Abstract
The growing prevalence of obesity and related health problems warrants immediate need for effective weight control interventions. Quantitative energy balance models serve as powerful tools to assist in these interventions, as a result of their ability to accurately predict individual weight change based on reliable measurements of energy intake and energy expenditure. However, the data collected in most existing weight interventions is self-monitored; these measurements often have significant noise or experience losses resulting from participant non-adherence, which in turn, limits accurate model estimation. To address this issue, we develop a Kalman filter-based estimation algorithm for a practical scenario where on-line state estimation for weight, or energy intake/expenditure is still possible despite correlated partial data losses. To account for non-linearities in the models, an algorithm based on extended Kalman filtering is also developed for sequential state estimation in the presence of missing data. Simulation studies are presented to illustrate the performance of the algorithms and the potential benefits of these techniques in real-life interventions.
Original language | English (US) |
---|---|
Pages (from-to) | 13532-13537 |
Number of pages | 6 |
Journal | 20th IFAC World Congress |
Volume | 50 |
Issue number | 1 |
DOIs | |
State | Published - Jul 2017 |
All Science Journal Classification (ASJC) codes
- Control and Systems Engineering
Access to Document
Other files and links
Fingerprint
Dive into the research topics of 'State Estimation Under Correlated Partial Measurement Losses: Implications for Weight Control Interventions'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver
}
In: 20th IFAC World Congress, Vol. 50, No. 1, 07.2017, p. 13532-13537.
Research output: Contribution to journal › Article › peer-review
TY - JOUR
T1 - State Estimation Under Correlated Partial Measurement Losses
T2 - Implications for Weight Control Interventions
AU - Guo, Penghong
AU - Rivera, Daniel E.
AU - Savage, Jennifer S.
AU - Downs, Danielle S.
N1 - Funding Information: ∗∗∗ School for Engineering of Matter, Transport, and Energy, Arizona School for Engineering of Matter, Transport, and Energy, Arizona StpaetnegUhonnivge.grusiot@y,aTsue.medpue,, dAaZni8e5l.2r8iv1erUaS@Aa.su(.ee-dmua)il: State University, Tempe, AZ 85281 USA. (e-mail: CenterpefonrghCohnigld.ghuooo@d aOsube.esdituy, Rdeasneiaerl.crhivaenrad@tahseuD.eedpua)rtment of ∗∗∗∗CenterpefornghCohingld.guhoo@odasOube.esdityu,Rdanieseaelr.crhivandera@theasuD.eedpua)rtment of N∗u∗tCrietniotneralfoSrciCenhciledsh,oPodenOnbSestaittye RUensievaerrcshitya,ndUnthiveeDrseitpyarPtmarekn,tPoAf, Nutritional Sciences, Penn State University, University Park, PA, Nutritional Sciences, Penn State University, University Park, PA, Exercise PsycholUogSyAL. a(beomraatiol:ryjf,s1D9e5p@arptsmu.eendtu)of Kinesiology, Penn ∗∗∗∗∗∗Exercise PsycholUogSyAL.a(ebomaratil:oryjfs1, D9e5p@partmsu.entedu)of Kinesiology, Penn ∗S∗t∗aEtexeUrnciisveerPsistyyc,hUolnoigvyerLsaitbyorPaatorrky,,PDAe,pUarStAm.en(etmofaiKl:dinseds1i1o@lopgsyu,.Pedeun)n State University, University Park, PA, USA. (email:dsd11@psu.edu) State University, University Park, PA, USA. (email:dsd11@psu.edu) Abstract: The growing prevalence of obesiffifl and relaffied healffih problems warranffis immediaffie Abstract: The growing prevalence of obesiffifl and relaffied healffih problems warranffis immediaffie Abstract: The growing prevalence of obesiffifl and relaffied healffih problems warranffis immediaffie naseepdowfoerrfuelffeffcioffoiivlse ffiwoeaigshsiffsi fficoinnffirffiohlesienffiienrffvieernvfefiinoffniios.nsQ, uaasnaffiiffiraeffsiiuvleffi eonfefrfihgfelirbaablainlicffifel fmfiooadceclsursaeffirevlfel as powerful ffiools ffio assisffi in ffihese inffiervenffiions, as a resulffi of ffiheir abiliffifl ffio accuraffielfl as powerful ffiools ffio assisffi in ffihese inffiervenffiions, as a resulffi of ffiheir abiliffifl ffio accuraffielfl pexrepdeincdffiififinudreiv. iHdouwalevweeri,gffhihffiecdhaaffniagceobllaecseffieddoin