Obstacles to High-Dimensional Particle Filtering

Snyder, Chris; Bengtsson, Thomas; Bickel, Peter; Anderson, Jeff
December 2008
Monthly Weather Review;Dec2008, Vol. 136 Issue 12, p4629
Academic Journal
Particle filters are ensemble-based assimilation schemes that, unlike the ensemble Kalman filter, employ a fully nonlinear and non-Gaussian analysis step to compute the probability distribution function (pdf) of a system’s state conditioned on a set of observations. Evidence is provided that the ensemble size required for a successful particle filter scales exponentially with the problem size. For the simple example in which each component of the state vector is independent, Gaussian, and of unit variance and the observations are of each state component separately with independent, Gaussian errors, simulations indicate that the required ensemble size scales exponentially with the state dimension. In this example, the particle filter requires at least 1011 members when applied to a 200-dimensional state. Asymptotic results, following the work of Bengtsson, Bickel, and collaborators, are provided for two cases: one in which each prior state component is independent and identically distributed, and one in which both the prior pdf and the observation errors are Gaussian. The asymptotic theory reveals that, in both cases, the required ensemble size scales exponentially with the variance of the observation log likelihood rather than with the state dimension per se.


Related Articles

  • Simulations of two types of El Niño events by an optimal forcing vector approach. Duan, Wansuo; Tian, Ben; Xu, Hui // Climate Dynamics;Sep2014, Vol. 43 Issue 5/6, p1677 

    In this paper, an optimal forcing vector (OFV) approach is proposed. The OFV offsets tendency errors and optimizes the agreement of the model simulation with observation. We apply the OFV approach to the well-known Zebiak-Cane model and simulate several observed eastern Pacific (EP) El Niño...

  • Potential of an ensemble Kalman smoother for stratospheric chemical-dynamical data assimilation. Milewski, Thomas; Bourqui, MichelS. // Tellus: Series A;2013, Vol. 65, p1 

    A new stratospheric ensemble Kalman smoother (EnKS) system is introduced, and the potential of assimilating posterior stratospheric observations to better constrain the whole model state at analysis time is investigated. A set of idealised perfect-model Observation System Simulation Experiments...

  • Controlling Overestimation of Error Covariance in Ensemble Kalman Filters with Sparse Observations: A Variance-Limiting Kalman Filter. Gottwald, Georg A.; Mitchell, Lewis; Reich, Sebastian // Monthly Weather Review;Aug2011, Vol. 139 Issue 8, p2650 

    The problem of an ensemble Kalman filter when only partial observations are available is considered. In particular, the situation is investigated where the observational space consists of variables that are directly observable with known observational error, and of variables of which only their...

  • Observation impact in data assimilation: the effect of non-Gaussian observation error. Fowler, Alison; Jan Van Leeuwen, Peter // Tellus: Series A;2013, Vol. 65, p1 

    Data assimilation methods which avoid the assumption of Gaussian error statistics are being developed for geoscience applications. We investigate how the relaxation of the Gaussian assumption affects the impact observations have within the assimilation process. The effect of non-Gaussian...

  • Finite-Horizon Robust Kalman Filter for Uncertain Attitude Estimation System with Star Sensor Measurement Delays. Hua-Ming Qian; Wei Huang; Biao Liu // Abstract & Applied Analysis;2014, p1 

    This paper addresses the robust Kalman filtering problem for uncertain attitude estimation system with star sensor measurement delays. Combined with the misalignment errors and scale factor errors of gyros in the process model and the misalignment errors of star sensors in the measurement model,...

  • A Hybrid ETKF–3DVAR Data Assimilation Scheme for the WRF Model. Part II: Real Observation Experiments. Xuguang Wang; Barker, Dale M.; Snyder, Chris; Hamill, Thomas M. // Monthly Weather Review;Dec2008, Vol. 136 Issue 12, p5132 

    The hybrid ensemble transform Kalman filter–three-dimensional variational data assimilation (ETKF–3DVAR) system developed for the Weather Research and Forecasting (WRF) Model was further tested with real observations, as a follow-up for the observation system simulation experiment...

  • Alignment Error Models and Ensemble-Based Data Assimilation. Lawson, W. Gregory; Hansen, James A. // Monthly Weather Review;Jun2005, Vol. 133 Issue 6, p1687 

    The concept of alternative error models is suggested as a means to redefine estimation problems with non-Gaussian additive errors so that familiar and near-optimal Gaussian-based methods may still be applied successfully. The specific example of a mixed error model including both alignment...

  • Ensemble Kalman filter for model updating - A special issue. Nævdal, Geir; Hanea, Remus; Oliver, Dean; Vallès, Brice // Computational Geosciences;Mar2011, Vol. 15 Issue 2, p223 

    The article discusses various reports published within the issue, including one by P. Sakov and L. Bertino on the relation between two common localization methods for the ensemble Kalman filter (EnKF), another by J. D. Kepert on balance-aware covariance localization for atmospheric and oceanic...

  • A Multi Swarm Particle Filter for Mobile Robot Localization. Havangi, Ramazan; Nekoui, Mohammad Ali; Teshnehlab, Mohammad // International Journal of Computer Science Issues (IJCSI);May2010, Vol. 7 Issue 3, p15 

    Particle filter (PF) is widely used in mobile robot localization, since it is suitable for the nonlinear non- Gaussian system. Localization based on PF, However, degenerates over time. This degeneracy is due to the fact that a particle set estimating the pose of the robot looses its diversity....


Read the Article


Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics