Results 1  10
of
102
A tutorial on particle filters for online nonlinear/nonGaussian Bayesian tracking
 IEEE TRANSACTIONS ON SIGNAL PROCESSING
, 2002
"... Increasingly, for many application areas, it is becoming important to include elements of nonlinearity and nonGaussianity in order to model accurately the underlying dynamics of a physical system. Moreover, it is typically crucial to process data online as it arrives, both from the point of view o ..."
Abstract

Cited by 1137 (2 self)
 Add to MetaCart
Increasingly, for many application areas, it is becoming important to include elements of nonlinearity and nonGaussianity in order to model accurately the underlying dynamics of a physical system. Moreover, it is typically crucial to process data online as it arrives, both from the point of view of storage costs as well as for rapid adaptation to changing signal characteristics. In this paper, we review both optimal and suboptimal Bayesian algorithms for nonlinear/nonGaussian tracking problems, with a focus on particle filters. Particle filters are sequential Monte Carlo methods based on point mass (or “particle”) representations of probability densities, which can be applied to any statespace model and which generalize the traditional Kalman filtering methods. Several variants of the particle filter such as SIR, ASIR, and RPF are introduced within a generic framework of the sequential importance sampling (SIS) algorithm. These are discussed and compared with the standard EKF through an illustrative example.
Dynamic Bayesian Networks: Representation, Inference and Learning
, 2002
"... Modelling sequential data is important in many areas of science and engineering. Hidden Markov models (HMMs) and Kalman filter models (KFMs) are popular for this because they are simple and flexible. For example, HMMs have been used for speech recognition and biosequence analysis, and KFMs have bee ..."
Abstract

Cited by 564 (3 self)
 Add to MetaCart
Modelling sequential data is important in many areas of science and engineering. Hidden Markov models (HMMs) and Kalman filter models (KFMs) are popular for this because they are simple and flexible. For example, HMMs have been used for speech recognition and biosequence analysis, and KFMs have been used for problems ranging from tracking planes and missiles to predicting the economy. However, HMMs
and KFMs are limited in their “expressive power”. Dynamic Bayesian Networks (DBNs) generalize HMMs by allowing the state space to be represented in factored form, instead of as a single discrete random variable. DBNs generalize KFMs by allowing arbitrary probability distributions, not just (unimodal) linearGaussian. In this thesis, I will discuss how to represent many different kinds of models as DBNs, how to perform exact and approximate inference in DBNs, and how to learn DBN models from sequential data.
In particular, the main novel technical contributions of this thesis are as follows: a way of representing
Hierarchical HMMs as DBNs, which enables inference to be done in O(T) time instead of O(T 3), where T is the length of the sequence; an exact smoothing algorithm that takes O(log T) space instead of O(T); a simple way of using the junction tree algorithm for online inference in DBNs; new complexity bounds on exact online inference in DBNs; a new deterministic approximate inference algorithm called factored frontier; an analysis of the relationship between the BK algorithm and loopy belief propagation; a way of
applying RaoBlackwellised particle filtering to DBNs in general, and the SLAM (simultaneous localization
and mapping) problem in particular; a way of extending the structural EM algorithm to DBNs; and a variety of different applications of DBNs. However, perhaps the main value of the thesis is its catholic presentation of the field of sequential data modelling.
RaoBlackwellised Particle Filtering for Dynamic Bayesian Networks
"... Particle filters (PFs) are powerful samplingbased inference/learning algorithms for dynamic Bayesian networks (DBNs). They allow us to treat, in a principled way, any type of probability distribution, nonlinearity and nonstationarity. They have appeared in several fields under such names as “conde ..."
Abstract

