Search

With growing security threats, many online and offine frameworks have been proposed for anomaly detection in video sequences. However, existing online anomaly detection techniques are either computationally very expensive or lack desirable accuracy. This research work proposes a novel particle filtering based framework for online anomaly detection which detects video frames with anomalous activities based upon the posterior probability of activities in a video sequence. The proposed method also detects anomalous regions in anomalous video frames. We propose novel prediction and measurement models to accurately detect anomalous video frames and anomalous regions in video frames. Novel prediction model for particle prediction and likelihood model for assigning weights to these particles are proposed. These models efficiently utilise variable sized cell structure which creates variable sized sub-regions of scenes in video frames. Furthermore, they efficiently extract and utilise information from the video frame in the form of size, motion and location features. The proposed framework is tested on UCSD and LIVE datasets and compared with the existing state-of-the-art algorithms in the literature. The proposed anomaly detection algorithm outperforms the state-of-the art algorithms in terms of reduced Equal Error Rate (EER) with comparatively lesser processing time.

Human hands engage in interactive activities in many practical working scenarios, among which the interactions between human hands and objects are the most common. Tracking the movement of the human hand during hand-object interactions is an important research task that is also challenging due to the high-dimensionality and occlusions. In this paper, we track hand-object interactions from depth observations with a model-based method. To overcome the difficulties of optimum searching in the hand-object high-dimensional space, we propose a new algorithm - collaborative differential evolution filtering (CoDEF) - for tracking hand-object interactions. The proposed CoDEF algorithm integrates the differential evolution (DE) algorithm into a particle filtering (PF) framework to accelerate the convergence of particles. Particles are driven to the regions with a high probability by optimizing the matching error under the current observation with DE. To decompose the state space and decrease the complexity of optimum searching, CoDEF tracks the movement of the hand and object by using two collaborative trackers. Based on the proposed CoDEF algorithm, we develop a model-based tracking system with 3D graphic techniques. According to the experimental results, the proposed CoDEF algorithm can achieve robust tracking of hand-object interactions using fewer particles.

EURASIP journal on advances in signal processing 2019(1), 52 (2019). doi:10.1186/s13634-019-0643-3, Published by Springer, Heidelberg

Anomaly detection constitutes a fundamental step in developing self-aware autonomous agents capable of continuously learning from new situations, as it enables to distinguish novel experiences from already encountered ones. This paper combines Dynamic Bayesian Networks (DBNs) and Neural Networks (NNs) and proposes a method for detecting anomalies in video data at different abstraction levels. We use a Variational Autoencoder (VAE) to reduce the dimensionality of video frames, and Optical Flows between subsequent images, generating a latent space that captures both visual and dynamical information and that is comparable to low-dimensional sensory data (e.g., positioning, steering angle). An Adapted Markov Jump Particle Filter is employed to predict the following frames and detect anomalies in video data. Our method's evaluation is executed using different video data from a semi-autonomous vehicle performing different tasks in a closed environment. Tests on benchmark anomaly detection datasets have additionally been conducted.

Advanced control and driving assistance systems play a major role in modern vehicles, ensuring higher standards of safety and performance. Their correct operation extensively depends on the knowled...

One of the most significant research trends in the last decades of the aeronautic industry is the effort to move towards the design and the production of “more electric aircraft”. Within this framework, the application of the electrical technology to flight control systems has seen a progressive, although slow, increase: starting with the introduction of fly-by-wire and proceeding with the partial replacement of the traditional hydraulic/electro-hydraulic actuators with purely electro-mechanical ones. This evolution allowed to obtain more flexible solutions, reduced installation issues and enhanced aircraft control capability. Electro-Mechanical Actuators (EMAs) are however far from being a mature technology and still suffer from several safety issues, which can be partially limited by increasing the complexity of their design and hence their production costs. The development of a robust Prognostics and Health Management (PHM) system could provide a way to prevent the occurrence of a critical failure without resorting to complex device design. This paper deals with the first part of the study of a comprehensive PHM system for EMAs employed as primary flight control actuators; the peculiarities of the application are presented and discussed, while a novel approach, based on short pre-flight/post-flight health monitoring tests, is proposed. Turn-to-turn short in the electric motor windings is identified as the most common electrical degradation and a particle filtering framework for anomaly detection and prognosis featuring a self-tuning non-linear model is proposed. Features, anomaly detection and a prognostic algorithm are hence evaluated through state-of-the art performance metrics and their results discussed.

Map-matching is a popular method that uses spatial information to improve the accuracy of positioning methods. The performance of map matching methods is closely related to spatial characteristics. Although several studies have demonstrated that certain map matching algorithms are affected by some spatial structures (e.g., parallel paths), they focus on the analysis of single map matching method or few spatial structures. In this study, we explored how the most commonly-used four spatial characteristics (namely forks, open spaces, corners, and narrow corridors) affect three popular map matching methods, namely particle filtering (PF), hidden Markov model (HMM), and geometric methods. We first provide a theoretical analysis on how spatial characteristics affect the performance of map matching methods, and then evaluate these effects through experiments. We found that corners and narrow corridors are helpful in improving the positioning accuracy, while forks and open spaces often lead to a larger positioning error. We hope that our findings are helpful for future researchers in choosing proper map matching algorithms with considering the spatial characteristics.

