System analysis and synthesis via orthogonal polynomial series and Fourier series (Q1070976): Difference between revisions

This paper provides a review of some recent results on solving systems analysis and synthesis problems via orthogonal polynomial series and Fourier series. The results are unified in the sense that a unique approach is followed namely that of reducing the problems at hand (viz. analysis, identification, sensitivity, optimal control) to the problem of solving a system of algebraic equations. The key point of this unified approach is the so-called integral property of the basis vector \(\phi\) (t), i.e. \[ \int^{t}_{\alpha}...\int^{t}_{\alpha}\phi (\sigma)(d\sigma)^ k\simeq P^ k\phi (t),\quad k\text{-times integrated,} \] where P is a square constant matrix called the operational matrix of integration, and \(\phi(t) = [\phi_ 0(t), \ldots, \phi_{n-1}(t)]^ T\) with \(\phi_ 0(t),...,\phi_{n-1}(t)\) being the basis functions which are orthogonal (or orthonormal) on a certain interval [\(\alpha\),\(\beta\) ]. The basis function used so far can be grouped into three categories, namely, piecewise constant functions (Walsh and block-pulse functions), polynomial functions (Chebyshev, Legendre, Laguerre, Hermite, etc.) and sine-cosine functions.