mreolisaffbi eleximsffiienagsuwreemighenffiffiisnffoiefrveennefrfiigofnl sinisffiaskeelf-amnodneiffnioerregdf∆l expendiffiure. However, ffihe daffia collecffied in mosffi exisffiing weighffi inffiervenffiions is self-moniffiored∆ expendiffiure. However, ffihe daffia collecffied in mosffi exisffiing weighffi inffiervenffiions is self-moniffiored∆ ffihese measuremenffis offfien have significanffi noise or experience losses resulffiing from parffiicipanffi naoKna-almdhaenrfeinlffcieer,-bwahsiecdheinsffififimuranff,iiolinmaiflfigsoarcicffihumrafffieormaopdrealceffsiifcfiiaml ascffieionna.rTioowahdedrresosnf-filhiniseisssffiuafefie, weseffiidmevaeffiiloonp a Kalman filffier-based esffiimaffiion algoriffihm for a pracffiical scenario where on-line sffiaffie esffiimaffiion a Kalman filffier-based esffiimaffiion algoriffihm for a pracffiical scenario where on-line sffiaffie esffiimaffiion for weighffi, or energfl inffiake/expendiffiure is sffiill possible despiffie correlaffied parffiial daffia losses. Tisoaalscocoduenvffielfoopr endonfo-lrinseaqruifefiinefsfiiainl sffifhfiaeffiemeosdffieimls,afafiinonalignorffiihffihemprbeasesendceonofemxffiiesnsidnegd dKaaffialm. aSnimfiulflfiaerffiiinong To accounffi for non-lineariffiies in ffihe models, an algoriffihm based on exffiended Kalman filffiering issffiuadlsieosdaerveeplorpeseednfffieodr fsfioeqiulleunsffffiiriaalffiesfffifaihffeiepeesrffifiomrmaffiaionnceinofffifhfiheeparlegsoerniffcihemosfamnidssffiihnegpdoafffifeian.ffiiSailmbuelnaeffifiioffnis sffiudies are presenffied ffio illusffiraffie ffihe performance of ffihe algoriffihms and ffihe poffienffiial benefiffis sffiudies are presenffied ffio illusffiraffie ffihe performance of ffihe algoriffihms and ffihe poffienffiial benefiffis of ffihese ffiechniques in real-life inffiervenffiions. © 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Keywords: Esffiimaffiion∆ Filffiering∆ Kalman filffier∆ Exffiended Kalman filffier∆ Inffiermiffiffienffi Keywords: Esffiimaffiion∆ Filffiering∆ Kalman filffier∆ Exffiended Kalman filffier∆ Inffiermiffiffienffi Keywords: Esffiimaffiion∆ Filffiering∆ Kalman filffier∆ Exffiended Kalman filffier∆ Inffiermiffiffienffi measuremenffis∆ Mulffiiple missing measuremenffis∆ Obesiffifl∆ Inffiervenffiions. measuremenffis∆ Mulffiiple missing measuremenffis∆ Obesiffifl∆ Inffiervenffiions. 1. INTRODUCTION weighffi inffiervenffiion efforffis [Sffieinberg effi al. (2013)∆ Vesco 1. INTRODUCTION weighffi inffiervenffiion efforffis [Sffieinberg effi al. (2013)∆ Vesco 1. INTRODUCTION weighffi inffiervenffiion efforffis [Sffieinberg effi al. (2013)∆ Vesco Obesiffiflhasbecomeaworldwidehealffihconcerndueffioiffis pfafiraffili.ci(p2a0n1ffi6s)’],dwieeffiaigrhflffiicnoffinafkfireolainsdmcaanloargiecdebxfplemndoinffiiuffiroerining Obesiffifl has become a worldwide healffih concern due ffio iffis parffiicipanffis’ dieffiarfl inffiake and caloric expendiffiure in Obesiffifl has become a worldwide healffih concern due ffio iffis parffiicipanffis’ dieffiarfl inffiake and caloric expendiffiure in hAicgchorpdrienvgalfefioncfefihaenNdarffeiiloanffiaeldHadeavleffrihseahnedalNffihufcfiroifnfiisoenquEexnacmes-. adhdflisfifiicoiannffsiocffaihneirexwaemiginhfeficihfapnagreffi.icBipaasnefdfisomneffeihffiesffiehemireacsaulorerisc, According ffio ffihe Naffiional Healffih and Nuffiriffiion Exam-phflsicians can examine if parffiicipanffis meeffi ffiheir caloric According ffio ffihe Naffiional Healffih and Nuffiriffiion Exam-phflsicians can examine if parffiicipanffis meeffi ffiheir caloric ipnraefvfiiaolnenSceurovfebflei(nNgHoAveNrEwSei)ghcfofi nodr uocbffeiesde (inOW20/1O1-B2∆01d2e,finffiehde poroavlsi,desdo. Tffihoabffi efffuiffierfrfihuenr dheresaffliaffihndchouownseinlidnigvidaudavliscemcaainffiabine prevalence of being overweighffi or obese 2(OW/OB∆ defined provided. To beffiffier undersffiand how individuals mainffiain prevalence of being overweighffi or obese 2(OW/OB∆ defined provided. To beffiffier undersffiand how individuals mainffiain aasdualfbfisodinfl fmfihaessUiSn,deinxc[lBudMinI]g≥342.59%kg/omf 