Best Linear Time-Varying Approximation of a General Class of Nonlinear Time-Varying Systems

This article presents a method for estimating linear time-varying approximation of general class nonlinear (NLTV) systems. It starts from noisy measurements the response NLTV system to special periodic excitation signals. These are subject measurement noise, process and trend. The proposed is two-step procedure. First, disturbing noise variance quantified. Next, using this knowledge, dynamics estimated together with distortions. latter split into even odd contributions. As result, signal-to-nonlinear-distortion ratio allows one decide whether or not justifiable application at hand. algorithm fully automatic in sense that user only has choose upper bounds on number basis functions used modeling signal. obtained best difference between actual predicted by uncorrelated input. Therefore, it called (BLTVA). Finally, theory validated simulation example illustrated two examples: crystallographic pitting corrosion aluminum copper electrorefining.