For target tracking in underwater wireless sensor networks (WSNs), the contributions of the measured values of each sensor node are different for data fusion, so a better weighted nodes fusion and participation planning mechanism can obtain better tracking performance. A distributed particle filter based target tracking algorithm with Grubbs criterion and mutual information entropy weighted fusion (GMIEW) is proposed in this paper. The Grubbs criterion is adopted to analyze and verify the information obtained by sensor nodes before the information fusion, and accordingly some interference information or error information can be excluded from the data set. In the process of calculating importance weight in particle filter, dynamic weighting factor is introduced. The mutual information entropy between the measured value of the sensor nodes and the target state is used to reflect the amount of target information provided by sensor nodes, thus a dynamic weighting factor corresponding to each node can be obtained. The simulation results show that the proposed algorithm effectively improves the accuracy of prediction of target tracking system.

Microscopic images of neuronal cells provide essential structural information about the key constituents of the brain and form the basis of many neuroscientific studies. Computational analyses of the morphological properties of the captured neurons require first converting the structural information into digital tree-like reconstructions. Many dedicated computational methods and corresponding software tools have been and are continuously being developed with the aim to automate this step while achieving human-comparable reconstruction accuracy. This pursuit is hampered by the immense diversity and intricacy of neuronal morphologies as well as the often low quality and ambiguity of the images. Here we present a novel method we developed in an effort to improve the robustness of digital reconstruction against these complicating factors. The method is based on probabilistic filtering by sequential Monte Carlo estimation and uses prediction and update models designed specifically for tracing neuronal branches in microscopic image stacks. Moreover, it uses multiple probabilistic traces to arrive at a more robust, ensemble reconstruction. The proposed method was evaluated on fluorescence microscopy image stacks of single neurons and dense neuronal networks with expert manual annotations serving as the gold standard, as well as on synthetic images with known ground truth. The results indicate that our method performs well under varying experimental conditions and compares favorably to state-of-the-art alternative methods. Electronic supplementary material The online version of this article (10.1007/s12021-018-9407-8) contains supplementary material, which is available to authorized users.

This article develops a sequential Bayesian learning method to estimate the parameters and recover the state variables for generalized autoregressive conditional heteroscedasticity (GARCH) models, which are commonly used in the financial time-series analysis. This simulation-based method combines particle-filtering technology with a Markov chain Monte Carlo algorithm when the model is non-linear and the number of observed variables is relatively sparse. We compare the performance of the sequential Bayesian learning approach with the numerical maximum likelihood estimation (NMLE) in estimating models based on S&P 500 return rates. Our research concludes that the sequential parameter learning approach performs more robustly and accurately than the NMLE, by taking into account the uncertainty of the model. We also carry out simulation studies to confirm that the sequential Bayesian learning method is extremely reliable for GARCH models.

International audience; We propose an analytic, time-variant model that conservatively evaluates the increase in reliability achievable when a component is equipped with a Prognostics and Health Management system of known performance metrics. The reliability model builds on metrics of literature and is applicable to different industrial contexts. A simulated case study concerning crack propagation in a mechanical component is considered to validate the proposed model.

In this paper, we propose a particle filtering-based method for predicting the remaining storage life (RSL) of electromagnetic relays. The RSL prediction problem here addressed has the following three distinctive features: i) limited measurement data available; ii) incomplete run-to-failure data; and iii) no model available for the physical degradation process. Then, to develop the method for RSL prediction, storage testing and degradation mechanism analysis have been carried out to obtain the knowledge and information needed to develop the physical model that supports the RSL prediction procedure. We discuss the three main steps of the proposed prediction method: parameter estimation, model validation and RSL prediction. Data from nine relays are used for estimating the initial parameter values distribution and data from one relay are used for RSL prediction. The RSL prediction results are compared with those obtained by a nonlinear curve-fitting method and a basic particle filtering algorithm. The comparison shows that the proposed method is more effective in predicting the RSL than the other methods.

A deterioration model plays an important role to predict the valid total maintenance cost for sustainable maintenance of bridges. In the current state-of-the-art, the deterioration model has regression parameters as a probabilistic process by an initially determined mean and standard deviation, called an existing model. However, the existing model has difficulty to predict maintenance costs accurately, because it cannot reflect an information based on structural damage at an operational stage. In this research, updating the probabilistic deterioration model is presented for the prediction of pre-stressed concrete I-type (PSCI) girder bridges using a particle filtering technique which is an advanced Bayesian updating method based on big data analysis. The method enables predicting maintenance cost fitted in the current structural status, which includes the recent information by inspection with bridge-monitoring. The method is adapted in the Mokdo Bridge which is currently being used for evaluating the efficiency of maintenance cost by effects on updated probabilistic values with two different scenarios. As the result, it is shown that the proposed method is effective in predicting maintenance costs.

Sensors 21(6), 2114 (2021). doi:10.3390/s21062114 special issue: "Special Issue "Industrial Applications of Smart Sensors and Smart Data in Cyber-Physical Systems" / Special Issue Editors: Prof. Dr. Oliver Niggemann, Guest Editor ; Prof. Dr. Michael Heizmann, Guest Editor", Published by MDPI, Basel

There is a rising interest in studying the degree of connection and the causal relationships between brain regions, as a growing body of evidence suggests that features of these interactions could play a role as markers in a host of neurological diseases. The vast majority of brain connectivity studies treats the brain network as stationary. New insights on the temporal behaviour of these connections could significantly improve our understanding of brain networking in both physiology and pathology. In this paper, we propose the application of a computational methodology, named Particle Filter (PF), to functional Magnetic Resonance Imaging (fMRI) data. The PF algorithm aims to estimate time-varying hidden variables of a given observational model through a Sequential Monte Carlo approach. The fMRI data are represented as a first-order linear time-varying Vector Autoregression model (VAR). On simulated time series, the PF approach effectively detected and enabled to follow time-varying hidden parameters and it captured causal relationships among signals. The method was also applied to real fMRI data and provided similar results to those obtained by using a different proxy measure of causal dependency, that is, correlation between delayed time series. Interestingly, the PF approach also enabled to detect statistically significant changes in the cause-effect relationships between areas, which correlated with the underlying stimulation pattern delivered to subjects during the fMRI acquisition.

International audience; Detect and tracking of maneuvering target is a complicated dynamic state estimation problem whose difficulty is increased in case of high noise environments or low signal-to-noise ratio (SNR). In this case, the track-before-detect filter (TBDF) that uses unthresholded measurements considers as an effective method for detecting and tracking a single target under low SNR conditions. Nevertheless, the performance of the algorithm will be affected with severe loss because of the mismatching of target model during maneuver. In this paper, to resolve the target maneuvers, we propose an application of particle filtering which depends on track before detect (PF - TBD) algorithm in order to track the maneuvering target. We employ the Constant Acceleration (CA) model and Coordinate Turn model (CT). Our simulation results show that the detection and tracking of maneuvering target performance of TBD-PF has been improved using the proposed algorithm.

In this paper, we propose a particle-filter-based superpixel (PFS) segmentation method that extends the original tracking problem as a region clustering problem. The basic idea is to approximate superpixel centers by multiple particles to obtain high intra-region similarity. Specifically, we firstly use a density cluster to initialize single-group particles and introduce the association rule for mining other initial candidate particles. In propagation, particles are transferred to neighboring local regions by a moving step aiming to update local candidate particles with a lower energy cost. We evaluate all particles on the basis of their cluster similarity and estimate the largest particles as the final superpixel centers. The proposed method can locate cluster centers in diverse feature space, which alleviates the risk of a local optimum and produces better segmentation performance. Experimental results on the Berkeley segmentation 500 dataset (BSD500) demonstrate that our method outperforms seven state-of-the-art methods.

The conventional direction of arrival (DOA) estimation algorithm with static sources assumption usually estimates the source angles of two adjacent moments independently and the correlation of the moments is not considered. In this article, we focus on the DOA estimation of moving sources and a modified particle filtering (MPF) algorithm is proposed with state space model of single acoustic vector sensor. Although the particle filtering (PF) algorithm has been introduced for acoustic vector sensor applications, it is not suitable for the case that one dimension angle of source is estimated with large deviation, the two dimension angles (pitch angle and azimuth angle) cannot be simultaneously employed to update the state through resampling processing of PF algorithm. To solve the problems mentioned above, the MPF algorithm is proposed in which the state estimation of previous moment is introduced to the particle sampling of present moment to improve the importance function. Moreover, the independent relationship of pitch angle and azimuth angle is considered and the two dimension angles are sampled and evaluated, respectively. Then, the MUSIC spectrum function is used as the “likehood” function of the MPF algorithm, and the modified PF-MUSIC (MPF-MUSIC) algorithm is proposed to improve the root mean square error (RMSE) and the probability of convergence. The theoretical analysis and the simulation results validate the effectiveness and feasibility of the two proposed algorithms.

In this paper, a tracking method based on sequential Bayesian inference is proposed. The proposed method focuses on solving both the problem of tracking under partial occlusions and the problem of non-rigid object tracking in real-time on a desktop personal computer (PC). The proposed method is mainly composed of two parts: (1) modeling the target object using elastic structure of local patches for robust performance; and (2) efficient hierarchical diffusion method to perform the tracking procedure in real-time. The elastic structure of local patches allows the proposed method to handle partial occlusions and non-rigid deformations through the relationship among neighboring patches. The proposed hierarchical diffusion method generates samples from the region where the posterior is concentrated to reduce computation time. The method is extensively tested on a number of challenging image sequences with occlusion and non-rigid deformation. The experimental results show the real-time capability and the robustness of the proposed method under various situations. Display Omitted We propose a tracking method to solve both the problem of partial occlusions and non-rigid deformations in real-time.The target object is modeled through an elastic structure of local patches for robust performance.Hierarchical diffusion method is proposed to obtain an acceptable solution in real time.Extensive evaluation shows that the proposed method outperforms state-of-the-art.

