This paper addresses a distributed filter over wireless sensor networks for optimal estimation. A distributed filter over the networks allows all local estimators to calculate optimal estimates with a scalable communication cost. Outputs of the distributed filter for the optimal estimation can be denoted as a solution of a consensus optimization problem. Thus, the distributed filter is designed based on distributed alternating direction method of multipliers (ADMM). The remarkable points of the distributed filter based on the ADMM are that: the distributed filter has a faster convergence rate than distributed subgradient projection algorithm; the weight, which is optimized by a semidefinite programming problem, accelerates the convergence rate of the proposed method.
Ryosuke ADACHI
Yamaguchi University
Yuji WAKASA
Yamaguchi University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Ryosuke ADACHI, Yuji WAKASA, "Distributed Filter Using ADMM for Optimal Estimation Over Wireless Sensor Network" in IEICE TRANSACTIONS on Fundamentals,
vol. E105-A, no. 11, pp. 1458-1465, November 2022, doi: 10.1587/transfun.2021KEP0008.
Abstract: This paper addresses a distributed filter over wireless sensor networks for optimal estimation. A distributed filter over the networks allows all local estimators to calculate optimal estimates with a scalable communication cost. Outputs of the distributed filter for the optimal estimation can be denoted as a solution of a consensus optimization problem. Thus, the distributed filter is designed based on distributed alternating direction method of multipliers (ADMM). The remarkable points of the distributed filter based on the ADMM are that: the distributed filter has a faster convergence rate than distributed subgradient projection algorithm; the weight, which is optimized by a semidefinite programming problem, accelerates the convergence rate of the proposed method.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2021KEP0008/_p
Copy
@ARTICLE{e105-a_11_1458,
author={Ryosuke ADACHI, Yuji WAKASA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Distributed Filter Using ADMM for Optimal Estimation Over Wireless Sensor Network},
year={2022},
volume={E105-A},
number={11},
pages={1458-1465},
abstract={This paper addresses a distributed filter over wireless sensor networks for optimal estimation. A distributed filter over the networks allows all local estimators to calculate optimal estimates with a scalable communication cost. Outputs of the distributed filter for the optimal estimation can be denoted as a solution of a consensus optimization problem. Thus, the distributed filter is designed based on distributed alternating direction method of multipliers (ADMM). The remarkable points of the distributed filter based on the ADMM are that: the distributed filter has a faster convergence rate than distributed subgradient projection algorithm; the weight, which is optimized by a semidefinite programming problem, accelerates the convergence rate of the proposed method.},
keywords={},
doi={10.1587/transfun.2021KEP0008},
ISSN={1745-1337},
month={November},}
Copy
TY - JOUR
TI - Distributed Filter Using ADMM for Optimal Estimation Over Wireless Sensor Network
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1458
EP - 1465
AU - Ryosuke ADACHI
AU - Yuji WAKASA
PY - 2022
DO - 10.1587/transfun.2021KEP0008
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E105-A
IS - 11
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - November 2022
AB - This paper addresses a distributed filter over wireless sensor networks for optimal estimation. A distributed filter over the networks allows all local estimators to calculate optimal estimates with a scalable communication cost. Outputs of the distributed filter for the optimal estimation can be denoted as a solution of a consensus optimization problem. Thus, the distributed filter is designed based on distributed alternating direction method of multipliers (ADMM). The remarkable points of the distributed filter based on the ADMM are that: the distributed filter has a faster convergence rate than distributed subgradient projection algorithm; the weight, which is optimized by a semidefinite programming problem, accelerates the convergence rate of the proposed method.
ER -