A Distributed Quaternion Kalman Filter With Applications to Smart Grid and Target Tracking

IEEE Control Syst. Lett.

2019

Self Cite

This paper presents a fully distributed approach for tracking state vector sequences over sensor networks in presence of unknown actuations. The problem arises in large-scale systems where modeling the full dynamics becomes impractical. In this work, the network only considers a subsection of the overall system which it can detect while accounting other inputs as unknown actuations. First a centralized technique that can consolidate all the available observation information is introduced. Then, operations of this optimal centralized solution are decomposed in a manner to allow their implementation in a distributed fashion while allowing each agent to retain an estimate of both the state vector and unknown actuations. The filter is derived in both diffusion and consensus formulations. The diffusion formulation is intended as a cost-effective solution, while the consensus formulation trades implementation complexity for accuracy.

Section: A Diffusion Formulationmentioning

confidence: 99%

Distributed Kalman Filtering in Presence of Unknown Outer Network Actuations

IEEE Control Syst. Lett.

2019

Self Cite

“…The ACF described in (18)- (20) allows the update operation in (16) to be performed in a distributed manner. The operations of such a distributed fractional least mean square (DFLMS) filter are summarized in Algorithm 1, whereĤ l,n denotes the estimate of H obtained at time instant n at node l. For the case of α = 2, as α → 2; then, the proposed DFLMS (Algorithm 1) simplifies into the distributed least mean square (DLMS) in [34,35].…”

Section: Distributed Adaptive Approachmentioning

confidence: 99%

Distributed Adaptive Filtering of $\alpha$-Stable Signals

IEEE Signal Process. Lett.

Mandic

2018

Self Cite

A cost-effective framework for distributed adaptive filtering of α-stable signals over sensor networks is proposed. First, the filtering paradigm of α-stable signals through multiple observations made over a network of sensors is revisited and an optimal solution is formulated. Then, an adaptive gradient descent based algorithm for distributed real-time filtering of αstable signals via multi-agent networks is derived. This not only provides an approximation of the formulated optimal solution, but also a cost-effective algorithm which scales with the size of the network. Moreover, performance of the derived algorithm is analyzed and convergence conditions are established.Index Terms-Sensor networks, distributed adaptive filtering, consensus fusion, fractional differential, α-stable random signals.

“…and ζ l,n = Υ l,n l,n−1|n−1 , whereas ξ l,n =p(ψ l,n |yi,1:n) (I − G l,n H l,n ) ν l,n χ l,n =p(ψ l,n |yi,1:n)G l,n ω l,n . Now, the following typical conditions in Kalman filtering analysis are held to be true [14,30,31]:…”

Section: Convergence Analysismentioning

confidence: 99%

“…ν l,n } are controllable. Then, if the diffusion coefficients {p(ψ l,n |yi,1:n) : ∀l, i ∈ N } are held constant, that is the cluster structure of the network becomes time invariant, from the framework introduced in [14,30], it follows that the matrices {M l,n|n : l ∈ N } become time invariant. Moreover, from Algorithm 1, time invariant matrices {M l,n|n : l ∈ N } result in the matrices {G l,n : l ∈ N } also becoming time invariant and therefore C i,n|n as expressed in (25) converges.…”

Section: Convergence Analysismentioning

confidence: 99%

“…In recent years, distributed learning and optimization techniques have become the prevailing method for implementing signal processing and control operations over large-scale sensor networks [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16]. In this context, due to the optimality of the Kalman filter for tracking linear Gaussian systems and the flexibility of the state space representation for modeling a wide range of real-world dynamic systems, a great deal of interest has been shown in developing distributed Kalman filtering frameworks [4,[8][9][10][11][12][13][14][15], where in-network cooperation between agents of the sensor network is used to enhance various performance criteria, such as accuracy. However, cooperation is not beneficial in all incidences.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Kalman Filtering and Clustering in Sensor Networks

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Koivunen

2018

Self Cite

In this work, a distributed Kalman filtering and clustering framework for sensor networks tasked with tracking multiple state vector sequences is developed. This is achieved through recursively updating the likelihood of a state vector estimation from one agent offering valid information about the state vector of its neighbors, given the available observation data. These likelihoods then form the diffusion coefficients, used for information fusion over the sensor network. For rigour, the mean and mean square behavior of the developed Kalman filtering and clustering framework is analyzed, convergence criteria are established, and the performance of the developed framework is demonstrated in a simulation example.Index Terms-Adaptive learning over networks, distributed Kalman filtering, adaptive clustering, multi-task sensor networks.