International audience; Particle Filtering (PF) is a model-based, filtering technique, which has drawn the attention of the Prognostic and Health Management (PHM) community due to its applicability to nonlinear models with non-additive and non-Gaussian noise. When multiple physical models can describe the evolution of the degradation of a component, the PF approach can be based on Multiple Swarms (MS) of particles, each one evolving according to a different model, from which to select the most accurate a posteriori distribution. However, MS are highly computational demanding due to the large number of particles to simulate. In this work, to tackle the problem we have developed a PF approach based on the introduction of an augmented discrete state identifying the physical model describing the component evolution, which allows to detect the occurrence of abnormal conditions and identifying the degradation mechanism causing it. A crack growth degradation problem has been considered to prove the effectiveness of the proposed method in the detection of the crack initiation and the identification of the occurring degradation mechanism. The comparison of the obtained results with that of a literature MS method and of an empirical statistical test has shown that the proposed method provides both an early detection of the crack initiation, and an accurate and early identification of the degradation mechanism. A reduction of the computational cost is also achieved. 2

Algorithms for approximate Bayesian inference, such as those based on sampling (i.e., Monte Carlo methods), provide a natural source of models of how people may deal with uncertainty with limited cognitive resources. Here, we consider the idea that individual differences in working memory capacity (WMC) may be usefully modeled in terms of the number of samples, or “particles,” available to perform inference. To test this idea, we focus on two recent experiments that report positive associations between WMC and two distinct aspects of categorization performance: the ability to learn novel categories, and the ability to switch between different categorization strategies (“knowledge restructuring”). In favor of the idea of modeling WMC as a number of particles, we show that a single model can reproduce both experimental results by varying the number of particles—increasing the number of particles leads to both faster category learning and improved strategy‐switching. Furthermore, when we fit the model to individual participants, we found a positive association between WMC and best‐fit number of particles for strategy switching. However, no association between WMC and best‐fit number of particles was found for category learning. These results are discussed in the context of the general challenge of disentangling the contributions of different potential sources of behavioral variability.\ud \ud

For nuclear power plants (NPPs) to have long lifetimes, ageing is a major issue. Currently, ageing management for NPP systems is based on correlations built from generic experimental data. However, each system has its own characteristics, operational history, and environment. To account for this, it is possible to resort to prognostics that predicts the future state and time to failure (TTF) of the target system by updating the generic correlation with specific information of the target system. In this paper, we present an application of particle filtering for the prediction of degradation in steam generator tubes. With a case study, we also show how the prediction results vary depending on the uncertainty of the measurement data. Keywords: Prognostics, Particle filtering, Model-based method, Steam generator tube rupture, Nuclear power plant

The estimation of the 3D trajectory, the class and the dimensions of a vehicle represents three relevant tasks for traffic monitoring. They are usually performed by separate sub-systems and only few existing algorithms cope with the three tasks at the same time. However, if these tasks are integrated, the trajectory estimation enforces the classification with temporal consistency, and at the same time, the estimation of the vehicle class and dimensions can be used to increase the trajectory estimate accuracy. In this work, we propose an algorithm to estimate the 3D trajectory, the class and the dimensions of vehicles simultaneously by means of a Backward-Simulation Particle Smoother whose state contains both continuous (vehicle pose and dimensions), and discrete (vehicle class) quantities. To integrate the class estimate in the Particle Smoother we model the class prediction as a Markov Chain. We performed experimental tests on both simulated and real datasets; they show that the pose and dimension estimation reaches centimeter-accuracy and the classification accuracy is higher than 95 %.

This paper presents an autonomous docking system with novel integrated algorithms for mobile self-reconfigurable robots equipped with inexpensive sensors. A novel docking algorithm was developed to determine the initial distance and orientation of the two modules, and sensor models were established through experiments. Both Extended Kalman filter (EKF) and particle filter (PF) were deployed to fuse the measurements from IR and encoders and provide accurate estimates of orientation and distance. Simulation experiments were carried out first and then real experiments were conducted to verify the feasibility and good performance of the proposed docking algorithm and system. The proposed system provides a robust and reliable docking solution using low cost sensors.

The deployment of networks of sensors and development of pertinent information processing techniques can facilitate the requirement of situational awareness present in many defense/surveillance systems. Sensors allow the collection and distributed processing of information in a variety of environments whose structure is not known and is dynamically changing with time. A distributed dynamic data driven (DDDAS-based) framework is developed in this paper to address distributed multi-threat tracking under limited sensor resources. The acquired sensor data will be used to control the sensing part of the sensor network, and utilize only the sensing devices that acquire good quality measurements about the present targets. The DDDAS-based concept will be utilized to enable efficient sensor activation of only those parts of the network located close to a target/object. A novel combination of stochastic filtering techniques, drift homotopy and sparsity-inducing canonical correlation analysis (S-CCA) is utilized to dynamically identify the target-informative sensors and utilize them to perform improved drift-based particle filtering techniques that will allow robust, stable and accurate distributed tracking of multiple objects. Numerical tests demonstrate the effectiveness of the novel framework.

