Information Geometric Approach to Multisensor Estimation Fusion

Distributed estimation fusion is concerned with the combination of local estimates from multiple distributed sensors to produce a fused result. In this paper, we characterize local estimates as posterior probability densities, and assume that they all belong to a parametric family. Our starting point is to consider this family as a Riemannian manifold by introducing the Fisher information metric. From the perspective of information geometry, the fused density is formulated as an informative barycenter in the space of probability densities and sought by minimizing the sum of its squared geodesic distances from the local posterior densities. Under Gaussian assumptions, a geodesic projection (GP) method and a Siegel distance (SD) method in the information-geometric framework are proposed to tackle the problem. The GP method gives a fusion result in accord with the covariance intersection estimate but under an information-geometric criterion, while the SD method appears to achieve a better approximation of the informative barycenter. Numerical examples are provided to demonstrate the performance of the proposed estimation fusion algorithms.

[1]  Ronald P. S. Mahler,et al.  Optimal/robust distributed data fusion: a unified approach , 2000, SPIE Defense + Commercial Sensing.

[2]  Shun-ichi Amari,et al.  Information geometry of the EM and em algorithms for neural networks , 1995, Neural Networks.

[3]  Stevan Stevic,et al.  Geometric Mean , 2011, International Encyclopedia of Statistical Science.

[4]  James Llinas,et al.  Multisensor Data Fusion , 1990 .

[5]  Josep M. Oller,et al.  A biplot method for multivariate normal populations with unequal covariance matrices , 2002 .

[6]  L. Skovgaard.  A Riemannian geometry of the multivariate normal model , 1984 .

[7]  J. M. Oller,et al.  AN EXPLICIT SOLUTION OF INFORMATION GEODESIC EQUATIONS FOR THE MULTIVARIATE NORMAL MODEL , 1991 .

[8]  Bernard R Gelbaum,et al.  The variational theory of geodesics , 1967 .

[9]  Hiroto Inoue.  Group Theoretical Study on Geodesics for the Elliptical Models , 2015, GSI.

[10]  Giorgio Battistelli,et al.  Kullback-Leibler average, consensus on probability densities, and distributed state estimation with guaranteed stability , 2014, Autom..

[11]  Masato Wakayama,et al.  Remarks on geodesics for multivariate normal models , 2011 .

[12]  N. Čencov.  Statistical Decision Rules and Optimal Inference , 2000 .

[13]  Chongzhao Han,et al.  Optimal Linear Estimation Fusion — Part I : Unified Fusion Rules , 2001 .

[14]  Frederic Barbaresco,et al.  Tracking quality monitoring based on information geometry and geodesic shooting , 2016, 2016 17th International Radar Symposium (IRS).

[15]  L. M. MILNE-THOMSON,et al.  An Introduction to Differential Geometry with Use of the Tensor Calculus , 1942, Nature.

[16]  C. Udriste,et al.  Geometric Modeling in Probability and Statistics , 2014 .

[17]  Y. Bar-Shalom.  On the track-to-track correlation problem , 1981 .

[18]  Aaas News,et al.  Book Reviews , 1893, Buffalo Medical and Surgical Journal.

[19]  S. Filippi,et al.  Information Geometry and Sequential Monte Carlo , 2012, 1212.0764.

[20]  Sueli I. Rodrigues Costa,et al.  Fisher information distance: a geometrical reading? , 2012, Discret. Appl. Math..

[21]  Simon J. Julier,et al.  An Empirical Study into the Use of Chernoff Information for Robust, Distributed Fusion of Gaussian Mixture Models , 2006, 2006 9th International Conference on Information Fusion.

[22]  Asuka Takatsu.  Wasserstein geometry of Gaussian measures , 2011 .

[23]  Jie Zhou,et al.  Information-geometric methods for distributed multi-sensor estimation fusion , 2016, 2016 19th International Conference on Information Fusion (FUSION).

[24]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .

[25]  M. Eilders.  Decentralized Riemannian Particle Filtering with Applications to Multi-Agent Localization , 2012 .

[26]  R. W. R. Darling,et al.  Geometrically Intrinsic Nonlinear Recursive Filters I: Algorithms , 1998 .

[27]  Jimmie D. Lawson,et al.  The Geometric Mean, Matrices, Metrics, and More , 2001, Am. Math. Mon..

[28]  F. Barbaresco,et al.  Radar detection using Siegel distance between autoregressive processes, application to HF and X-band radar , 2008, 2008 IEEE Radar Conference.

[29]  J. A. Todd,et al.  An Introduction to Differential Geometry with Use of the Tensor Calculus , 1941, The Mathematical Gazette.

[30]  Christian Germain,et al.  An M-Estimator for Robust Centroid Estimation on the Manifold of Covariance Matrices , 2016, IEEE Signal Processing Letters.

[31]  G. Battistelli,et al.  An Information-Theoretic Approach to Distributed State Estimation , 2011 .

[32]  L. Armijo.  Minimization of functions having Lipschitz continuous first partial derivatives. , 1966 .

[33]  Wojtek J. Krzanowski.  Rao's Distance Between Normal Populations That Have Common Principal Components , 1996 .

[34]  Bruno Pelletier,et al.  Informative barycentres in statistics , 2005 .

