Performance of certain decentralized distributed change detection procedures

Alexander G. Tartakovsky, Hongjoong Kim

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

18 Citations (Scopus)

Abstract

We compare several decentralized change-point detection procedures for multisensor distributed systems when the information available for decision-making is distributed across a set of sensors. Asymptotically optimal procedures for two scenarios are presented In the first scenario, the sensors send quantized versions of their observations to a fusion center where change detection is performed based on all the sensor messages. If, in particular, the quantizers are binary, then the proposed binary CUSUM detection test is optimal in the class of tests with binary quantized data. In the second scenario, the sensors perform local change detection using the CUSUM procedures and send their final decisions to the fusion center for combining The decision in favor of the change occurrence is made whenever CUSUM statistics at all sensors exceed thresholds. The latter decentralized procedure has the same first order asymptotic (as the false alarm rate is low) minimax operating characteristics as the globally optimal centralized detection procedure that has access to all the sensor observations. However, the presented Monte Carlo experiments for the Poisson example show that despite the fact that the procedure with local decisions is globally asymptotically optimal for a low false alarm rate, it performs worse than the procedure with binary quantization unless the false alarm rate is extremely low. In addition, two voting-type local decision based detection procedures are proposed and evaluated Applications to network security (rapid detection of computer intrusions) are discussed.

Original languageEnglish
Title of host publication2006 9th International Conference on Information Fusion, FUSION
DOIs
Publication statusPublished - 2006 Dec 1
Event2006 9th International Conference on Information Fusion, FUSION - Florence, Italy
Duration: 2006 Jul 102006 Jul 13

Other

Other2006 9th International Conference on Information Fusion, FUSION
CountryItaly
CityFlorence
Period06/7/1006/7/13

Fingerprint

Sensors
Fusion reactions
Network security
Decision making
Statistics
Experiments

Keywords

  • Change-point sequential detection
  • CUSUM test
  • Distributed multisensor decisions
  • Intrusion detection
  • Local decisions
  • Optimal fusion
  • Quickest detection

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems

Cite this

Tartakovsky, A. G., & Kim, H. (2006). Performance of certain decentralized distributed change detection procedures. In 2006 9th International Conference on Information Fusion, FUSION [4086098] https://doi.org/10.1109/ICIF.2006.301812

Performance of certain decentralized distributed change detection procedures. / Tartakovsky, Alexander G.; Kim, Hongjoong.

2006 9th International Conference on Information Fusion, FUSION. 2006. 4086098.

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

Tartakovsky, AG & Kim, H 2006, Performance of certain decentralized distributed change detection procedures. in 2006 9th International Conference on Information Fusion, FUSION., 4086098, 2006 9th International Conference on Information Fusion, FUSION, Florence, Italy, 06/7/10. https://doi.org/10.1109/ICIF.2006.301812
Tartakovsky AG, Kim H. Performance of certain decentralized distributed change detection procedures. In 2006 9th International Conference on Information Fusion, FUSION. 2006. 4086098 https://doi.org/10.1109/ICIF.2006.301812
Tartakovsky, Alexander G. ; Kim, Hongjoong. / Performance of certain decentralized distributed change detection procedures. 2006 9th International Conference on Information Fusion, FUSION. 2006.
@inproceedings{11015d451211454e981395c31a71394f,
title = "Performance of certain decentralized distributed change detection procedures",
abstract = "We compare several decentralized change-point detection procedures for multisensor distributed systems when the information available for decision-making is distributed across a set of sensors. Asymptotically optimal procedures for two scenarios are presented In the first scenario, the sensors send quantized versions of their observations to a fusion center where change detection is performed based on all the sensor messages. If, in particular, the quantizers are binary, then the proposed binary CUSUM detection test is optimal in the class of tests with binary quantized data. In the second scenario, the sensors perform local change detection using the CUSUM procedures and send their final decisions to the fusion center for combining The decision in favor of the change occurrence is made whenever CUSUM statistics at all sensors exceed thresholds. The latter decentralized procedure has the same first order asymptotic (as the false alarm rate is low) minimax operating characteristics as the globally optimal centralized detection procedure that has access to all the sensor observations. However, the presented Monte Carlo experiments for the Poisson example show that despite the fact that the procedure with local decisions is globally asymptotically optimal for a low false alarm rate, it performs worse than the procedure with binary quantization unless the false alarm rate is extremely low. In addition, two voting-type local decision based detection procedures are proposed and evaluated Applications to network security (rapid detection of computer intrusions) are discussed.",
keywords = "Change-point sequential detection, CUSUM test, Distributed multisensor decisions, Intrusion detection, Local decisions, Optimal fusion, Quickest detection",
author = "Tartakovsky, {Alexander G.} and Hongjoong Kim",
year = "2006",
month = "12",
day = "1",
doi = "10.1109/ICIF.2006.301812",
language = "English",
isbn = "1424409535",
booktitle = "2006 9th International Conference on Information Fusion, FUSION",

