Low complexity source localization algorithms in sensor networks

Junichi Shirahama, Tomoaki Ohtsuki, Toshinobu Kaneko

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

One typical use of sensor networks is monitoring targets. The sensor networks classify, detect, locate, and track targets. The ML (Maximum likelihood) estimation algorithm is one of the estimation algorithms of target location. The ML estimation algorithm has high accuracy to estimate target location. However, the calculation amount of the ML estimation algorithm is large. The EM (Expectation Maximization) algorithm is proposed to reduce the complexity of the ML estimation algorithm. However, the EM algorithm sometimes traps into local minimum. These conventional algorithms to estimate target location use all the sensors' receiving signals. The transmission signal from the target is attenuated with distance. In particular, the effects of noise on the received signals of the sensors far apart from the target are large. The received signals thus do not help a lot to improve the estimation accuracy. In this paper, we propose the new algorithm to estimate a target location with a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. Moreover, we propose the low complexity source localization method, where we use only the sensors' information with receiving energy higher than threshold. From the simulation results, we show that the proposed algorithm has a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. We also show that proposed method can reduce the calculation amount while keeping the estimation accuracy by setting threshold appropriately in the ML estimation algorithm and the proposed algorithm.

Original languageEnglish
Title of host publicationPE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks
EditorsM. Ould-Khaoua, M. Takai
Pages130-136
Number of pages7
DOIs
Publication statusPublished - 2005
EventPE-WASUN'05 - Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks - Montreal, QB, Canada
Duration: 2005 Oct 102005 Oct 13

Other

OtherPE-WASUN'05 - Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks
CountryCanada
CityMontreal, QB
Period05/10/1005/10/13

Fingerprint

Sensor networks
Maximum likelihood estimation
Sensors

Keywords

  • EM Algorithm
  • ML Estimation Algorithm
  • Sensor networks
  • Source localization

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Shirahama, J., Ohtsuki, T., & Kaneko, T. (2005). Low complexity source localization algorithms in sensor networks. In M. Ould-Khaoua, & M. Takai (Eds.), PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks (pp. 130-136) https://doi.org/10.1145/1089803.1089977

Low complexity source localization algorithms in sensor networks. / Shirahama, Junichi; Ohtsuki, Tomoaki; Kaneko, Toshinobu.

PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks. ed. / M. Ould-Khaoua; M. Takai. 2005. p. 130-136.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Shirahama, J, Ohtsuki, T & Kaneko, T 2005, Low complexity source localization algorithms in sensor networks. in M Ould-Khaoua & M Takai (eds), PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks. pp. 130-136, PE-WASUN'05 - Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks, Montreal, QB, Canada, 05/10/10. https://doi.org/10.1145/1089803.1089977
Shirahama J, Ohtsuki T, Kaneko T. Low complexity source localization algorithms in sensor networks. In Ould-Khaoua M, Takai M, editors, PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks. 2005. p. 130-136 https://doi.org/10.1145/1089803.1089977
Shirahama, Junichi ; Ohtsuki, Tomoaki ; Kaneko, Toshinobu. / Low complexity source localization algorithms in sensor networks. PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks. editor / M. Ould-Khaoua ; M. Takai. 2005. pp. 130-136
@inproceedings{8263fde031e34fa3b713a946fbc554c0,
title = "Low complexity source localization algorithms in sensor networks",
abstract = "One typical use of sensor networks is monitoring targets. The sensor networks classify, detect, locate, and track targets. The ML (Maximum likelihood) estimation algorithm is one of the estimation algorithms of target location. The ML estimation algorithm has high accuracy to estimate target location. However, the calculation amount of the ML estimation algorithm is large. The EM (Expectation Maximization) algorithm is proposed to reduce the complexity of the ML estimation algorithm. However, the EM algorithm sometimes traps into local minimum. These conventional algorithms to estimate target location use all the sensors' receiving signals. The transmission signal from the target is attenuated with distance. In particular, the effects of noise on the received signals of the sensors far apart from the target are large. The received signals thus do not help a lot to improve the estimation accuracy. In this paper, we propose the new algorithm to estimate a target location with a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. Moreover, we propose the low complexity source localization method, where we use only the sensors' information with receiving energy higher than threshold. From the simulation results, we show that the proposed algorithm has a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. We also show that proposed method can reduce the calculation amount while keeping the estimation accuracy by setting threshold appropriately in the ML estimation algorithm and the proposed algorithm.",
keywords = "EM Algorithm, ML Estimation Algorithm, Sensor networks, Source localization",
author = "Junichi Shirahama and Tomoaki Ohtsuki and Toshinobu Kaneko",
year = "2005",
doi = "10.1145/1089803.1089977",
language = "English",
isbn = "1595931821",
pages = "130--136",
editor = "M. Ould-Khaoua and M. Takai",
booktitle = "PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks",

}

TY - GEN

T1 - Low complexity source localization algorithms in sensor networks

AU - Shirahama, Junichi

AU - Ohtsuki, Tomoaki

AU - Kaneko, Toshinobu

PY - 2005

Y1 - 2005

N2 - One typical use of sensor networks is monitoring targets. The sensor networks classify, detect, locate, and track targets. The ML (Maximum likelihood) estimation algorithm is one of the estimation algorithms of target location. The ML estimation algorithm has high accuracy to estimate target location. However, the calculation amount of the ML estimation algorithm is large. The EM (Expectation Maximization) algorithm is proposed to reduce the complexity of the ML estimation algorithm. However, the EM algorithm sometimes traps into local minimum. These conventional algorithms to estimate target location use all the sensors' receiving signals. The transmission signal from the target is attenuated with distance. In particular, the effects of noise on the received signals of the sensors far apart from the target are large. The received signals thus do not help a lot to improve the estimation accuracy. In this paper, we propose the new algorithm to estimate a target location with a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. Moreover, we propose the low complexity source localization method, where we use only the sensors' information with receiving energy higher than threshold. From the simulation results, we show that the proposed algorithm has a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. We also show that proposed method can reduce the calculation amount while keeping the estimation accuracy by setting threshold appropriately in the ML estimation algorithm and the proposed algorithm.

AB - One typical use of sensor networks is monitoring targets. The sensor networks classify, detect, locate, and track targets. The ML (Maximum likelihood) estimation algorithm is one of the estimation algorithms of target location. The ML estimation algorithm has high accuracy to estimate target location. However, the calculation amount of the ML estimation algorithm is large. The EM (Expectation Maximization) algorithm is proposed to reduce the complexity of the ML estimation algorithm. However, the EM algorithm sometimes traps into local minimum. These conventional algorithms to estimate target location use all the sensors' receiving signals. The transmission signal from the target is attenuated with distance. In particular, the effects of noise on the received signals of the sensors far apart from the target are large. The received signals thus do not help a lot to improve the estimation accuracy. In this paper, we propose the new algorithm to estimate a target location with a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. Moreover, we propose the low complexity source localization method, where we use only the sensors' information with receiving energy higher than threshold. From the simulation results, we show that the proposed algorithm has a smaller amount of calculation than the ML estimation algorithm and higher estimation accuracy than the EM algorithm. We also show that proposed method can reduce the calculation amount while keeping the estimation accuracy by setting threshold appropriately in the ML estimation algorithm and the proposed algorithm.

KW - EM Algorithm

KW - ML Estimation Algorithm

KW - Sensor networks

KW - Source localization

UR - http://www.scopus.com/inward/record.url?scp=31844441596&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=31844441596&partnerID=8YFLogxK

U2 - 10.1145/1089803.1089977

DO - 10.1145/1089803.1089977

M3 - Conference contribution

AN - SCOPUS:31844441596

SN - 1595931821

SP - 130

EP - 136

BT - PE-WASUN'05 - Proceedings of the Second ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks

A2 - Ould-Khaoua, M.

A2 - Takai, M.

ER -