[35]  Josep M. Oller,et al.  What does intrinsic mean in statistical estimation , 2006 .

[36]  Uwe D. Hanebeck,et al.  Minimum Covariance Bounds for the Fusion under Unknown Correlations , 2015, IEEE Signal Processing Letters.

[37]  J. Burbea.  Informative Geometry of Probability Spaces , 1984 .

[38]  Marc Arnaudon,et al.  Riemannian Medians and Means With Applications to Radar Signal Processing , 2013, IEEE Journal of Selected Topics in Signal Processing.

[39]  R. Bhatia,et al.  Riemannian geometry and matrix geometric means , 2006 .

[40]  Frank Chongwoo Park,et al.  DTI Segmentation and Fiber Tracking Using Metrics on Multivariate Normal Distributions , 2013, Journal of Mathematical Imaging and Vision.

[41]  Josep M. Oller,et al.  A distance between multivariate normal distributions based in an embedding into the Siegel group , 1990 .

[42]  Maher Moakher,et al.  A Differential Geometric Approach to the Geometric Mean of Symmetric Positive-Definite Matrices , 2005, SIAM J. Matrix Anal. Appl..

[43]  Yimin Wang,et al.  Distributed estimation fusion under unknown cross-correlation: An analytic center approach , 2010, 2010 13th International Conference on Information Fusion.

[44]  J. Jost.  Riemannian geometry and geometric analysis , 1995 .

[45]  Xavier Pennec,et al.  Intrinsic Statistics on Riemannian Manifolds: Basic Tools for Geometric Measurements , 2006, Journal of Mathematical Imaging and Vision.

[46]  S. Amari,et al.  Information geometry of estimating functions in semi-parametric statistical models , 1997 .

[47]  Shun-ichi Amari,et al.  Methods of information geometry , 2000 .

[48]  Uwe D. Hanebeck,et al.  Multi-sensor distributed estimation fusion using minimum distance sum , 2014, 17th International Conference on Information Fusion (FUSION).

[49]  Shun-ichi Amari,et al.  Information geometry of Boltzmann machines , 1992, IEEE Trans. Neural Networks.

[50]  X. R. Li,et al.  A fast and fault-tolerant convex combination fusion algorithm under unknown cross-correlation , 2009, 2009 12th International Conference on Information Fusion.

[51]  Yonina C. Eldar,et al.  A Minimax Chebyshev Estimator for Bounded Error Estimation , 2008, IEEE Transactions on Signal Processing.

[52]  R. W. R. Darling,et al.  Geometrically Intrinsic Nonlinear Recursive Filters II: Foundations , 1998 .

[53]  Mark R. Morelande,et al.  Information geometry of target tracking sensor networks , 2013, Inf. Fusion.

[54]  Jeffrey K. Uhlmann,et al.  Using Exponential Mixture Models for Suboptimal Distributed Data Fusion , 2006, 2006 IEEE Nonlinear Statistical Signal Processing Workshop.

[55]  W. Marsden.  I and J , 2012 .

[56]  Kaare Brandt Petersen,et al.  The Matrix Cookbook , 2006 .

[57]  Jeffrey K. Uhlmann,et al.  A non-divergent estimation algorithm in the presence of unknown correlations , 1997, Proceedings of the 1997 American Control Conference (Cat. No.97CH36041).

[58]  Simon J. Julier,et al.  On conservative fusion of information with unknown non-Gaussian dependence , 2012, 2012 15th International Conference on Information Fusion.

[59]  Chongzhao Han,et al.  Optimal linear estimation fusion .I. Unified fusion rules , 2003, IEEE Trans. Inf. Theory.

[60]  Pramod K. Varshney,et al.  Multisensor Data Fusion , 1997, IEA/AIE.

[61]  M. Hurley.  An information theoretic justification for covariance intersection and its generalization , 2002, Proceedings of the Fifth International Conference on Information Fusion. FUSION 2002. (IEEE Cat.No.02EX5997).

[62]  Yimin Wang,et al.  Distributed Estimation Fusion with Unavailable Cross-Correlation , 2012, IEEE Transactions on Aerospace and Electronic Systems.

[63]  Yaakov Bar-Shalom,et al.  The Effect of the Common Process Noise on the Two-Sensor Fused-Track Covariance , 1986, IEEE Transactions on Aerospace and Electronic Systems.

[64]  R. Vogel.  The geometric mean? , 2020, Communications in Statistics - Theory and Methods.

[65]  Josep M. Oller,et al.  A distance between elliptical distributions based in an embedding into the Siegel group , 2002 .

[66]  Rachid Deriche,et al.  Statistics on the Manifold of Multivariate Normal Distributions: Theory and Application to Diffusion Tensor MRI Processing , 2006, Journal of Mathematical Imaging and Vision.

[67]  Rajendra Bhatia,et al.  The Riemannian Mean of Positive Matrices , 2013 .

[68]  Adrian N. Bishop.  Information fusion via the wasserstein barycenter in the space of probability measures: Direct fusion of empirical measures and Gaussian fusion with unknown correlation , 2014, 17th International Conference on Information Fusion (FUSION).

[69]  Suresh Venkatasubramanian,et al.  Robust statistics on Riemannian manifolds via the geometric median , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.