liaoyw2016/gp-adf

Gaussian Process Assumed Density Filter (Code for ICML 2009 paper)

gp-adf

% These matlab functions contain the core of the GP-ADF-algorithm as described in % Deisenroth, Huber, Hanebeck: Analytic Moment-based Gaussian Process Filtering, % International Conference on Machine Learning (ICML), 2009. % % functions included: % % eps2pdf.m: plot/print utility % gpf.m: the function that does the high-level filtering (one time step) % gpPt.m: GP predictions with uncertain inputs (transition dynamics) % gpPo.m: GP predicitons with uncertain inputs (observation function) % gpukf.m: GP-UKF implementation % maha.m: computes the pairwise squared Mahalanobis distance between to % sets of vectors % scaledSymmetricSigmaPoints.m: compute sigma points for UKF % sim_scalar.m scalar toy example to compare filters % trainf.m: wrapper to train multiple-target GPs % ukf_add.m: UKF for additive Gaussian noise % % % You should be able to run the GP-ADF % % % example call: % sim_scalar % % % (C) Copyright 2009-2016, Marc Peter Deisenroth % % http://wp.doc.ic.ac.uk/sml % 2016-07-19