Showing papers by "Luigi Chisci published in 2021"

Centralized Cooperative Sensor Fusion for Dynamic Sensor Network With Limited Field-of-View via Labeled Multi-Bernoulli Filter

Abstract: This paper presents a new solution for multi-target tracking over a network of sensors with limited spatial coverage. The proposed solution is based on the centralized data fusion architecture. The main contribution of the paper is the introduction of a new track-to-track fusion approach in which the posterior distributions of multi-target states, reported by various sensor nodes, are fused in a way that the redundant information are combined and the rest complement each other. The proposed solution is formulated within the labeled random finite set framework in which the fused posterior incorporates all the state and label information provided by multiple sensor nodes. The performance of the proposed method is evaluated via simulation experiments that involve challenging tracking scenarios. The proposed method is implemented using sequential Monte Carlo method and the results confirm its effectiveness.

Fusion-Based Multidetection Multitarget Tracking With Random Finite Sets

Abstract: Multidetection (MD) systems are characterized by multiple observation modes (OMs), and hence, simultaneously produce multiple measurements for each target. The key challenge in exploiting MD systems for multitarget tracking (MTT), compared to single-detection (SD) systems, is the significant amount of extra computational burden involved in order to solve the resulting multidimensional assignment problem among measurements, targets, and OMs. This article presents a novel computationally efficient MTT framework for MD systems, wherein the multitarget state is modeled as a random finite set (RFS), and a bank of OM-dependent MTT RFS filters with SD model are employed to recursively provide OM-dependent posteriors. The latter, which contain both real and false targets, are then suitably fused so as to enhance consensus on the true targets while weakening trust on the existence of the false ones. In this way, the computational complexity is significantly reduced compared to existing MTT algorithms with the MD model. Two representative RFS filters, i.e., unlabeled probability hypothesis density (PHD) and labeled multi-Bernoulli (LMB), are considered in the proposed framework and the computational complexity of the resulting MD MTT algorithms is analyzed. Performance of the proposed approach is assessed by simulation experiments in both over-the-horizon-radar (OTHR) and single-frequency-network passive radar (SFN-PR) MTT applications.

PHD-SLAM 2.0: Efficient SLAM in the Presence of Missdetections and Clutter

Abstract: This article addresses simultaneous localization and mapping (SLAM) via probability hypothesis density (PHD) filtering. The resulting approach, named PHD-SLAM, has demonstrated its effectiveness, especially when measurements provided by the sensors onboard the vehicle are highly contaminated by missdetections and clutter. However, since the proposal distribution (PD) of standard PHD-SLAM does not take into account most recently received measurements, a huge amount of particles are typically needed in order to achieve satisfactory performance. In this article, a new PD, which aims to approximate the vehicle pose posterior, is proposed for PHD-SLAM. The resulting algorithm, named PHD-SLAM 2.0, allows for drastically reducing the number of particles, and hence, the computational burden, while preserving the SLAM performance. The computational complexity of PHD-SLAM 2.0 is analyzed, and its performance is assessed via both simulated and real-data experiments.

Multi-sensor joint target detection, tracking and classification via Bernoulli filter.

Abstract: This paper focuses on \\textit{joint detection, tracking and classification} (JDTC) of a target via multi-sensor fusion. The target can be present or not, can belong to different classes, and depending on its class can behave according to different kinematic modes. Accordingly, it is modeled as a suitably extended Bernoulli \\textit{random finite set} (RFS) uniquely characterized by existence, classification, class-conditioned mode and class\\&mode-conditioned state probability distributions. By designing suitable centralized and distributed rules for fusing information on target existence, class, mode and state from different sensor nodes, novel \\textit{centralized} and \\textit{distributed} JDTC \\textit{Bernoulli filters} (C-JDTC-BF and D-JDTC-BF), are proposed. The performance of the proposed JDTC-BF approach is evaluated by means of simulation experiments.

A Variational Bayes Moving Horizon Estimation Adaptive Filter with Guaranteed Stability

Abstract: This paper addresses state estimation of linear systems with special attention on unknown process and measurement noise covariances, aiming to enhance estimation accuracy while preserving the stability guarantee of the Kalman filter. To this end, the full information estimation problem over a finite interval is firstly addressed. Then, a novel adaptive variational Bayesian (VB) moving horizon estimation (MHE) method is proposed, exploiting VB inference, MHE and Monte Carlo integration with importance sampling for joint estimation of the unknown process and measurement noise covariances, along with the state trajectory over a moving window of fixed length. Further, it is proved that the proposed adaptive VB MHE filter ensures mean-square boundedness of the estimation error with any number of importance samples and VB iterations, as well as for any window length. Finally, simulation results on a target tracking example demonstrate the effectiveness of the VB MHE filter with enhanced estimation accuracy and convergence properties compared to the conventional non-adaptive Kalman filter and other existing adaptive filters.

Principled information fusion for multi-view multi-agent surveillance systems.

Abstract: A key objective of multi-agent surveillance systems is to monitor a much larger region than the limited field-of-view (FoV) of any individual agent by successfully exploiting cooperation among multi-view agents. Whenever either a centralized or a distributed approach is pursued, this goal cannot be achieved unless an appropriately designed fusion strategy is adopted. This paper presents a novel principled information fusion approach for dealing with multi-view multi-agent case, on the basis of Generalized Covariance Intersection (GCI). The proposed method can be used to perform multi-object tracking on both a centralized and a distributed peer-to-peer sensor network. Simulation experiments on realistic multi-object tracking scenarios demonstrate effectiveness of the proposed solution.