A new method of set membership identification for nonlinear system parameters (Q2726057)

The authors consider the estimation of the parameter vector \(\theta \in {\mathbb R}^p\) contained in the following nonlinear system NEWLINE\[NEWLINEz_i=f_i(\theta)+v_i,\quad i=1,\ldots,N,NEWLINE\]NEWLINE where \(z_i\in {\mathbb R}^m\) and \(v_i\in {\mathbb R}^m\) are the output and measurement noise, respectively, and \(f_i(\theta)\) is a nonlinear differentiable function of \(\theta\). Assuming that the noise is bounded, the set membership identification method is used to estimate the parameter \(\theta\). They present a two-steps method, that is, the center of the feasible solution set (FSS), first, is estimated and then the size of FSS is determined. In order to illustrate their algorithm and its effectiveness, they give a numerical simulation example, which describes the longitudinal dynamics of a flying vehicle.