}

TY - GEN

T1 - Performance of certain decentralized distributed change detection procedures

AU - Tartakovsky, Alexander G.

AU - Kim, Hongjoong

PY - 2006/12/1

Y1 - 2006/12/1

N2 - We compare several decentralized change-point detection procedures for multisensor distributed systems when the information available for decision-making is distributed across a set of sensors. Asymptotically optimal procedures for two scenarios are presented In the first scenario, the sensors send quantized versions of their observations to a fusion center where change detection is performed based on all the sensor messages. If, in particular, the quantizers are binary, then the proposed binary CUSUM detection test is optimal in the class of tests with binary quantized data. In the second scenario, the sensors perform local change detection using the CUSUM procedures and send their final decisions to the fusion center for combining The decision in favor of the change occurrence is made whenever CUSUM statistics at all sensors exceed thresholds. The latter decentralized procedure has the same first order asymptotic (as the false alarm rate is low) minimax operating characteristics as the globally optimal centralized detection procedure that has access to all the sensor observations. However, the presented Monte Carlo experiments for the Poisson example show that despite the fact that the procedure with local decisions is globally asymptotically optimal for a low false alarm rate, it performs worse than the procedure with binary quantization unless the false alarm rate is extremely low. In addition, two voting-type local decision based detection procedures are proposed and evaluated Applications to network security (rapid detection of computer intrusions) are discussed.

AB - We compare several decentralized change-point detection procedures for multisensor distributed systems when the information available for decision-making is distributed across a set of sensors. Asymptotically optimal procedures for two scenarios are presented In the first scenario, the sensors send quantized versions of their observations to a fusion center where change detection is performed based on all the sensor messages. If, in particular, the quantizers are binary, then the proposed binary CUSUM detection test is optimal in the class of tests with binary quantized data. In the second scenario, the sensors perform local change detection using the CUSUM procedures and send their final decisions to the fusion center for combining The decision in favor of the change occurrence is made whenever CUSUM statistics at all sensors exceed thresholds. The latter decentralized procedure has the same first order asymptotic (as the false alarm rate is low) minimax operating characteristics as the globally optimal centralized detection procedure that has access to all the sensor observations. However, the presented Monte Carlo experiments for the Poisson example show that despite the fact that the procedure with local decisions is globally asymptotically optimal for a low false alarm rate, it performs worse than the procedure with binary quantization unless the false alarm rate is extremely low. In addition, two voting-type local decision based detection procedures are proposed and evaluated Applications to network security (rapid detection of computer intrusions) are discussed.

KW - Change-point sequential detection

KW - CUSUM test

KW - Distributed multisensor decisions

KW - Intrusion detection

KW - Local decisions

KW - Optimal fusion

KW - Quickest detection

UR - http://www.scopus.com/inward/record.url?scp=50149104586&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=50149104586&partnerID=8YFLogxK

U2 - 10.1109/ICIF.2006.301812

DO - 10.1109/ICIF.2006.301812

M3 - Conference contribution

AN - SCOPUS:50149104586

SN - 1424409535

SN - 9781424409532

BT - 2006 9th International Conference on Information Fusion, FUSION

ER -