Convergence of a recursive rational interpolation-based identification algorithm

This paper investigates the convergence properties of an identification algorithm based on recursive rational interpolation. The algorithm utilizes the moving discrete Fourier transform (MDFT), which is a recursive form of the DFT, to efficiently monitor certain points in the spectra of the system input and output signals. Convergence of parameter estimates to their true values is established for the algorithm, and persistent excitation conditions are also given.<
>