2a)diuslff6is8.b5e%inagmcoonng-orcluorsaeffiewlfeligphrffei,ddicfflfinawmeiigchfefincehrgafnlgbealhaanvcee bmeoendelwseflfilhaeffsifficaabn-adulffis in ffihe US, including 34.9%2 of adulffis being con-accuraffielfl predicffi weighffi change have been well esffiab-adulffis in ffihe US, including 34.9%2 of adulffis being con-accuraffielfl predicffi weighffi change have been well esffiab-sidered as obese (BMI ≥ 30 kg/m22) [Ogden effi al. (2014)]. lished [Thomas effi al. (2009)∆ Guo effi al. (2016)], which sidered as obese (BMI ≥ 30 kg/m ) [Ogden effi al. (2014)]. lished [Thomas effi al. (2009)∆ Guo effi al. (2016)], which High BMI is significanffilfl associaffied wiffih increased risks of if used as ffiools ffio assisffi inffiervenffiions, will help phflsicians caidrdiffiiioevsa[sBcauslaffiiredniseeffiaasle.s(,2d0i1a4b)e]f.fiePs,aarenndffiaolffihoebrescilfifinflimcaalfcloamffeocrf-fi anifdfihfobseffifefiffrierpaaffsiiseensffsi iandgheffirhencoeuffffiicoodmieefsfi oorf ewxeeirgchisffei rpelgaunlsa.ffiion bidiffiies [Basffiien effi al. (2014)]. Parenffial obesiffifl mafl affecffi and fosffier paffiienffi adherence ffio dieffi or exercise plans. bidiffiies [Basffiien effi al. (2014)]. Parenffial obesiffifl mafl affecffi Ancdcufroasffiermpoadffieielnpffireaddihceffirioenncreqffiouidreiesffireolriaebxleermciseeaspulraenms.enffis ffihe offspring obesiffifl ffihrough herediffifl [Wu and Suzuki Accuraffiemodelpredicffiionrequiresreliablemeasuremenffis ffihe offspring obesiffifl ffihrough herediffifl [Wu and Suzuki Afccffiuhreaffkieemfl oddeeffilerpmreidniacnffiiffiosnorfeqeunierregsflrebliaalbalnecme e[aHsaulrlemeffei nafflis. (w2006)2e0i0g6h)ffi]]..coThTnhffirereorlefefioronrffie,ee,rvffiffiehhnerefrfiieeonexesxiiffsffisioffisspraaevgrgerneaffiefafifafinnndeedeerdedffuorocreeffecffieoffbeecsffiiiffvviflee. ofccffiurahe ffiekemfloddeeffilerpremidicnanffiioffisnofrequireenergflsrebliaalbleancmee[aHalsurel meeffinalffis. (2006)]. Therefore, ffihere exisffis a greaffi need for effecffiive o2f0f1fih2e)].kIenflredaelf-fielirfmewineainghffisffiionfffieernveerngffifilonbsa,lhaonwcev[eHr,aflfilheefdfiaaffila. weighffi conffirol inffiervenffiions ffio prevenffi and reduce obesiffifl. of(20ffi1he2)].kIenflredaeffil-erlifemiwneaignhffisffi iofnffieenrverengflffiionbsal, hancoweev[eHalr, ffihel effidaalffia. Bodfl weighffi change resulffis from an imbalance beffiween is usuallfl self-reporffied or self-moniffiored bfl free-living par-Bodfl weighffi change resulffis from an imbalance beffiween (2is0u1su2)].allflInserelf-arel-lifeporffiewedigohrffisienlf-ffiemrvoeniffionffiiorensd,bhfolwfreevee-r,livingffihe dapaffiar- Bnoedrfglflweinigffiahkffiechaanndgeenreesruglfflfisefxrpomendainffiuirme.baWlaenigcehffibeloffiwsseeins ifisicuipsuaanlfflisflvsiealfe-rleecpffoirroffnieidcodresveilcfe-sm,ownhifificohrepdrbofdlufrcees-lsiivginnigficpaanrf-fi Bodfl weighffi change resulffis from an imbalance beffiween is usuallfl self-reporffied or self-moniffiored bfl free-living par-encheriegvfledinwffiahkeen ffaionffidal energfl inexppueffinidsilffeiusrsef.fihaWnefifigohffiaffil leonsesrgifsl ffiiociispeaninffisffivhieadelaeffciaffircoonllieccdffiieovnic,easl,ownghicwhiffpihromdiuscsiensgsidganffiiaficdauneffi achieved when ffioffial energfl inpuffi is less ffihan ffioffial energfl noise in ffihe daffia collecffiion, along wiffih missing daffia due achieved when ffioffial energfl inpuffi is less ffihan ffioffial energfl noise in ffihe daffia collecffiion, along wiffih missing daffia due ★ouffipuffi, and vice versa. Hence, in mosffi of ffihe exisffiing ffio forgeffifulness or lack of parffiicipanffi adherence ffio inffier-ouSffiuppupffoi,rtafnodr thviiscewovrkerhsaas. bHeeenncper,oviindedmboysfftihoefNfafihtieoneaxl iHsfefiianrtg, ffio forgeffifulness or lack of parffiicipanffi adherence ffio inffier-★★ Support for this work has been provided by the National Heart, imenpffliiaounssi.bIlfeadumeeaffiosusriegmniefincffai nisfficeornrosird,eirffiehdaassffipohbflesidolioscgaicradleldfl LuSnugp, paonrdt