Collision avoidance is one of the high-level safety objectives and requires a complete and reliable description of maritime traffic situation. A combined use of data provided by independent data sources is an approach to improve the accuracy and integrity of traffic situation related information. In this paper the authors study the usage of radar images for automatic identification system (AIS) and radar fusion. Therefore they simulate synthetic radar images and evaluate the tracking performance of the particle filter algorithm as the most promising filter approach. During the filter process the algorithm estimates the target position and velocity which they finally compare with the known position of the simulation. This approach allows the performance analysis of the particle filter for vessel tracking on radar images. In a second extended simulation they add the respective AIS information of the target vessel and study the gained level of improvement for the particle filter approach. The work of this paper is integrated in the research and development activities of DLR Institute of Communications and Navigation dealing with the introduction of data and system integrity into the maritime traffic system. One of the aimed objectives is the automatic assessment of the traffic situation aboard a vessel including integrity information.

International audience; In this paper we present two datasets of instrumental gestures performed with expressive variations: ve violinists performing standard pedagogical phrases with variation in dynamics and tempo; and two pianists performing a repertoire piece with variations in tempo, dynamics and articulation. We show the utility of these datasets by highlighting the diierent movement qualities embedded in both datasets. In addition, for the violin dataset, we report on gesture recognition tests using two state-of-the-art realtime gesture recognizers. We believe that these resources create opportunities for further research on the understanding of complex human movements through computational methods.

Date of Conference: 15-18 May 2017 Conference Name: IEEE 25th Signal Processing and Communications Applications Conference, SIU 2017 In this paper, we investigate online nonlinear regression and introduce novel algorithms based on the long short term memory (LSTM) networks. We first put the underlying architecture in a nonlinear state space form and introduce highly efficient particle filtering (PF) based updates, as well as, extended Kalman filter (EKF) based updates. Our PF based training method guarantees convergence to the optimal parameter estimation under certain assumptions. We achieve this performance with a computational complexity in the order of the first order gradient based methods by controlling the number of particles. The experimental results illustrate significant performance improvements achieved by the introduced algorithms with respect to the conventional methods.

The problem of damage detection in beams from time domain vibration measurements is investigated. A cantilever beam subjected to dynamic point loadings is considered. A measurable damage in the form of a crack-like defect is introduced into the beam. Using Bayesian principles, a particle filtering algorithm applied on the experimentally measured tip accelerations is used to estimate the changes in damping and in the flexural rigidity of the beam. Subsequently, the size and location of the crack are estimated using principles of fracture mechanics and particle filtering. The proposed method bypasses the need for baseline measurements of the vibration response of the undamaged beam. � 2014, Springer-Verlag Berlin Heidelberg.

Fault diagnosis permits computational redundancy, which renders a system sustainable and eventually leads to hardware cost reduction. To achieve the posterior distribution computation needed for fault diagnosis along with motion estimation, we suggest a particle filtering (PF)-based state-segmentation approach. Here, both a continuous state vector and fault states are segmented accordingly to allow flexible reasoning for fault diagnosis and motion estimation. For each segmented space, an attempt is made to construct a corresponding posterior distribution independently, resulting in a reduction of the number of particles. Our experimental simulation demonstrates fault diagnosis among billions of fault states. Our state-segmentation approach reduced 98% of particles compared with the ordinal PF approach., 資料番号: PA1510083000

Percutaneous image-guided tumor ablation is a minimally invasive surgical procedure for the treatment of malignant tumors using a needle-shaped ablation probe. Automating the insertion of a needle by using a robot could increase the accuracy and decrease the execution time of the procedure. Extracting the needle tip position from the ultrasound (US) images is of paramount importance for verifying that the needle is not approaching any forbidden regions (e.g., major vessels and ribs), and could also be used as a direct feedback signal to the robot inserting the needle. A method for estimating the needle tip has previously been developed combining a modified Hough transform, image filters, and machine learning. This paper improves that method by introducing a dynamic selection of the region of interest in the US images and filtering the tracking results using either a Kalman filter or a particle filter. Experiments where a biopsy needle has been inserted into a phantom by a robot have been conducted, guided by an infrared tracking system. The proposed method has been accurately evaluated by comparing its estimations with the needle tip's positions manually detected by a physician in the US images. The results show a significant improvement in precision and more than 85% reduction of 95th percentile of the error compared with the previous automatic approaches. The method runs in real time with a frame rate of 35.4 frames/s. The increased robustness and accuracy can make our algorithm usable in autonomous surgical systems for needle insertion.

