Inference with the Unscented Kalman Filter and optimization of sigma points

Michela Eugenia Pasetto; Umberto NoÃ¨; Alessandra Luati; Dirk Husmeier

Open Conference Systems, STATISTICS AND DATA SCIENCE: NEW CHALLENGES, NEW GENERATIONS

Michela Eugenia Pasetto, Umberto NoÃ¨, Alessandra Luati, Dirk Husmeier

Last modified: 2017-05-20

Abstract

We investigate the accuracy of inference in a chaotic dynamical system (Duffing oscillator) with the Unscented Kalman Filter and quantify the dependence on the sample size and the signal to noise ratio.In order to improve convergence to the true parameters in the case of a bad initialization of the algorithm, we optimize the location of sigma points with Bayesian optimisation.