一种基于无线传感器网络的生理信号采集系统_英文_

Journal of Southeast U niversity (Eng lish Edition) V o.l 26,N o .1,pp .73-77M ar .2010 I SSN 1003 7985Physiological signal acquisition syste mbased on w ireless sensor net worksQ iuW enjiao Zhang Y ongkui(K ey L abo ra t o ry o f Ch ild D eve l opm ent and L earn i ng Sc ience o fM i n istry o f Education ,Southea st U n i v ersity ,N anji ng 210096,Ch i na)Abstr act :B ase d on w ireless sensor net w orks ,a physi o l ogicalsi gnal acquisiti on syste m is pr oposed .The syste m is use d i n classroo m education in order t o understand t he physi o l ogical c hanges i n t he students .I n t he s y ste m,the biolog i cal electrical si gnal relate d to student atte n ti on a nd e m oti on states ca n be m easur ed by electrocar d i ogr aphy signals .The b i oel ectri cal signal is digitali zed at a 200H z s a m p ling rate and is tra n s m itted by t he Z ig B ee protoc o.l S i m u lta neously ,t he B l uet ooth technology is als o e m be dde d i n the nodes so as to m eet the high sa m pli ng rate a nd the high ba nd w i d t h tra n s m issi on .The syste m can i m p l e m ent them on it o ri ng tas k s for 30students ,and t he experi m ental resu lts of usi ng t he syste m i n t he cl assroo m are propose d .Finall y ,t he a pplicati ons of w irel ess se n s or net w orks used in e ducati on is als o d iscussed .K ey w or ds :w ir eless sensor net w or k ;physi o l ogical si gna;l e du cati onRecei ved 2009 07 20.Biographies :Q i u W en ji ao (1969 ),m ale ,graduate ;Zh ang Y on gku i (corres ponding au t ho r),m ale ,do ctor ,ass ociat e pro fess or ,yzhangb @s .cn .Foundati on ite m:The N ati ona lNatural S ci en ce Foundation of Ch i na(No.60775057).C itati on :Q iu W en ji ao,Zh ang Y on gku.i Phy si o l og i ca l signa l acqu isition s y st e m based on w i rel ess sen s or net w ork s[J].J ournal o f SoutheastU niver sit y (E ngli sh Ed iti on ),2010,26(1):73 77.Although the attent i on state and e mot i ona l state of stude n ts dur i ng the lear n i ng process belong to the psyc holog ica l do m a i n,it can be i ndirectly m easured by physi o l og i cal signals ,suc h as e lectr ocard i ography si gnals and pulse signals .Students and teachers m ental states of attention and e mo ti on in the classroo m m ay be chang i ng dur i ng teach i ng progress ,wh ich can physi o l og i cally acti vate the sy m pathetic and parasy mpathetic division o f theautono m ic nervous syste m (AN S )[1 2].T hese m easurable physi o l og i cal signals can be recorded for the analyses of psyc holog ica l ar ousal of social re w ar ds and punish m ents ,such as posit i ve feedbac k by the teacher s praises or higher exa m i nat i on grades ,even the a mount of a scholarship ,al though the arousal of the autono m ic nervous syste m reac t i ons m ay be a co m plex i ndirect relationship to e m o t i on [3 4].Interact i ons bet w een states of the autono m ic nerv ous syste m and cognit i ve perfor m ance have a long trad ition of being a top ic of psycho log ical research .C lassic concepts fro m mo tivat i onal psycho logy have suggested an i nverted u shaped association bet w een unspec ific activat i on and m ental funct i oning [5 6].A ccor d i ng to this ,the best funct i onal condit i ons are expected atm idrange arousa,l and both over ar ousa l and under arousal are acco m panied by declines in perf or m ance .Cardiovasc u lar psychophysiology has also contributed to this li ne of research ,the respective models re l at i ng changes in card i ovascular act i v ity to facilitation,i nh i b it i on of i nfor m ation processi ng [7]or ener get icm ob iliza t i on o f the organis m when faced w ith a situat i on requ iri ngbehav i oral ad j ust m ent [8].H o w ever ,although this certai n ly const itutes a benefic i a l appr oach ,e m p irical work i n th isf i e l d re m a i ns re lati vely sparse [9].Both the sy m pathetic syste m and the parasy mpathetic syste m contri bute to cardiovascular regulat i on [10].Sy mpathetic i n fl uences are trans m itted through efferent fibres to the si nus node ,the m yo cardiu m and the vascularm usc u lature ,and their activati on leads to an i ncrease i n heart rate ,cardiac contract ility and vascular tone .Parasy mpathetic i nfl ue nces arew i de l y ,but not co mpletely ,restri cted to the m odulat i on of heart rate through i nhibiting sinus node