Adaptive split/merge-based Gaussian mixture model approach for uncertainty propagation

This paper presents an adaptive splitting and merging scheme for dynamic selection of Gaussian kernels in a Gaussian mixture model. The Gaussian kernel in the Gaussian mixture model is split into multiple components if the Kolmogorov equation error exceeds a prescribed threshold. Two different splitting mechanisms are presented in this work. The first splitting mechanism corresponds to splitting one Gaussian kernel in all directions, whereas the second splitting mechanism corresponds to splitting in only the direction of maximum nonlinearity. The state transition matrix in conjunction with unscented transformation is used to compute the departure from linearity, and hence approximate the direction of maximum nonlinearity. The merging mechanism exploits the angle between eigenvectors corresponding to the maximum eigenvalue of covariance matrices corresponding to two different Gaussian kernels to find candidate components for merging. Finally, a sparse approximation problem is defined to provide a mechanism...

[1]  Mrinal Kumar,et al.  Numerical solution of high dimensional stationary Fokker-Planck equations via tensor decomposition and Chebyshev spectral differentiation , 2014, Comput. Math. Appl..

[2]  Puneet Singla,et al.  Computation of probabilistic hazard maps and source parameter estimation for volcanic ash transport and dispersion , 2014, J. Comput. Phys..

[3]  Y. K. Lin,et al.  On exact stationary solutions of equivalent non-linear stochastic systems , 1988 .

[4]  Jian-Qiao Sun,et al.  The closed-form solution of the reduced Fokker-Planck-Kolmogorov equation for nonlinear systems , 2016, Commun. Nonlinear Sci. Numer. Simul..

[5]  P. Green,et al.  On Bayesian Analysis of Mixtures with an Unknown Number of Components (with discussion) , 1997 .

[6]  Nando de Freitas,et al.  Sequential Monte Carlo Methods in Practice , 2001, Statistics for Engineering and Information Science.

[7]  Aubrey B. Poore,et al.  Gaussian Sum Filters for Space Surveillance: Theory and Simulations , 2011 .

[8]  James L. Beck,et al.  Approximate Solutions for Nonlinear Random Vibration Problems , 1996 .

[9]  Puneet Singla,et al.  Uncertainty Propagation in Dynamic Systems using Polynomial Chaos based Multiresolution Approach , 2012 .

[10]  Sebastian Thrun,et al.  Probabilistic robotics , 2002, CACM.

[11]  P. Spanos,et al.  Random vibration and statistical linearization , 1990 .

[12]  Puneet Singla,et al.  Conjugate Unscented Transformation-Based Approach for Accurate Conjunction Analysis , 2015 .

[13]  Suman Chakravorty,et al.  The Partition of Unity Finite Element Approach to the Stationary Fokker-Planck Equation , 2006 .

[14]  Henryk Wozniakowski,et al.  When Are Quasi-Monte Carlo Algorithms Efficient for High Dimensional Integrals? , 1998, J. Complex..

[15]  Solomon Kullback,et al.  Information Theory and Statistics , 1960 .

[16]  J. Junkins,et al.  How Nonlinear is It? A Tutorial on Nonlinearity of Orbit and Attitude Dynamics , 2003 .

[17]  G. Uhlenbeck,et al.  On the Theory of the Brownian Motion II , 1945 .

[18]  G. Muscolino,et al.  Stationary and non-stationary probability density function for non-linear oscillators , 1997 .

[19]  Herman Bruyninckx,et al.  Comment on "A new method for the nonlinear transformation of means and covariances in filters and estimators" [with authors' reply] , 2002, IEEE Trans. Autom. Control..

[20]  Fred Daum,et al.  Comments on "Finite-dimensional filters with nonlinear drift" [and addendum] , 1998 .

[21]  Kyle J. DeMars,et al.  Entropy-Based Approach for Uncertainty Propagation of Nonlinear Dynamical Systems , 2013 .

[22]  Jemin George,et al.  Spacecraft Attitude Estimation Using Adaptive Gaussian Sum Filter , 2009 .

[23]  Suman Chakravorty,et al.  The partition of unity finite element approach with hp-refinement for the stationary Fokker–Planck equation , 2009 .

[24]  Tarunraj Singh,et al.  Conjugate Unscented Transform and its Application to Filtering and Stochastic Integral Calculation , 2012 .

[25]  Puneet Singla,et al.  An Approach for Nonlinear Uncertainty Propagation: Application to Orbital Mechanics , 2009 .

