Unscented von Mises–Fisher Filtering

We introduce the unscented von Mises-Fisher filter (UvMFF), a nonlinear filtering algorithm for dynamic state estimation on the $n$-dimensional unit hypersphere. Estimation problems on the unit hypersphere occur in computer vision, e.g., when using omnidirectional cameras, as well as in signal processing. As approaches in literature are limited to very simple system and measurement models, we propose a deterministic sampling scheme on the unit hypersphere, which allows us to handle nonlinear system and measurement models. The proposed approach can be seen as a hyperspherical variant of the unscented Kalman filter (UKF). The advantages of the novel method are shown by means of simulations.

[1]  Suvrit Sra,et al.  A short note on parameter approximation for von Mises-Fisher distributions: and a fast implementation of Is(x) , 2012, Comput. Stat..

[2]  R. Fisher.  Dispersion on a sphere , 1953, Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences.

[3]  Gerhard Kurz,et al.  Unscented Orientation Estimation Based on the Bingham Distribution , 2013, IEEE Transactions on Automatic Control.

[4]  Paris Smaragdis,et al.  Multiple speaker tracking with the Factorial von Mises-Fisher Filter , 2014, 2014 IEEE International Workshop on Machine Learning for Signal Processing (MLSP).

[5]  Eric Lengyel,et al.  Mathematics for 3D Game Programming and Computer Graphics, Second Edition , 2001 .

[6]  Florent Chatelain,et al.  Von Mises-Fisher approximation of multiple scattering process on the hypersphere , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[7]  Ivan Markovic,et al.  Moving object detection, tracking and following using an omnidirectional camera on a mobile robot , 2014, 2014 IEEE International Conference on Robotics and Automation (ICRA).

[8]  Uwe D. Hanebeck,et al.  Dirac mixture approximation of multivariate Gaussian densities , 2009, Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference.

[9]  Inderjit S. Dhillon,et al.  Clustering on the Unit Hypersphere using von Mises-Fisher Distributions , 2005, J. Mach. Learn. Res..

[10]  Jeffrey K. Uhlmann,et al.  Unscented filtering and nonlinear estimation , 2004, Proceedings of the IEEE.

[11]  Gerhard Kurz,et al.  Recursive Bingham filter for directional estimation involving 180 degree symmetry , 2014 .

[12]  Gerhard Kurz,et al.  Recursive nonlinear filtering for angular data based on circular distributions , 2013, 2013 American Control Conference.

[13]  Philipp Hennig,et al.  Using an Infinite Von Mises-Fisher Mixture Model to Cluster Treatment Beam Directions in External Radiation Therapy , 2010, 2010 Ninth International Conference on Machine Learning and Applications.

[14]  Milton Abramowitz,et al.  Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables , 1964 .

[15]  Gerhard Kurz,et al.  Recursive Bayesian filtering in circular state spaces , 2015, IEEE Aerospace and Electronic Systems Magazine.

[16]  Leslie Pack Kaelbling,et al.  Tracking 3-D Rotations with the Quaternion Bingham Filter , 2013 .

[17]  Alessandro Chiuso,et al.  Visual tracking of points as estimation on the unit sphere , 1997, Block Island Workshop on Vision and Control.

[18]  Gerhard Kurz,et al.  Nonlinear measurement update for estimation of angular systems based on circular distributions , 2014, 2014 American Control Conference.

[19]  Alfred O. Hero,et al.  Parameter Estimation in Spherical Symmetry Groups , 2015, IEEE Signal Processing Letters.

[20]  Nicholas I. Fisher,et al.  Statistical Analysis of Spherical Data. , 1987 .

[21]  Gerhard Kurz,et al.  Deterministic approximation of circular densities with symmetric Dirac mixtures based on two circular moments , 2014, 17th International Conference on Information Fusion (FUSION).

[22]  M. Abramowitz,et al.  Handbook of Mathematical Functions With Formulas, Graphs and Mathematical Tables (National Bureau of Standards Applied Mathematics Series No. 55) , 1965 .

[23]  Gerhard Kurz,et al.  Non-identity measurement models for orientation estimation based on directional statistics , 2015, 2015 18th International Conference on Information Fusion (Fusion).

[24]  Gerhard Kurz,et al.  Stochastic sampling of the hyperspherical von mises–fisher distribution without rejection methods , 2015, 2015 Sensor Data Fusion: Trends, Solutions, Applications (SDF).

[25]  Ivan Markovic,et al.  Direction-only tracking of moving objects on the unit sphere via probabilistic data association , 2014, 17th International Conference on Information Fusion (FUSION).

[26]  Carl-Fredrik Westin,et al.  Hyperspherical von Mises-Fisher Mixture (HvMF) Modelling of High Angular Resolution Diffusion MRI , 2007, MICCAI.