Identification of fractional-order continuous-time hybrid Box-Jenkins models using refined instrumental variable continuous-time fractional-order method

Abstract

This paper illustrates the identification of a Box-Jenkins model from sampled input and output data. This is achieved by extending a refined instrumental variable continuous-time (RIVC) method to a refined instrumental variable continuous-time fractional-order (RIVCF) method. The model is a hybrid of continuous and discrete-time as well as fractional and integer-orders. The model consists of a fractional-order linear continuous-time (FLC) transfer function and noise. The FLC transfer function represents the noise free system and the noise represents an integer-order discrete-time autoregressive moving average (ARMA). Monte Carlo simulation analysis is applied for illustrating the performance of the proposed RIVCF method.