late 18th century, and since then it has been applied in a vast number of areas. In every case the motivation is the same: to estimate the set of parameters that best fits a model to an observed phenomenon.
