Bayesian Nonparametric Inference of Switching Linear Dynamical Systems

Abstract

Many complex dynamical phenomena can be effectively modeled by a system that switches among a set of conditionally linear dynamical modes. We consider two such models: the switching linear dynamical system (SLDS) and the switching vector autoregressive (VAR) process. Our Bayesian nonparametric approach utilizes a hierarchical Dirichlet process prior to learn an unknown number of persistent, smooth dynamical modes. We additionally employ automatic relevance determination to infer a sparse set of dynamic dependencies allowing us to learn SLDS with varying state dimension or switching VAR processes with varying autoregressive order. We develop a sampling algorithm that combines a truncated approximation to the Dirichlet process with efficient joint sampling of the mode and state sequences. The utility and flexibility of our model are demonstrated on synthetic data, sequences of dancing honey bees, the IBOVESPA stock index, and a maneuvering target tracking application.

50 pages, 7 figures

References in corpus (1)

• Bayesian Inference for Linear Dynamic Models with Dirichlet Process Mixtures

Cited by in corpus (41)

• Composing graphical models with neural networks for structured representations and fast inference
• Dynamical Variational Autoencoders: A Comprehensive Review
• Social Interactions for Autonomous Driving: A Review and Perspectives
• Joint modeling of multiple time series via the beta process with application to motion capture segmentation
• Extracting Traffic Primitives Directly from Naturalistically Logged Data for Self-Driving Applications
• ICON: an adaptation of infinite HMMs for time traces with drift
• Motion Blur Filtering: A Statistical Approach for Extracting Confinement Forces and Diffusivity from a Single Blurred Trajectory
• Stochastic Variational Inference
• Inferring Latent States and Refining Force Estimates via Hierarchical Dirichlet Process Modeling in Single Particle Tracking Experiments
• Modeling the Complex Dynamics and Changing Correlations of Epileptic Events
• Learning Self-Awareness for Autonomous Vehicles: Exploring Multisensory Incremental Models
• Joint Modeling of Multiple Related Time Series via the Beta Process
• Identification of jump Markov linear models using particle filters
• A Hidden Absorbing Semi-Markov Model for Informatively Censored Temporal Data: Learning and Inference
• Sparsistent Estimation of Time-Varying Discrete Markov Random Fields
• Driving Style Analysis Using Primitive Driving Patterns With Bayesian Nonparametric Approaches
• Dynamical System Segmentation for Information Measures in Motion
• Inference for multiple object tracking: A Bayesian nonparametric approach
• The Greedy Dirichlet Process Filter - An Online Clustering Multi-Target Tracker
• Policy Optimization for Markovian Jump Linear Quadratic Control: Gradient-Based Methods and Global Convergence
• Policy Learning of MDPs with Mixed Continuous/Discrete Variables: A Case Study on Model-Free Control of Markovian Jump Systems
• Learning Human Activities and Object Affordances from RGB-D Videos
• Stochastic Sequential Neural Networks with Structured Inference
• Mixed Membership Models for Time Series
• Statistical inference for dynamical systems: a review
• Global Convergence Using Policy Gradient Methods for Model-free Markovian Jump Linear Quadratic Control
• A Tempt to Unify Heterogeneous Driving Databases using Traffic Primitives
• Use of Bayesian Nonparametric methods for Estimating the Measurements in High Clutter
• A Semi-Markov Switching Linear Gaussian Model for Censored Physiological Data
• A nonparametrically corrected likelihood for Bayesian spectral analysis of multivariate time series
• Identification of Piecewise Affine State-Space Models via Expectation Maximization
• Data Driven Estimation of Stochastic Switched Linear Systems of Unknown Order
• Uncertainty Quantification of Discrete Association Problems in Image Sequence-based Tracking
• Bayesian Nonparametric Density Autoregression with Lag Selection
• Bayesian Parameter Identification for Jump Markov Linear Systems
• A Stochastic Hybrid Framework for Driver Behavior Modeling Based on Hierarchical Dirichlet Process
• Probabilistic Trajectory Segmentation by Means of Hierarchical Dirichlet Process Switching Linear Dynamical Systems
• An Adaptive Online HDP-HMM for Segmentation and Classification of Sequential Data
• Dynamic network identification from non-stationary vector autoregressive time series
• Probabilistic Segmentation via Total Variation Regularization
• Robust Particle Filtering via Bayesian Nonparametric Outlier Modeling