Efficient parameter estimation for a class of linear continuous-time systems using the HMF-method

Specifically, the main purpose of this paper is to analyse the use of a frequency weighted least-squares (WLS) fit via the Hartley modulating function (HMF-) method to estimate the parameters of various linear continuous-time systems and to investigate computational considerations associated with the proposed identification scheme. The proposed method does not require the observation of all state variables. It assumes the a priori knowledge of the system structure with unknown parameters. The advantage of the HMF-method is that it handles very elegant the time-derivative problem. This approach converts the differential equation describing the system into a set of linear algebraic equations. The unknown model parameters can then be estimated in the frequency domain by a WLS algorithm in the presence of high measurement noises. A root mean square (RMS) normalized error criterion is applied to select the optimal mode or numerically significant spectral elements in the parameter estimation. The performance of the proposed approach is examined by means of several numerical examples through Monte Carlo simulations.