An enhanced bootstrap filtering method for non-intrusive load monitoring

Weicong Kong; Zhaoyang Dong; Yan Xu; David Hill

doi:10.1109/PESGM.2016.7741487

An enhanced bootstrap filtering method for non-intrusive load monitoring

Weicong Kong
, Zhaoyang Dong
, Yan Xu
, David Hill

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

12 Citations (Scopus)

Abstract

Non-intrusive load monitoring (NILM) aims to estimate the power or energy consumption for a collection of different appliances connected to a single power inlet, with only aggregated power profile being known. Such estimation is highly valuable for many potential applications in future smart grid. This paper proposes a bootstrap filtering based solver to work with smart meter data to solve the NILM problem. The weight updating process is the focus in our solver. It is shown that some slight change applied to the weight updating process can significantly enhance the computation efficiency. Evaluation of our approach is given based on results from a popular public dataset. It is also demonstrated that some common evaluation metrics are more appropriate than others. The extra merit to attacking NILM using bootstrap filtering is also illustrated. © 2016 IEEE.

Original language	English
Title of host publication	2016 IEEE Power and Energy Society General Meeting, PESGM 2016
Publisher	IEEE Computer Society
Volume	2016-November
ISBN (Print)	9781509041688
DOIs	https://doi.org/10.1109/PESGM.2016.7741487
Publication status	Published - 10 Nov 2016
Externally published	Yes
Event	2016 IEEE Power and Energy Society General Meeting, PESGM 2016 - Boston, United States Duration: 17 Jul 2016 → 21 Jul 2016

Publication series

Name	IEEE Power and Energy Society General Meeting
Volume	2016-November
ISSN (Print)	1944-9925
ISSN (Electronic)	1944-9933

Conference

Conference	2016 IEEE Power and Energy Society General Meeting, PESGM 2016
Place	United States
City	Boston
Period	17/07/16 → 21/07/16

Bibliographical note

Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].

Funding

This research was supported by Australian Endeavour Postgraduate Scholarship, and the funding from the Faculty of Engineering & Information Technologies, The University of Sydney, under the Faculty Research Cluster Program. This research was also supported by the Research Grants Council Theme-based Scheme through Project No. T23- 701/14N, Hong Kong, and supported by China Southern Power Grid Company through project WYKJ00000027

Research Keywords

Bootstrap filter
Factorial hidden Markov model
Hidden Markov model
Non-intrusive load monitoring

RGC Funding Information

RGC-funded

Access Output

10.1109/PESGM.2016.7741487

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{3dd3ab34714245208d1ac766ed3933e4,

title = "An enhanced bootstrap filtering method for non-intrusive load monitoring",

abstract = "Non-intrusive load monitoring (NILM) aims to estimate the power or energy consumption for a collection of different appliances connected to a single power inlet, with only aggregated power profile being known. Such estimation is highly valuable for many potential applications in future smart grid. This paper proposes a bootstrap filtering based solver to work with smart meter data to solve the NILM problem. The weight updating process is the focus in our solver. It is shown that some slight change applied to the weight updating process can significantly enhance the computation efficiency. Evaluation of our approach is given based on results from a popular public dataset. It is also demonstrated that some common evaluation metrics are more appropriate than others. The extra merit to attacking NILM using bootstrap filtering is also illustrated. {\textcopyright} 2016 IEEE.",

keywords = "Bootstrap filter, Factorial hidden Markov model, Hidden Markov model, Non-intrusive load monitoring",

author = "Weicong Kong and Zhaoyang Dong and Yan Xu and David Hill",

note = "Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].; 2016 IEEE Power and Energy Society General Meeting, PESGM 2016 ; Conference date: 17-07-2016 Through 21-07-2016",