Expert knowledge in the form of mathematical models can be considered sufficient statistics of all prior experimentation in the domain, embodying generic or abstract knowledge of it. When used in a probabilistic framework, such models provide a sound foundation for data mining, inference, and decision making under uncertainty. We describe a methodology for encapsulating knowledge in the form of ordinary differential equations (ODEs) in dynamic Bayesian networks (DBNs). The resulting DBN framework can handle both data and model uncertainty in a principled manner, can be used for temporal data mining with noisy and missing data, and can be used to re-estimate model parameters automatically using data streams. A standard assumption when performing inference in DBNs is that time steps are fixed. Generally, the time step chosen is small enough to capture the dynamics of the most rapidly changing variable. This can result in DBNs having a natural time step that is very short, leading to inefficient inference; this is particularly an issue for DBNs derived from ODEs and for systems where the dynamics are not uniform over time. We propose an alternative to the fixed time step inference used in standard DBNs. In our algorithm, the DBN automatically adapts the time step lengths to suit the dynamics in each step. The resulting system allows us to efficiently infer probable values of hidden variables using multiple time series of evidence, some of which may be sparse, noisy or incomplete. We evaluate our approach with a DBN based on a variant of the van der Pol oscillator, and demonstrate an example where it gives more accurate results than the standard approach, but using only one tenth the number of time steps. We also apply our approach to a real-world example in critical care medicine. By incorporating knowledge in the form of an existing ODE model, we have built a DBN framework for efficiently predicting individualised patient responses using the available bedside and lab data.

International audience; Multi-object tracking is a difficult problem underlying many computer vision applications. In this work, we focus on sediment transport experiments in a flow were sediments are represented by spherical calibrated beads. The aim is to track all beads over long time sequences to obtain sediment velocities and concentration. Classical algorithms used in fluid mechanics fail to track the beads over long sequences with a high precision because they incorrectly handle both miss-detections and detector imprecision. Our contribution is to propose a particle filter-based algorithm including an adapted multiple motion model. Additionally, this algorithm integrates several improvements to account for the lack of precision of the detector. The evaluation was made using a test sequence with a dedicated ground-truth. The results show that the method outperforms state-of-the-art concurrent algorithms.

International audience; Interference are among the most penalizing error sources in global positioning system (GPS) navigation. So far, many effort has been devoted to developing GPS receivers more robust to the radio-frequency environment. Contrary to previous approaches, this paper does not aim at improving the estimation of the GPS pseudoranges between the mobile and the GPS satellites in the presence of interference. As an alternative, we propose to model interference effects as variance jumps affecting the GPS measurements which can be directly detected and compensated at the level of the navigation algorithm. Since the joint detection/estimation of the interference errors and motion parameters is a highly non linear problem, a particle filtering technique is used. An original particle filter is developed to improve the detection performance while ensuring a good accuracy of the positioning solution.

When the dimensions of polycrystalline structures become comparable to the average grain size, some reliability issues can be reported for the moving parts of inertial microelectromechanical systems (MEMS). Not only the overall behavior of the device turns out to be affected by a large scattering, but also the sensitivity to imperfections gets enhanced. In this work, through on-chip tests, we experimentally investigate the behavior of thin polysilicon samples using standard electrostatic actuation/sensing. The discrepancy between the target and actual responses of each sample has then been exploited to identify: (i) the overall stiffness of the film and, according to standard continuum elasticity, a morphology-based value of its Young’s modulus; (ii) the relevant over-etch induced by the fabrication process. To properly account for the aforementioned stochastic features at the micro-scale, the identification procedure has been based on particle filtering. A simple analytical reduced-order model of the moving structure has been also developed to account for the nonlinearities in the electrical field, up to pull-in. Results are reported for a set of ten film samples of constant slenderness, and the effects of different actuation mechanisms on the identified micromechanical features are thoroughly discussed.

This paper is concerned with cooperative Terrain Aided Navigation of a network of aircraft using fusion of Radar Altimeter and inter-node range measurements. State inference is performed using a Rao-Blackwellized Particle Filter with online measurement noise statistics estimation. For terrain coverage measurement noise parameter identification, an online Expectation Maximization algorithm is proposed, where local sufficient statistics at each node are calculated in the E-step, which are then distributed to neighboring nodes using a random gossip algorithm to perform the M-step at each node. Simulation results show that improvement on positioning and calibration performance can be achieved compared to a non-cooperative approach.

International audience; This paper addresses the problem of 3D tracking of human gesture for buying behavior estimation. The top view of the customers, which has been little treated for human tracking, is exploited in this particular context. This point of view avoids occlusion except for those of the arms. We propose an hybrid 3D-2D tracking method based on the particle filtering framework, which uses the exclusion principle to separate the observation related to each customer and deals with multi-person tracking. The head and shoulders are tracked in the 2D space, while the arms are tracked in the 3D space: these are the spaces where they are the most descriptive. We validate our method both experimentally, so as to obtain qualitative results, and on-site. We demonstrated that it makes a good estimation for various cases and situations in real-time ((Formula presented.)40 fps).

