Consensus-based algorithms for distributed filtering

Consensus‐based multiple‐model Bayesian filtering for distributed tracking

“…Although such algorithms can perform comparably to a centralised filter when a large number of consensus iterations are carried out during each sampling interval [19], their performance can degrade substantially when only a few consensus iterations can be afforded. In fact, it is known that likelihood consensus algorithms require a sufficiently large number of consensus rounds per sampling interval to achieve stable distributed state estimation (DSE) [12]. The present paper provides significant follow-ups of [18], in that a different consensus approach is adopted, which provides comparable performance with a remarkably smaller amount of consensus iterations and, hence, turns out to be more energy-efficient since it involves a smaller communication overhead.…”

Consensus-Based Linear and Nonlinear Filtering

“…As shown in [16], this approach, that will be referred to as consensus on measurements (CM), can be extended to nonlinear systems by following the EKF paradigm. To this end, the information filter of Table I can be exploited, with the only difference that the virtual measurements y j t has to be redefined in terms of the local state predictionsx j t|t−1 instead of the centralized onex t|t−1 (which is not available in a distributed setting).…”

“…Different choices for the combination weights are discussed, leading to different properties. Following [16], where preliminary results on the subject can be found, the nonlinear case is treated by resorting to the Extended Kalman Filter (EKF) paradigm. The main contribution of this note is a stability analysis of the proposed family of HCMCI filters in the case of linear systems.…”

Survey on Estimation

“…Recently, Battistelli and Chisci (2014) introduced a generalized consensus on probability densities which opens up the possibility to perform in a fully distributed and scalable way any Bayesian estimation task over a sensor network. As by-products, this approach allowed to derive consensus Kalman filters with guaranteed stability under minimal requirements of system observability and network connectivity (Battistelli et al 2011(Battistelli et al , 2012Battistelli and Chisci 2014), consensus nonlinear filters (Battistelli et al 2012), and a consensus CPHD filter for distributed multitarget tracking (Battistelli et al 2013). Despite these interesting preliminary results, networked estimation is still a very active research area with many open problems related to energy efficiency, estimation performance optimality, robustness with respect to delays and/or data losses, etc.…”

“…1981Grimble, etc. -1988 Robust (e.g., H 1 ) filtering (Poor and Looze 1981;Darragh and Looze 1984;Verdú and Poor 1984;Grimble 1988;Hassibi et al 1999;Simon 2006 (Dai 1987(Dai , 1989Nikoukhah et al 1992;Chisci and Zappa 1992) Stankovic et al 2009;Battistelli et al 2011Battistelli et al , 2012Battistelli et al , 2013Battistelli and Chisci 2014) for networked estimation to be included in the KF framework. For these reasons, the KF is, and continues to be, an extremely effective and easy-to-implement tool for a great variety of practical tasks, e.g., to detect signals in noise or to estimate unmeasurable quantities from accessible observables.…”