Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks

Abstract

Particle filters (PFs) are powerful sampling-based inference/learning algorithms for dynamic Bayesian networks (DBNs). They allow us to treat, in a principled way, any type of probability distribution, nonlinearity and non-stationarity. They have appeared in several fields under such names as "condensation", "sequential Monte Carlo" and "survival of the fittest". In this paper, we show how we can exploit the structure of the DBN to increase the efficiency of particle filtering, using a technique known as Rao-Blackwellisation. Essentially, this samples some of the variables, and marginalizes out the rest exactly, using the Kalman filter, HMM filter, junction tree algorithm, or any other finite dimensional optimal filter. We show that Rao-Blackwellised particle filters (RBPFs) lead to more accurate estimates than standard PFs. We demonstrate RBPFs on two problems, namely non-stationary online regression with radial basis function networks and robot localization and map building. We also discuss other potential application areas and provide references to some finite dimensional optimal filters.

Appears in Proceedings of the Sixteenth Conference on Uncertainty in Artificial Intelligence (UAI2000)

References in corpus (3)

• Context-Specific Independence in Bayesian Networks
• Tractable Inference for Complex Stochastic Processes
• Random Algorithms for the Loop Cutset Problem

Cited by in corpus (13)

• Fighting Sample Degeneracy and Impoverishment in Particle Filters: A Review of Intelligent Approaches
• Towards Collaborative Simultaneous Localization and Mapping: a Survey of the Current Research Landscape
• Gaussian Process Autonomous Mapping and Exploration for Range Sensing Mobile Robots
• Monte Carlo Sampling Methods for Approximating Interactive POMDPs
• Particle Filtering with Invertible Particle Flow
• Intrusion Detection using Continuous Time Bayesian Networks
• Improved and Scalable Online Learning of Spatial Concepts and Language Models with Mapping
• A Whole Brain Probabilistic Generative Model: Toward Realizing Cognitive Architectures for Developmental Robots
• Tracking People in Highly Dynamic Industrial Environments
• Autonomous search for a diffusive source in an unknown environment
• Sequential Estimation of Gaussian Process-based Deep State-Space Models
• Active Exploration based on Information Gain by Particle Filter for Efficient Spatial Concept Formation
• Exploiting Causality for Selective Belief Filtering in Dynamic Bayesian Networks