Particle filtering (PF) is being applied successfully in nonlinear and/or non-Gaussian system failure prognosis. However, for failure prediction of many complex systems whose dynamic state evolution models involve time-varying parameters, the traditional PF-based prognosis framework will probably generate serious deviations in results since it implements prediction through iterative calculation using the state models. To address the problem, this paper develops a novel integrated PF-LSSVR framework based on PF and least squares support vector regression (LSSVR) for nonlinear system failure prognosis. This approach employs LSSVR for long-term observation series prediction and applies PF-based dual estimation to collaboratively estimate the values of system states and parameters of the corresponding future time instances. Meantime, the propagation of prediction uncertainty is emphatically taken into account. Therefore, PF-LSSVR avoids over-dependency on system state models in prediction phase. With a two-sided failure definition, the probability distribution of system remaining useful life (RUL) is accessed and the corresponding methods of calculating performance evaluation metrics are put forward. The PF-LSSVR framework is applied to a three-vessel water tank system failure prognosis and it has much higher prediction accuracy and confidence level than traditional PF-based framework.

In this paper, joint identification for structural systems, characterized by severe nonlinearities (softening) in the constitutive model, is pursued via the Sigma-Point Kalman Filter (S-PKF) and the Particle Filter (PF). Since a formal proof of the effects of softening in a stochastic structural system on the accuracy and stability of the filters is still missing, we comparatively assess the performances of S-PKF and PF. We show that the PF displays a higher convergence rate towards steady-state model calibrations and the S-PKF is less sensitive to the measurement noise. Both S-PKF and PF are robust, even if they tend to get unstable when a structural failure is triggered.

International audience; Map-matching means determining the location of a mobile with respect to a road network description stored in a digital map. This problem is usually addressed using GPS-like fixes. Unfortunately, there are many situations in urban areas where few satellites are visible because of outages due to tall buildings. In this paper, map-matching is solved using raw GPS measurements (pseudoranges and Doppler measurements), avoiding the necessity to compute a global position. The problem is formalized in a general Bayesian framework in order to handle noise, which is able to perform multi-hypothesis map-matching when there is not enough information to make unambiguous decisions. This tightly-coupled GPS-Map fusion has to cope simultaneously with identifying the road and estimating the mobile's position on that road. A marginalized particle filter is proposed for solving this hybrid estimation problem efficiently. Real experimental results are reported to show that this approach can be initialized with fewer than four satellites. It is also able to track the location with two satellites only, once the road selection has been solved.

Particle filter-based acoustic source tracking algorithms track (online and in real-time) the position of a sound source-a person speaking in a room-based on the current data from a distributed microphone array as well as the previously recorded data. This paper develops a multi-target tracking (MTT) methodology to allow for an unknown and time-varying number of speakers in a fully probabilistic manner and in doing so does not resort to independent modules for new target proposal or target number estimation as in previous works. The approach uses the concept of an existence grid to propose possible regions of activity before tracking is carried out with a variable dimension particle filter-which also explicitly supports the concept of a null particle, containing no target states, when no speakers are active. Examples demonstrate typical tracking performance in a number of different scenarios with simultaneously active speech sources.

The paper presents an algorithm for estimating a pedestrian location in an urban environment. The algorithm is based on the particle filter and uses different data sources: a GPS receiver, inertial sensors, probability maps and a stereo camera. Inertial sensors are used to estimate a relative displacement of a pedestrian. A gyroscope estimates a change in the heading direction. An accelerometer is used to count a pedestrian's steps and their lengths. The so-called probability maps help to limit GPS inaccuracy by imposing constraints on pedestrian kinematics, e.g., it is assumed that a pedestrian cannot cross buildings, fences etc. This limits position inaccuracy to ca. 10 m. Incorporation of depth estimates derived from a stereo camera that are compared to the 3D model of an environment has enabled further reduction of positioning errors. As a result, for 90% of the time, the algorithm is able to estimate a pedestrian location with an error smaller than 2 m, compared to an error of 6.5 m for a navigation based solely on GPS.

A large number of particles are needed to improve the precision of particle filtering of fast simultaneous localization and mapping (FastSLAM) for uninhabited aerial vehicle (UAV). Simulated annealing (SA) method is proposed to solve this problem. Through simulated annealing algorithm, the particle's prediction is updated and proposal distribution is adjusted in FastSLAM, so the particles are concentrated around the true pose. Stratified resampling is used to solve the degeneracy of particles. At last, the proposed algorithm is tested for UAV FastSLAM. The result shows this method is effective and feasible.

A method derived from the adaptive Sampling Importance Resampling particle filter is proposed to solve vehicle detection and tracking problems in laser range finder data. The originality of this approach lies in a joint detection and tracking of the objects. To this end, the solution is based on a matched filter which uses a predefined vehicle model. The non-linearity of the state equations is tackled by sequential Monte Carlo methods which are here the basis of our solution. A central point here is to calculate the weights of the matched particle filter, according to the vehicle model. The proposed approach is then applied to synthetic data from a road scenario. The efficiency of the method is shown in terms of estimation accuracies and detection.

