Algorithmic Techniques for Modeling Nonlinear Functions
Algorithmic Techniques are derived and applied to the problem of modeling nonlinear functions. Two general approaches are considered. The first uses zero order quantizers with adaptable parameters to approximate the given nonlinearity. The second makes use of algorithmic techniques to establish successively better estimates of the coefficients of a Chebyshev polynomial representation of a nonlinear function, with particular emphasis on a hysteresis nonlinearity. Algorithms are derived, convergence considered and results given for each case.
KeywordsConvergence Algorithm Algorithmic Technique Hysteresis Nonlinearity Initial Gain Continuous Monotonic Function
Unable to display preview. Download preview PDF.
- 1.“Learning the Shape of a Class of pdf’s” Lester A. Gerhardt and Kenneth W. Drake, IEEE Transactions on SMC, July 1972.Google Scholar
- 2.“Nonlinear Problems in Random Theory,” N. Wiener, The Technology Press of M.I.T., Cambridge, Mass., and John Wiley and Sons, New York, 1949.Google Scholar
- 3.“System Identification and Pattern Recognition,” Rob Roy and James Sherman, IFAC, Prague, 1967.Google Scholar
- 4.“Harmonic Analysis for Nonlinear Characteristics,” Laurel J. Lewis, Transaction of the AIEE, Vol. 74, 1955, pp. 693–700.Google Scholar
- 5.“Nonlinear Network and Systems,” Richard Clay, Wiley- Interscience, 1971, pp. 126–129.Google Scholar