Normalizing Flow-Based Differentiable Particle Filters

作     者：Chen, Xiongjie Li, Yunpeng 

作者机构：Kings Coll London Fac Dent Oral & Craniofacial Sci Ctr Oral Clin & Translat Sci London SE1 9RT England 

出 版 物：《IEEE TRANSACTIONS ON SIGNAL PROCESSING》 (IEEE Trans Signal Process)

年 卷 期：2025年第73卷

页      面：493-507页

核心收录：

学科分类：0808[工学-电气工程] 08[工学] 

主　　题：Particle filters Neural networks State-space methods Particle measurements Parameter estimation Monte Carlo methods Atmospheric measurements Proposals Maximum likelihood estimation Bayes methods Sequential Monte Carlo differentiable particle filters normalizing flows parameter estimation machine learning 

摘      要：Recently, there has been a surge of interest in incorporating neural networks into particle filters, e.g. differentiable particle filters, to perform joint sequential state estimation and model learning for nonlinear non-Gaussian state-space models in complex environments. Existing differentiable particle filters are mostly constructed with vanilla neural networks that do not allow density estimation. As a result, they are either restricted to a bootstrap particle filtering framework or employ predefined distribution families (e.g. Gaussian distributions), limiting their performance in more complex real-world scenarios. In this paper we present a differentiable particle filtering framework that uses (conditional) normalizing flows to build its dynamic model, proposal distribution, and measurement model. This not only enables valid probability densities but also allows the proposed method to adaptively learn these modules in a flexible way, without being restricted to predefined distribution families. We derive the theoretical properties of the proposed filters and evaluate the proposed normalizing flow-based differentiable particle filters performance through a series of numerical experiments.