[26]  Kyle J. DeMars,et al.  Probabilistic Initial Orbit Determination Using Gaussian Mixture Models , 2013 .

[27]  T. Singh,et al.  Uncertainty Propagation for Nonlinear Dynamic Systems Using Gaussian Mixture Models , 2008 .

[28]  Vivek Vittaldev,et al.  Space Object Collision Probability via Monte Carlo on the Graphics Processing Unit , 2017 .

[29]  John L. Junkins,et al.  Non-Gaussian error propagation in orbital mechanics , 1996 .

[30]  Geoffrey E. Hinton,et al.  SMEM Algorithm for Mixture Models , 1998, Neural Computation.

[31]  M. Giles,et al.  Algorithm Developments for Discrete Adjoint Methods , 2003 .

[32]  A. Jazwinski.  Stochastic Processes and Filtering Theory , 1970 .

[33]  A.R. Runnalls,et al.  A Kullback-Leibler Approach to Gaussian Mixture Reduction , 2007 .

[34]  Stephen P. Boyd,et al.  Enhancing Sparsity by Reweighted ℓ1 Minimization , 2007, 0711.1612.

[35]  A. T. Fuller,et al.  Analysis of nonlinear stochastic systems by means of the Fokker–Planck equation† , 1969 .

[36]  John L. Junkins,et al.  von Karman Lecture: Adventures on the Interface of Dynamics and Control , 1997 .

[37]  Derivative pricing with non-linear Fokker–Planck dynamics , 2003 .

[38]  Peter D. Scott,et al.  Adaptive Gaussian Sum Filter for Nonlinear Bayesian Estimation , 2011, IEEE Transactions on Automatic Control.

[39]  R. N. Iyengar,et al.  Study of the Random Vibration of Nonlinear Systems by the Gaussian Closure Technique , 1978 .

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

[41]  A. Sofi,et al.  Approximate solution of the Fokker–Planck–Kolmogorov equation , 2002 .

[42]  Puneet Singla,et al.  Nonlinear uncertainty propagation for perturbed two-body orbits , 2014 .

[43]  李昭祥,et al.  EXACT SOLUTIONS FOR STATIONARY RESPONSES OF SEVERAL CLASSES OF NONLINEAR SYSTEMS TO PARAMETRIC AND/OR EXTERNAL WHITE NOISE EXCITATIONS , 1990 .

[44]  Zhihua Zhang,et al.  EM algorithms for Gaussian mixtures with split-and-merge operation , 2003, Pattern Recognition.

[45]  F. Daum.  Bounds on performance for multiple target tracking , 1990 .

[46]  T. Singh,et al.  Polynomial-chaos-based Bayesian approach for state and parameter estimations , 2013 .

[47]  Gabriel Terejanu,et al.  An Adaptive Split-Merge Scheme for Uncertainty Propagation using Gaussian Mixture Models , 2011 .

[48]  S. Haykin,et al.  Cubature Kalman Filters , 2009, IEEE Transactions on Automatic Control.

[49]  Per Lötstedt,et al.  Fokker–Planck approximation of the master equation in molecular biology , 2009 .

[50]  H. Sorenson,et al.  Nonlinear Bayesian estimation using Gaussian sum approximations , 1972 .

[51]  Andrei N. Kolmogorov,et al.  On the Shannon theory of information transmission in the case of continuous signals , 1956, IRE Trans. Inf. Theory.

[52]  Puneet Singla,et al.  Polynomial Chaos Based Design of Robust Input Shapers , 2010 .

[53]  R. E. Kalman,et al.  A New Approach to Linear Filtering and Prediction Problems , 2002 .

[54]  Herman Bruyninckx,et al.  Kalman filters for non-linear systems: a comparison of performance , 2004 .

[55]  V. Benes.  Exact finite-dimensional filters for certain diffusions with nonlinear drift , 1981 .

[56]  F. Daum.  Nonlinear filters: beyond the Kalman filter , 2005, IEEE Aerospace and Electronic Systems Magazine.

[57]  Christian Musso,et al.  Improving Regularised Particle Filters , 2001, Sequential Monte Carlo Methods in Practice.

[58]  Kazufumi Ito,et al.  Gaussian filters for nonlinear filtering problems , 2000, IEEE Trans. Autom. Control..

[59]  Chris J. Budd,et al.  Monge-Ampére based moving mesh methods for numerical weather prediction, with applications to the Eady problem , 2013, J. Comput. Phys..