Situation-aware technologies enabled by multitarget tracking will lead to new services and applications in fields such as autonomous driving, indoor localization, robotic networks, crowd counting. In this tutorial paper, we advocate a recently proposed paradigm for scalable that is based on message passing or, more concretely, the loopy sum-product algorithm. This approach has advantages regarding estimation accuracy, computational complexity, implementation flexibility. Most importantly, it...

System-level simulations have become an indispensable tool for predicting the behavior of wireless cellular systems. As exact link-level modeling is unfeasible due to its huge complexity, mathematical abstraction required obtain equivalent results by less complexity. A particular problem in such approaches multiple coherent transmissions. Those arise multiple-input-multiple-output transmissions at every base station but nowadays so-called coordinated multipoint (CoMP) techniques very...

We propose a fast labeled multi-Bernoulli (LMB) filter that uses belief propagation for probabilistic data association. The complexity of our scales only linearly in the numbers Bernoulli components and measurements, while performance is comparable to or better than Gibbs sampler-based LMB filter.

Tracking an unknown number of low-observable objects is notoriously challenging. This letter proposes a sequential Bayesian estimation method based on the track-before-detect (TBD) approach. In TBD, raw sensor measurements are directly used by tracking algorithm without any preprocessing. Our proposed new statistical model that introduces object hypothesis for each data cell measurements. It allows to interact and contribute more than one cell. Based factor graph representing our model, we...

Joint probabilistic data association (JPDA) and multiple hypothesis tracking (MHT) introduced in the 70s, are still widely used methods for multitarget (MTT). Extensive studies over last few decades have revealed undesirable behavior of JPDA MHT type scenarios with targets close proximity. In particular, suffers from track coalescence effect, i.e., estimated tracks proximity tend to merge can become indistinguishable. Interestingly, MHT, an opposite effect called repulsion be observed. this...

In this article, we propose an efficient random finite set (RFS)-based algorithm for multiobject tracking, in which the object states are modeled by a combination of labeled multi-Bernoulli (LMB) RFS and Poisson RFS. The less computationally demanding part is used to track potential objects whose existence unlikely. Only if quantity characterizing plausibility above threshold, new Bernoulli component created, tracked more accurate but LMB algorithm. Conversely, transferred back corresponding...

Passive acoustic monitoring is widely used for detection and localization of marine mammals. Typically, pressure sensors are used, although several studies utilized vector (AVSs), that measure particle velocity can estimate azimuths to sources. The AVSs localize sources using a reduced number do not require precise time synchronization between sensors. However, when multiple animals calling concurrently, automated tracking individual still poses challenge, manual methods typically employed...

We propose a scalable track-before-detect (TBD) tracking method based on Poisson/multi-Bernoulli model. To limit computational complexity, we approximate the exact multi-Bernoulli mixture posterior probability density function (pdf) by pdf. Data association sum-product algorithm and recycling of Bernoulli components enable detection low-observable objects with limited resources. Our simulation results demonstrate significantly improved performance compared to state-of-the-art TBD method.

An analysis of the motion common carotid artery (CCA) provides effective indicators for cardiovascular diseases. Here, we propose a method tracking CCA wall from B-mode ultrasound video sequence. unscented Kalman filter based on suitably devised state-space model fuses measurements produced by an optical flow algorithm and localization algorithm. This approach compensates feature drift, which is detrimental effect in algorithms. The proposed demonstrated to outperform state-of-the-art flow.

In many multiobject tracking applications, including radar and sonar tracking, after prefiltering the received signal, measurement data is typically structured in cells. The cells, e.g., represent different range bearing values. However, conventional methods use so-called point measurements. Point measurements are provided by a preprocessing stage that applies threshold or detector breaks up cell's structure converting cell indexes into, We here propose Bayesian method processes have been...

We propose a distributed multisensor joint integrated probabilistic data association (JIPDA) filter for multiobject tracking in decentralized sensor networks. Conventional Chernoff fusion of the posterior distributions neighboring sensors presupposes correct "hard" objects tracked at sensors. To avoid detrimental effects an incorrect hard association, we develop method based on ("soft") object association. Our numerical results demonstrate significant performance gains relative to approach.

We consider a distributed labeled multi-Bernoulli (LMB) filter that uses the generalized covariance intersection technique for fusing local LMB distributions. A critical aspect of such filters is to correctly associate Bernoulli components describing same object at different sensors. Here, we improve on previously proposed association schemes by introducing probabilistic framework and algorithm (label) association. Instead enforcing hard association, propose compute probabilities use them in...

Analyzing the motion of common carotid artery (CCA) wall yields effective indicators for atherosclerosis. In this work, we propose a state-space model and tracking method estimating time-varying CCA radius from B-mode ultrasound sequence arbitrary length. We employ an unscented Kalman filter that fuses two sets measurements produced by optical flow algorithm localization algorithm. This fusion-and-tracking approach ensures feature drift, which tends to impair based methods, is compensated in...

Joint probabilistic data association (JPDA) filter methods and multiple hypothesis tracking (MHT) are widely used for multitarget (MTT). However, they known to exhibit undesirable behavior in scenarios with targets close proximity: JPDA suffer from the track coalescence effect, i.e., estimated tracks of proximity tend merge can become indistinguishable, MHT an opposite effect as repulsion. In this paper, we review discuss repulsion effects. We also consider a more recent methodology MTT that...

The following topics are dealt with: sensor fusion; target tracking; stochastic processes; radar Kalman filters; object mobile robots; optical radar; Bayes methods; image filtering.

Analyzing the motion of wall common carotid artery (CCA) yields effective indicators for atherosclerosis. In this work, we explore use multitarget tracking techniques estimating time-varying CCA radius from an ultrasound video sequence. We employ joint integrated probabilistic data association (JIPDA) filter to track a set "feature points" (FPs) located around cross section. Subsequently, estimate via non-linear least-squares method and Kalman filter. The application JIPDA is enabled by...

Tracking an unknown number of low-observable objects is notoriously challenging. This letter proposes a sequential Bayesian estimation method based on the track-before-detect (TBD) approach. In TBD, raw sensor measurements are directly used by tracking algorithm without any preprocessing. Our proposed new statistical model that introduces object hypothesis for each data cell measurements. It allows to interact and contribute more than one cell. Based factor graph representing our model, we...

We propose an efficient random finite set (RFS) based algorithm for multiobject tracking in which the object states are modeled by a combination of labeled multi-Bernoulli (LMB) RFS and Poisson RFS. The less computationally demanding part is used to track potential objects whose existence unlikely. Only if quantity characterizing plausibility above threshold, new Bernoulli component created tracked more accurate but LMB algorithm. Conversely, transferred back corresponding probability falls...

Joint probabilistic data association (JPDA) and multiple hypothesis tracking (MHT) introduced in the 70s, are still widely used methods for multitarget (MTT). Extensive studies over last few decades have revealed undesirable behavior of JPDA MHT type scenarios with targets close proximity. In particular, suffers from track coalescence effect, i.e., estimated tracks proximity tend to merge can become indistinguishable. Interestingly, MHT, an opposite effect called repulsion be observed. this...

We propose a scalable track-before-detect (TBD) tracking method based on Poisson/multi-Bernoulli model. To limit computational complexity, we approximate the exact multi-Bernoulli mixture posterior probability density function (pdf) by pdf. Data association sum-product algorithm and recycling of Bernoulli components enable detection low-observable objects with limited resources. Our simulation results demonstrate significantly improved performance compared to state-of-the-art TBD method.