Title :
Fourier-Hermite Kalman Filter
Author :
Sarmavuori, Juha ; Särkkä, Simo
Author_Institution :
Nokia Siemens Networks, Espoo, Finland
fDate :
6/1/2012 12:00:00 AM
Abstract :
In this note, we shall present a new class of Gaussian filters called Fourier-Hermite Kalman filters. Fourier-Hermite Kalman filters are based on expansion of nonlinear functions with the Fourier-Hermite series in same way as the traditional extended Kalman filter is based on the Taylor series. The first order truncation of the Fourier-Hermite series gives the previously known statistically linearized filter.
Keywords :
Fourier series; Gaussian processes; Kalman filters; nonlinear filters; nonlinear functions; Fourier-Hermite Kalman filter; Fourier-Hermite series; Gaussian filters; Taylor series; extended Kalman filter; nonlinear functions; statistically linearized filter; Approximation methods; Kalman filters; Numerical models; Polynomials; Stochastic processes; Taylor series; Vectors; Extended Kalman filtering (EKF); Fourier-Hermite series; nonlinear KF; statistical linearization;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.2011.2174667
