Cost-effective diffusion Kalman filtering with implicit measurement exchanges

Talebi, Sayed Pouria; Kanna, Sithan; Xia, Yili; Mandic, Danilo P.

doi:10.1109/icassp.2017.7952990

“…In Figure 2, the steady-state MSD of the proposed distributed Kalman filter is benchmarked against that of the diffusion Kalman filter (Algorithm 2 in [15]), the consensus Kalman filter (Algorithm 3 in [22]), and the distributed Kalman filter in [30]. Note that the proposed algorithm achieved a steadystate MSD close to that of the centralized Kalman filter and outperformed the consensus and diffusion schemes.…”

Section: A Distributed Filtering and Trackingmentioning

confidence: 97%

“…Note that the MSD at each node is the trace of its state vector estimation error covariance matrix obtainable in its transient and steady-state formulation from(29)and(30).…”

mentioning

confidence: 99%

Distributed Kalman Filtering and Control Through Embedded Average Consensus Information Fusion

Talebi

¹

,

Werner

²

2019

IEEE Trans. Automat. Contr.

Self Cite

View full text Add to dashboard Cite

This work presents a unified framework for distributed filtering and control of state-space processes. To this end, a distributed Kalman filtering algorithm is developed via decomposition of the optimal centralized Kalman filtering operations. This decomposition is orchestrated in a fashion so that each agent retains a Kalman style filtering operation and an estimate of the state vector. In this setting, the agents mirror the operations of the centralized Kalman filter in a distributed fashion through embedded average consensus fusion of local state vector estimates and their associated covariance information. For rigor, closed-form expressions for the mean and mean square error performance of the developed distributed Kalman filter are derived. More importantly, in contrast to current approaches, due to the comprehensive framework for fusion of the covariance information, a duality between the developed distributed Kalman filter and decentralized control is established. Thus, resulting in an effective and all inclusive distributed framework for filtering and control of state-space processes over a network of agents. The introduced theoretical concepts are validated using simulations that indicate a precise match between simulation results and the theoretical analysis. In addition, simulations indicate that performance levels comparable to that of the optimal centralized approaches are attainable.

show abstract

“…The optimal solution to this problem, in the mean square error sense, comes in the form of a centralized Kalman filter implemented at each node [15,29], the operations of which are summarized as: For node i ∈ N : Initialize with:…”

Section: Problem Formulationmentioning

confidence: 99%

“…wherex i,n|n−1 andx i,n|n denote the a priori and a posteriori estimates of xi,n, whereas Ci denotes the set of nodes in the same cluster as node i. From [29], if i and l are in the same cluster, we have M i,n|n = M l,n|n . Finally, note that it is assumed a central regulating unit that has prior knowledge of the network clustering structure organizes the distribution of information over the network.…”

Section: Problem Formulationmentioning

confidence: 99%

See 1 more Smart Citation

Kalman Filtering and Clustering in Sensor Networks

Talebi

¹

,

Werner

²

,

Koivunen

³

2018

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

Self Cite

View full text Add to dashboard 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.

show abstract

Distributed Kalman Filtering: Consensus, Diffusion, and Mixed

Talebi

¹

,

Werner

²

2018

2018 IEEE Conference on Control Technology and Applications (CCTA)

Self Cite

View full text Add to dashboard 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.

show abstract

Cost-effective diffusion Kalman filtering with implicit measurement exchanges

Cited by 10 publications

References 22 publications

Distributed Kalman Filtering and Control Through Embedded Average Consensus Information Fusion

Distributed Kalman Filtering and Control Through Embedded Average Consensus Information Fusion

Kalman Filtering and Clustering in Sensor Networks

Distributed Kalman Filtering: Consensus, Diffusion, and Mixed

Contact Info

Product

Resources

About