year = "2016",

month = nov,

day = "10",

doi = "10.1109/PESGM.2016.7741487",

language = "English",

isbn = "9781509041688",

volume = "2016-November",

series = "IEEE Power and Energy Society General Meeting",

publisher = "IEEE Computer Society",

booktitle = "2016 IEEE Power and Energy Society General Meeting, PESGM 2016",

address = "United States",

}

Kong, W, Dong, Z, Xu, Y & Hill, D 2016, An enhanced bootstrap filtering method for non-intrusive load monitoring. in 2016 IEEE Power and Energy Society General Meeting, PESGM 2016. vol. 2016-November, 7741487, IEEE Power and Energy Society General Meeting, vol. 2016-November, IEEE Computer Society, 2016 IEEE Power and Energy Society General Meeting, PESGM 2016, Boston, United States, 17/07/16. https://doi.org/10.1109/PESGM.2016.7741487

An enhanced bootstrap filtering method for non-intrusive load monitoring. / Kong, Weicong; Dong, Zhaoyang; Xu, Yan et al.
2016 IEEE Power and Energy Society General Meeting, PESGM 2016. Vol. 2016-November IEEE Computer Society, 2016. 7741487 (IEEE Power and Energy Society General Meeting; Vol. 2016-November).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - An enhanced bootstrap filtering method for non-intrusive load monitoring

AU - Kong, Weicong

AU - Dong, Zhaoyang

AU - Xu, Yan

AU - Hill, David

N1 - Publication details (e.g. title, author(s), publication statuses and dates) are captured on an “AS IS” and “AS AVAILABLE” basis at the time of record harvesting from the data source. Suggestions for further amendments or supplementary information can be sent to [email protected].

PY - 2016/11/10

Y1 - 2016/11/10

N2 - Non-intrusive load monitoring (NILM) aims to estimate the power or energy consumption for a collection of different appliances connected to a single power inlet, with only aggregated power profile being known. Such estimation is highly valuable for many potential applications in future smart grid. This paper proposes a bootstrap filtering based solver to work with smart meter data to solve the NILM problem. The weight updating process is the focus in our solver. It is shown that some slight change applied to the weight updating process can significantly enhance the computation efficiency. Evaluation of our approach is given based on results from a popular public dataset. It is also demonstrated that some common evaluation metrics are more appropriate than others. The extra merit to attacking NILM using bootstrap filtering is also illustrated. © 2016 IEEE.

AB - Non-intrusive load monitoring (NILM) aims to estimate the power or energy consumption for a collection of different appliances connected to a single power inlet, with only aggregated power profile being known. Such estimation is highly valuable for many potential applications in future smart grid. This paper proposes a bootstrap filtering based solver to work with smart meter data to solve the NILM problem. The weight updating process is the focus in our solver. It is shown that some slight change applied to the weight updating process can significantly enhance the computation efficiency. Evaluation of our approach is given based on results from a popular public dataset. It is also demonstrated that some common evaluation metrics are more appropriate than others. The extra merit to attacking NILM using bootstrap filtering is also illustrated. © 2016 IEEE.

KW - Bootstrap filter

KW - Factorial hidden Markov model

KW - Hidden Markov model

KW - Non-intrusive load monitoring

UR - https://www.scopus.com/pages/publications/85001820397

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85001820397&origin=recordpage

U2 - 10.1109/PESGM.2016.7741487

DO - 10.1109/PESGM.2016.7741487

M3 - RGC 32 - Refereed conference paper (with host publication)

SN - 9781509041688

VL - 2016-November

T3 - IEEE Power and Energy Society General Meeting

BT - 2016 IEEE Power and Energy Society General Meeting, PESGM 2016

PB - IEEE Computer Society

T2 - 2016 IEEE Power and Energy Society General Meeting, PESGM 2016

Y2 - 17 July 2016 through 21 July 2016

ER -

An enhanced bootstrap filtering method for non-intrusive load monitoring

Abstract

Publication series

Conference

Bibliographical note

Funding

Research Keywords

RGC Funding Information

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this