An index of system nonlinearity and its estimation

This paper proposes an index representing the degree of system nonlinearity. It is often important in system modeling to know whether the system under consideration can be regarded as a linear system. The proposed index takes values ranging from zero to one where the index takes the value zero if the system is linear and approaches to one as the output signal power which is unpredictable by the linear system increases. A method to estimate the index using a multilayer perceptron is proposed. A parametric system whose nonlinearity index naturally changes from zero to one according to the system parameter has been introduced and a computer simulation on that system confirmed the validity of the method.