A novel nonlinearity measure

Recent publications in adaptive filtering have shown great interest in the use of nonlinear structures. In particular the quadratic Volterra adaptive filter has proved very popular. In this paper we start by demonstrating the need for nonlinear adaptive filters. Subsequently we present a method for the estimation of the level of nonlinearity of a system so as to be able to select the appropriate adaptation algorithm. Lastly, we apply the nonlinear measure to two examples from the open literature.