Structural-parametric identification of the sum of hyperbolic functions with additive noise

Abstract

The paper proposes a two-step method for structural-parametric identification of hyperbolic functions with additive noise. At the first stage, a linear difference equation is associated with the sum of hyperbolic functions. For the estimation of the parameters of the difference equation, the method of generalized total least squares (GTLS). To solve the GTLS problem, an augmented system equivalent to a biased normal system is applied. In the second stage, the missing parameters are estimated. The simulation results showed that the use of GTLS in the first stage improves the accuracy of the estimate compared to Ordinary Least Squares (OLS).