Cited by 256 (10 self)
 Add to MetaCart
Particle filters (PFs) are powerful samplingbased inference/learning algorithms for dynamic Bayesian networks (DBNs). They allow us to treat, in a principled way, any type of probability distribution, nonlinearity and nonstationarity. 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 RaoBlackwellisation. 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 RaoBlackwellised particle filters (RBPFs) lead to more accurate estimates than standard PFs. We demonstrate RBPFs on two problems, namely nonstationary 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 Þnite dimensional optimal filters.
Particle Filters for Positioning, Navigation and Tracking
, 2002
"... A framework for positioning, navigation and tracking problems using particle filters (sequential Monte Carlo methods) is developed. It consists of a class of motion models and a general nonlinear measurement equation in position. A general algorithm is presented, which is parsimonious with the part ..."
Abstract

Cited by 120 (18 self)
 Add to MetaCart
A framework for positioning, navigation and tracking problems using particle filters (sequential Monte Carlo methods) is developed. It consists of a class of motion models and a general nonlinear measurement equation in position. A general algorithm is presented, which is parsimonious with the particle dimension. It is based on marginalization, enabling a Kalman filter to estimate all position derivatives, and the particle filter becomes lowdimensional. This is of utmost importance for highperformance realtime applications. Automotive and airborne applications illustrate numerically the advantage over classical Kalman filter based algorithms. Here the use of nonlinear models and nonGaussian noise is the main explanation for the improvement in accuracy. More specifically, we describe how the technique of map matching is used to match an aircraft's elevation profile to a digital elevation map, and a car's horizontal driven path to a street map. In both cases, realtime implementations are available, and tests have shown that the accuracy in both cases is comparable to satellite navigation (as GPS), but with higher integrity. Based on simulations, we also argue how the particle filter can be used for positioning based on cellular phone measurements, for integrated navigation in aircraft, and for target tracking in aircraft and cars. Finally, the particle filter enables a promising solution to the combined task of navigation and tracking, with possible application to airborne hunting and collision avoidance systems in cars.
An MCMCbased Particle Filter For Tracking Multiple Interacting Targets
 in Proc. ECCV
, 2003
"... We describe a Markov chain Monte Carlo based particle filter that effectively deals with interacting targets, i.e., targets that are influenced by the proximity and/or behavior of other targets. Such interactions cause problems for traditional approaches to the data association problem. In respon ..."
Abstract

Cited by 112 (6 self)
 Add to MetaCart
We describe a Markov chain Monte Carlo based particle filter that effectively deals with interacting targets, i.e., targets that are influenced by the proximity and/or behavior of other targets. Such interactions cause problems for traditional approaches to the data association problem. In response, we developed a joint tracker that includes a more sophisticated motion model to maintain the identity of targets throughout an interaction, drastically reducing tracker failures. The paper presents two main contributions: (1) we show how a Markov random field (MRF) motion prior, built on the fly at each time step, can substantially improve tracking when targets interact, and (2) we show how this can be done efficiently using Markov chain Monte Carlo (MCMC) sampling. We prove that incorporating an MRF to model interactions is equivalent to adding an additional interaction factor to the importance weights in a joint particle filter. Since a joint particle filter suffers from exponential complexity in the number of tracked targets, we replace the traditional importance sampling step in the particle filter with an MCMC sampling step. The resulting filter deals efficiently and effectively with complicated interactions when targets approach each other. We present both qualitative and quantitative results to substantiate the claims made in the paper, including a large scale experiment on a videosequence of over 10,000 frames in length.
Marginalized particle filters for mixed linear/nonlinear statespace models
 IEEE Transactions on Signal Processing
, 2005
"... Abstract—The particle filter offers a general numerical tool to approximate the posterior density function for the state in nonlinear and nonGaussian filtering problems. While the particle filter is fairly easy to implement and tune, its main drawback is that it is quite computer intensive, with th ..."
Abstract