This paper derives a particle filter algorithm within the Dempster–Shafer framework. Particle filtering is a well-established Bayesian Monte Carlo technique for estimating the current state of a hidden Markov process using a fixed number of samples. When dealing with incomplete information or qualitative assessments of uncertainty, however, Dempster–Shafer models with their explicit representation of ignorance often turn out to be more appropriate than Bayesian models.The contribution of this paper is twofold. First, the Dempster–Shafer formalism is applied to the problem of maintaining a belief distribution over the state space of a hidden Markov process by deriving the corresponding recursive update equations, which turn out to be a strict generalization of Bayesian filtering. Second, it is shown how the solution of these equations can be efficiently approximated via particle filtering based on importance sampling, which makes the Dempster–Shafer approach tractable even for large state spaces. The performance of the resulting algorithm is compared to exact evidential as well as Bayesian inference.

Particle Filter-based Acoustic Source Localization algorithms attempt to track the position of a sound source - one or more people speaking in a room - based on the current data from a microphone array as well as all previous data up to that point. This paper first discusses some of the inherent behavioral traits of the steered beamformer localization function. Using conclusions drawn from that study, a multitarget methodology for acoustic source tracking based on the Track Before Detect (TBD) framework is introduced. The algorithm also implicitly evaluates source activity using a variable appended to the state vector. Using the TBD methodology avoids the need to identify a set of source measurements and also allows for a vast increase in the number of particles used for a comparitive computational load which results in increased tracking stability in challenging recording environments. An evaluation of tracking performance is given using a set of real speech recordings with two simultaneously active speech sources.

The reliable and accurate indoor pedestrian positioning is one of the biggest challenges for location-based systems and applications. Most pedestrian positioning systems have drift error and large bias due to low-cost inertial sensors and random motions of human being, as well as unpredictable and time-varying radio-frequency (RF) signals used for position determination. To solve this problem, many indoor positioning approaches that integrate the user’s motion estimated by dead reckoning (DR) method and the location data obtained by RSS fingerprinting through Bayesian filter, such as the Kalman filter (KF), unscented Kalman filter (UKF), and particle filter (PF), have recently been proposed to achieve higher positioning accuracy in indoor environments. Among Bayesian filtering methods, PF is the most popular integrating approach and can provide the best localization performance. However, since PF uses a large number of particles for the high performance, it can lead to considerable computational cost. This paper presents an indoor positioning system implemented on a smartphone, which uses simple dead reckoning (DR), RSS fingerprinting using iBeacon and machine learning scheme, and improved KF. The core of the system is the enhanced KF called a sigma-point Kalman particle filter (SKPF), which localize the user leveraging both the unscented transform of UKF and the weighting method of PF. The SKPF algorithm proposed in this study is used to provide the enhanced positioning accuracy by fusing positional data obtained from both DR and fingerprinting with uncertainty. The SKPF algorithm can achieve better positioning accuracy than KF and UKF and comparable performance compared to PF, and it can provide higher computational efficiency compared with PF. iBeacon in our positioning system is used for energy-efficient localization and RSS fingerprinting. We aim to design the localization scheme that can realize the high positioning accuracy, computational efficiency, and energy efficiency through the SKPF and iBeacon indoors. Empirical experiments in real environments show that the use of the SKPF algorithm and iBeacon in our indoor localization scheme can achieve very satisfactory performance in terms of localization accuracy, computational cost, and energy efficiency.

International audience; In GPS navigation, distortion of the delay and phase of the received signal due to multipath propagation can degrade seriously the position estimation. In some applications such as aircraft landing, detecting the occurrence of multipath can be as important as mitigating its effect on the pseudodistance. This paper proposes deterministic particle filtering for joint multipath detection and navigation parameter estimation. The efficiency of the proposed method is illustrated by numerical simulations

A major challenge in applying Bayesian tracking methods for tracking 3D human body pose is the high dimensionality of the pose state space. It has been observed that the 3D human body pose parameters typically can be assumed to lie on a low-dimensional manifold embedded in the high-dimensional space. The goal of this work is to approximate the low-dimensional manifold so that a low-dimensional state vector can be obtained for efficient and effective Bayesian tracking. To achieve this goal, a globally coordinated mixture of factor analyzers is learned from motion capture data. Each factor analyzer in the mixture is a “locally linear dimensionality reducer” that approximates a part of the manifold. The global parametrization of the manifold is obtained by aligning these locally linear pieces in a global coordinate system. To enable automatic and optimal selection of the number of factor analyzers and the dimensionality of the manifold, a variational Bayesian formulation of the globally coordinated mixture of factor analyzers is proposed. The advantages of the proposed model are demonstrated in a multiple hypothesis tracker for tracking 3D human body pose. Quantitative comparisons on benchmark datasets show that the proposed method produces more accurate 3D pose estimates over time than those obtained from two previously proposed Bayesian tracking methods.