Implicit particle methods and their connection with variational data assimilation

Abstract

The implicit particle filter is a sequential Monte Carlo method for data assimilation that guides the particles to the high-probability regions via a sequence of steps that includes minimizations. We present a new and more general derivation of this approach and extend the method to particle smoothing as well as to data assimilation for perfect models. We show that the minimizations required by implicit particle methods are similar to the ones one encounters in variational data assimilation and explore the connection of implicit particle methods with variational data assimilation. In particular, we argue that existing variational codes can be converted into implicit particle methods at a low cost, often yielding better estimates, that are also equipped with quantitative measures of the uncertainty. A detailed example is presented.

References in corpus (3)

• Adaptive Importance Sampling in General Mixture Classes
• On adaptive resampling strategies for sequential Monte Carlo methods
• Sharp failure rates for the bootstrap particle filter in high dimensions

Cited by in corpus (14)

• An introduction to sampling via measure transport
• Limitations of polynomial chaos expansions in the Bayesian solution of inverse problems
• Conditions for successful data assimilation
• Parameter estimation by implicit sampling
• Model error covariance estimation in particle and ensemble Kalman filters using an online expectation-maximization algorithm
• Implicit Particle Filtering via a Bank of Nonlinear Kalman Filters
• Kernel embedding of maps for sequential Bayesian inference: The variational mapping particle filter
• Analysis of the ensemble Kalman filter for marginal and joint posteriors
• Path integral formulation of stochastic optimal control with generalized costs
• Iterative importance sampling algorithms for parameter estimation
• Nudging the particle filter
• Sampling, feasibility, and priors in Bayesian estimation
• Implicit sampling for path integral control, Monte Carlo localization, and SLAM
• Ensemble-based implicit sampling for Bayesian inverse problems with non-Gaussian priors