Cited by 62 (23 self)
 Add to MetaCart
Abstract—The particle filter offers a general numerical tool to approximate the posterior density function for the state in nonlinear and nonGaussian filtering problems. While the particle filter is fairly easy to implement and tune, its main drawback is that it is quite computer intensive, with the computational complexity increasing quickly with the state dimension. One remedy to this problem is to marginalize out the states appearing linearly in the dynamics. The result is that one Kalman filter is associated with each particle. The main contribution in this paper is the derivation of the details for the marginalized particle filter for a general nonlinear statespace model. Several important special cases occurring in typical signal processing applications will also be discussed. The marginalized particle filter is applied to an integrated navigation system for aircraft. It is demonstrated that the complete highdimensional system can be based on a particle filter using marginalization for all but three states. Excellent performance on real flight data is reported. Index Terms—Kalman filter, marginalization, navigation systems, nonlinear systems, particle filter, state estimation. I.
Pairwise Markov chains
 IEEE Transactions on Pattern Analysis and Machine Intelligence
, 2003
"... Abstract. The restoration of a hidden process X from an observed process Y is often performed in the framework of hidden Markov chains (HMC). HMC have been recently generalized to triplet Markov chains (TMC). In the TMC model one introduces a third random chain U and assumes that the triplet T = (X, ..."
Abstract

Cited by 48 (25 self)
 Add to MetaCart
Abstract. The restoration of a hidden process X from an observed process Y is often performed in the framework of hidden Markov chains (HMC). HMC have been recently generalized to triplet Markov chains (TMC). In the TMC model one introduces a third random chain U and assumes that the triplet T = (X, U, Y) is a Markov chain (MC). TMC generalize HMC but still enable the development of efficient Bayesian algorithms for restoring X from Y. This paper lists some recent results concerning TMC; in particular, we recall how TMC can be used to model hidden semiMarkov Chains or deal with nonstationary HMC.
Monte Carlo Filtering for MultiTarget Tracking and Data Association
 IEEE Transactions on Aerospace and Electronic Systems
, 2004
"... In this paper we present Monte Carlo methods for multitarget tracking and data association. The methods are applicable to general nonlinear and nonGaussian models for the target dynamics and measurement likelihood. We provide efficient solutions to two very pertinent problems: the data associat ..."
Abstract

Cited by 48 (2 self)
 Add to MetaCart
In this paper we present Monte Carlo methods for multitarget tracking and data association. The methods are applicable to general nonlinear and nonGaussian models for the target dynamics and measurement likelihood. We provide efficient solutions to two very pertinent problems: the data association problem that arises due to unlabelled measurements in the presence of clutter, and the curse of dimensionality that arises due to the increased size of the statespace associated with multiple targets. We develop a number of algorithms to achieve this. The first, which we will refer to as the Monte Carlo Joint Probabilistic Data Association Filter (MCJPDAF), is a generalisation of the strategy proposed in [1], [2]. As is the case for the JPDAF, the distributions of interest are the marginal filtering distributions for each of the targets, but these are approximated with particles rather than Gaussians. We also develop two extensions to the standard particle filtering methodology for tracking multiple targets. The first, which we will refer to as the Sequential Sampling Particle Filter (SSPF), samples the individual targets sequentially by utilising a factorisation of the importance weights. The second, which we will refer to as the Independent Partition Particle Filter (IPPF), assumes the associations to be independent over the individual targets, leading to an efficient componentwise sampling strategy to construct new particles. We evaluate and compare the proposed methods on a challenging synthetic tracking problem.
Mapbased multiple model tracking of a moving object
 Proceedings of eight RoboCup International Symposium
, 2004
"... Abstract. In this paper we propose an approach for tracking a moving target using RaoBlackwellised particle filters. Such filters represent posteriors over the target location by a mixture of Kalman filters, where each filter is conditioned on the discrete states of a particle filter. The discrete ..."
Abstract

Cited by 44 (1 self)
 Add to MetaCart
Abstract. In this paper we propose an approach for tracking a moving target using RaoBlackwellised particle filters. Such filters represent posteriors over the target location by a mixture of Kalman filters, where each filter is conditioned on the discrete states of a particle filter. The discrete states represent the nonlinear parts of the state estimation problem. In the context of target tracking, these are the nonlinear motion of the observing platform and the different motion models for the target. Using this representation, we show how to reason about physical interactions between the observing platform and the tracked object, as well as between the tracked object and the environment. The approach is implemented on a fourlegged AIBO robot and tested in the context of ball tracking in the RoboCup domain. 1