An efficient multiple particle filter based on the variational Bayesian approach

Variational Bayesian inference for network autoregression models

Detecting and tracking fiducial points successfully can generate necessary dynamic and deformable information for facial image interpretation tasks with numerous potential applications. In this paper we propose an automatic fiducial points tracking method using multiple Differential Evolution Markov Chain (DE-MC) particle filters with kernel correlation techniques. Fiducial points are initialized through the scale invariant feature based detectors. By taking the advantage of the ability to approximate complicated proposal distributions, multiple DE-MC particle filters are applied for fiducial points tracking by building a path connecting sampling with measurements, based on the fact that the posteriori depends on both the previous state and the current observation. A Kernel correlation analysis approach is proposed to find the detection likelihood with maximization of the similarity criterion between the target points and the candidate points. Sampling efficiency is improved and computational time is substantially reduced by making use of the intermediate results obtained in particle allocation.

The particle filter has attracted considerable attention in visual tracking due to its relaxation of the linear and Gaussian restrictions in the state space model. It is thus more flexible than the Kalman filter. However, the conventional particle filter uses system transition as the proposal distribution, leading to poor sampling efficiency and poor performance in visual tracking. It is not a trivial task to design satisfactory proposal distributions for the particle filter. In this paper, we introduce an improved particle filtering framework into visual tracking, which combines the unscented Kalman filter and the auxiliary particle filter. The efficient unscented auxiliary particle filter (UAPF) uses the unscented transformation to predict one-step ahead likelihood and produces more reasonable proposal distributions, thus reducing the number of particles required and substantially improving the tracking performance. Experiments on real video sequences demonstrate that the UAPF is computationally efficient and outperforms the conventional particle filter and the auxiliary particle filter.KeywordsParticle FilterImportance SamplingObservation ModelVisual TrackingUnscented Kalman FilterThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

A novel T-S fuzzy particle filtering algorithm based on fuzzy C-regression clustering

A novel particle filter for multiple target tracking with track-before-detect measurement models is proposed. Particle filters efficiently perform target tracking under nonlinear or non-Gaussian models. However, their application to multiple target tracking suffers from the curse of dimensionality. We introduce an efficient particle filter for multiple target tracking which deals with the curse of dimensionality better than previously developed methods. The proposed algorithm is tested and compared to other multiple target tracking particle filters.

Visual tracking is one of the key tasks in computer vision. The particle filter algorithm has been extensively used to tackle this problem due to its flexibility. However the conventional particle filter uses system transition as the proposal distribution, frequently resulting in poor priors for the filtering step. The main reason is that it is difficult, if not impossible, to accurately model the target's motion. Such a proposal distribution does not take into account the current observations. It is not a trivial task to devise a satisfactory proposal distribution for the particle filter. In this paper we advance a general augmented particle filtering framework for designing the optimal proposal distribution. The essential idea is to augment a second filter's estimate into the proposal distribution design. We then show that several existing improved particle filters can be rationalised within this general framework. Based on this framework we further propose variant algorithms for robust and efficient visual tracking. Experiments indicate that the augmented particle filters are more efficient and robust than the conventional particle filter.

Identification of jump Markov autoregressive exogenous systems with missing measurements

A distributed cubature Kalman filter (CKF) is proposed for multi-sensor systems with uncertain measurement noise statistics. Firstly, the CKF is exploited to handle nonlinear state estimation. By choosing inverse Wishart priors, the joint posterior distribution of the state vector and measurement noise covariance matrix is approximated to a free-form variational distribution based on the variational Bayesian approach. By communicating with neighbors, all the sensor nodes can approximate global estimation via performing arithmetic average fusion within the framework of consensus filtering. Simulations of a target tracking problem are performed to illustrate the effectiveness of the proposed algorithm.