BfolroothdisInwstoirtkuthea(sNbHeLenBIp)rotvhirdoeudghbygrtahnet NRa0t1ioHnaLl11H9e2a4r5t,. venffiions. If a measuremenffi is considered as phflsiologicallfl Support for this work has been provided by the National Heart, imorpalassuessisbmleednuffi,erffeiosuslifgfiinigficinanmffi oerrerodr,afififafi hmaisssffiiongbneedssis.cTarhdeesde Tuhnego,painndionBsloeoxdprIensssteidtuitne t(hNisHaLrBtiIc)letharroeutghhegaruatnhtoRrs0’1owHnL1a1n9d24d5o. implausible due ffio significanffi error, iffi has ffio be discarded Lung, and Blood Institute (NHLBI) through grant R01 HL119245. The opinions expressed in this article are the authors’ own and do fiomr iaffisasffieiossnms einnffffii,hreemsuelaffiisnugreidn dmafofiaremdigahffiaffi cmreisasffiienganeissss.uTe ihfeasne The opinions expressed in this article are the authors’ own and do limiffiaffiions in ffihe measured daffia mighffi creaffie an issue if an not necessarily reflect the views of NHLBI. limiffiaffiions in ffihe measured daffia mighffi creaffie an issue if an not necessarily reflect the views of NHLBI. 2405-8963 © 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Copyright © 2017 IFAC 14074 Peer review under responsibility of International Federation of Automatic Control. Copyright © 2017 IFAC 14074 10.1016/j.ifacol.2017.08.2347 Publisher Copyright: © 2017
PY - 2017/7
Y1 - 2017/7
N2 - The growing prevalence of obesity and related health problems warrants immediate need for effective weight control interventions. Quantitative energy balance models serve as powerful tools to assist in these interventions, as a result of their ability to accurately predict individual weight change based on reliable measurements of energy intake and energy expenditure. However, the data collected in most existing weight interventions is self-monitored; these measurements often have significant noise or experience losses resulting from participant non-adherence, which in turn, limits accurate model estimation. To address this issue, we develop a Kalman filter-based estimation algorithm for a practical scenario where on-line state estimation for weight, or energy intake/expenditure is still possible despite correlated partial data losses. To account for non-linearities in the models, an algorithm based on extended Kalman filtering is also developed for sequential state estimation in the presence of missing data. Simulation studies are presented to illustrate the performance of the algorithms and the potential benefits of these techniques in real-life interventions.
AB - The growing prevalence of obesity and related health problems warrants immediate need for effective weight control interventions. Quantitative energy balance models serve as powerful tools to assist in these interventions, as a result of their ability to accurately predict individual weight change based on reliable measurements of energy intake and energy expenditure. However, the data collected in most existing weight interventions is self-monitored; these measurements often have significant noise or experience losses resulting from participant non-adherence, which in turn, limits accurate model estimation. To address this issue, we develop a Kalman filter-based estimation algorithm for a practical scenario where on-line state estimation for weight, or energy intake/expenditure is still possible despite correlated partial data losses. To account for non-linearities in the models, an algorithm based on extended Kalman filtering is also developed for sequential state estimation in the presence of missing data. Simulation studies are presented to illustrate the performance of the algorithms and the potential benefits of these techniques in real-life interventions.
UR - http://www.scopus.com/inward/record.url?scp=85044315619&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85044315619&partnerID=8YFLogxK
U2 - 10.1016/j.ifacol.2017.08.2347
DO - 10.1016/j.ifacol.2017.08.2347
M3 - Article
C2 - 29242854
AN - SCOPUS:85044315619
SN - 2405-8963
VL - 50
SP - 13532
EP - 13537
JO - 20th IFAC World Congress
JF - 20th IFAC World Congress
IS - 1
ER -