Estimation of non-integer order model based on least-squares technique and instrumental variables: Hn model

Abstract

This contribution has as its goal to identify of fractional order systems using least-squares (LS) and instrumental variable (IV) technique. System estimation can be considered as a necessary step in control theory. We introduce, in this paper, a new technique that permits us to estimate non-integer model. This method employs a linearization technique to find a linear equation, and then estimates unknown parameters using least squares approach. In noisy output context, this method offers a biased estimation; the proposed solution is to use a new approach of instrumental variable technique. Different simulations test are presented.