This paper proposes a robust Poisson multi-Bernoulli mixture (PMBM) filter with inaccurate process and measurement noise covariances. A derivation of the robust PMBM filter is provided for jointly estimating the kinematic state, the predicted state covariance, and the measurement noise covariance. By modeling the augmented state as a Gaussian inverse Wishart inverse Wishart (GIWIW) distribution, a computationally feasible implementation of the robust PMBM (GIWIW-PMBM) filter is given for linear Gaussian systems. To guarantee the conjugacy of the GIWIW distribution, the variational Bayesian (VB) approach is employed to approximate the posterior density. Finally, simulation results show that the GIWIW-PMBM filter has the best overall performance compared to existing state-of-the-art filters regarding computational cost and filtering performance.

Localization is a basic technology for intelligent vehicle (IV), which is usually carried out by fusing multiple sensors. In order to achieve robust and accurate localization results, a novel adaptive multisensor fusion method is proposed. For each sensor, every measurement is identified by an indicator, which is used to recognize whether the measurement has useful information to improve the localization performance. A robust localization model of IV is then developed by using a variational Bayesian (VB) approach. Simulations and experiments using a real IV are used to demonstrate the potential and effectiveness of the proposed method.

An unbiased finite impulse response (UFIR) filtering algorithm is designed in the discrete-time state-space for industrial processes with unknown measurement data covariance. By assuming an inverse-Wishart distribution, the data noise covariance is recursively estimated using the variational Bayesian (VB) approach. The optimal averaging horizon length N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">opt</sub> is estimated in real time by incorporating the estimated data noise covariance into the full-horizon UFIR filter and specifying N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">opt</sub> at a point, where the estimation error covariance reaches a minimum. The proposed VB-UFIR algorithm is applied to a quadrupled water tank system and moving target tracking. It is demonstrated that the VB-UFIR filter self-estimates N <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">opt</sub> more accurately than known solutions. Furthermore, the VB-UFIR filter is not prone to divergence and produces more stable and more reliable estimates than the VB-Kalman filter.

The ocean environment is complex and changeable because of all kinds of noise interferences, such as salt cliffs, ships around and other electromagnetic interferences, so the measurement information is prone to be lost. It is difficult to describe the complex noise and the acquisition probability of measurement information. In this paper, a continuous discrete variational Bayesian filter (CD VBF) is proposed to solve the problems of the heavy tailed noises and measurement random loss for state estimation. The variational Bayesian (VB) approach can effectively estimate the state vector, scale matrices, degree of freedom (DOF) parameters, Bernoulli random variables and the acquisition probability of measurement information. The performances of the proposed algorithm and traditional algorithms are tested in simulations and underwater experiments. The simulation results illustrate that the CD VBF has better localization accuracy and robustness. The experimental results demonstrate that the localization accuracy is improved by CD VBF and an optimal solution of iteration number is acquired.

In this paper, we propose a novel simultaneous registration and fusion approach for tracking. This method is based on a recursive Variational Bayesian (RVB) algorithm, which is the online variant of the Variational Bayesian (VB) approach. Under the Bayesian framework, the states and parameters are recursively estimated. It is shown by simulation that the proposed RVB method has better estimation performance than the conventional approach.

Lifetime prognostics for power module bonding wires under non-constant stress conditions: A compound stochastic process degradation modeling approach

In this article, a novel robust Gaussian approximation smoother based on expectation–maximization (EM) algorithm is proposed for cooperative localization (CL) with faulty Doppler velocity log (DVL) and heavy-tailed measurement noise. In our model, the autonomous underwater vehicle (AUV) velocity information that is not available due to DVL failure and the bias in the underwater acoustic modem are considered as unknown inputs. Then, the Student’s t distribution is used to model the heavy-tailed measurement noise. An EM algorithm is also developed for the state-space model with heavy-tailed measurement noise. The state, noise covariance matrices, and auxiliary random variables are regarded as hidden variables to obtain the maximum likelihood estimation of unknown inputs. Gaussian smoother where modified process and measurement noise covariance are inferred by variational Bayesian (VB) approach is applied to estimate the state. The experimental results illustrate that given the heavy-tailed noise, the proposed method estimates the unknown input, and the state with a high level of accuracy. The proposed algorithm can be used as a backup algorithm for CL in cases where DVL is not available due to failure.