act i v ity .I n addition ,the cardiac baroreflex is involved .I n th i s negative feedbac k loop ,c han ges i n the act i vity of the arterial baroreceptors due to fl uctua tions i n b l ood pressure are responded w ith co m pensatory changes i n heart rate and contract ility .A co mplex net work of brai n ste m un its subserve cardi ovascular autono m ic contro,l e .g .,the nucleus of the so litar y tract(NT S),the dorsalm otor nucle us(DMN ),the nucleus a mbiguous(NA )and the rostralventrolateral medulla(RVL M )[11].B ilateral direct and i ndi rect c onnecti ons ex ist bet w een this net work and cortical areas ,which for m an m i portant li nk bet wee n cardi ovascularregulation and cogn ition [12-13].The present study am i s at in vest i gat i ng relationsh i ps a mong features of sy m pathetic ,para sy mpathetic and bar oreflex car d i ovascular control and atten tional perfor mance .To collect these physi ca l and mental states data i n real tm i e and objectivel y ,a hybrid w ireless sensor net w ork is desi gned as the subjects of a teac her and st ude nts are mob ile .For m edical and ho m e usage ,li ce nse free IS M (i ndustr y ,sci ence ,m edical)radio frequency(RF )o f 2 4G H z Z i gBee w irel ess sensor node technology is usef u l [14 18].1 Syste m D escripti o nA s a secondary ai m ,the study i nvestigates i nter i nd i v i d ual differences i n task i nduced cardiovascular m odulat i ons .Porges [19]postulated an associat i on bet w een resti ng cardiac vagal tone and the ex tent o f card i ovascular react i v i ty .Th is is consistent w ith stud i es that have revealed m ore pronounced heart rate responses to var i ous st i m u li i n chil dren and adults w ith higher baseline heart rate variabili ty [19 21].Cardiovascular reacti v ity to cogniti ve de m a nds m ayalso rel ate to task perfor m ance .Duschek [22]found a positive correlat i on bet wee n systo lic and diastolic b l ood pressure in creases duri ng the executi on o f five attenti on tasks a nd the perfor mance on each o f the m.I n i nfants ,greater decreases of RS A duri ng m ental testing are related to higher f unct i onallevels [21].Inter i ndividual differences in cardiovasc ular mod ulation possi bly reflect different degrees of autono m ic adj ustm e nt asw ell asmot i vat i on on a task ,both of wh i ch m ay contri bute to perfor mance .H o w ever ,i nc onsiste nt findi ngs ,.i e .,m issi ng or even i nverse associat i ons bet w een car d i ovascu l ar reactivity a nd mental perfor m a nce ,have also been repor te d [23 24].T hus ,the curre nt state of research does not a llo w defi nite conclusi ons .The hardware of t he w ireless se nsor net works(W SN s)consists of the data co llecting and trans m itt i ng parts which w ill be attached to students and teachers .To re spect t hese stude nts and teachers w illi ngness to try this ne w device ,the nodes can be c onnected by ECG leads ,but a lso can be rejecte d by the connecti on of leads .The syste m is au to configure d i n the classr oo m.The application of W S N s for physi olog ical si gnal acquisition in the cl assroo m is sho wn i n F ig 1.Fig 1 The sm i p l e star conf i gurati on o f w i reless nodes i n class room1 1 D ata collecti ng and trans m itt i ng partsT his part is desi gned to collect and trans m it physiolog i ca ldata fro m t he subjects [3].W e collect t w o kinds of physi o l og i cal signals :electr ocar d i o si gnals and pulse si gnals fro m the subjects in this research .AD620is chosen as the pre a mpli fier for the a m plification of the weak bi oelectrical signals of 10to 20mV .The ECG uses three standar d lm i b leads (left ar m,ri ght ar m,right leg).T he pulse sensor uses PVDF pie zoelectric fil m w ith a te m perature co mpensati on co mpone nt and an i ntegrated si gnal circuit .A fter pre a mplification of about 100tm i es ,the signal is subjected to l o w pass and hi gh pass filters (5to 100H z).T o reduce the 50H z po w er li ne i nterfere nce ,the addit i onal notc h f ilter is designed before the signal is pro m oted to a posi tive DC level of 1 5V.The har dware block diagra m and cir cuit are shown i n F ig 2and F ig 3.F ig 4sho w s the picture of the se nsor node .The a nal og si gnals ar e di gitali zed usi ng a n A t me ga128m icr ocontroller s on c hi p ADC at a 200Hz sa mpli ng ratebecauseFig 2 Phy si o l og i ca l signa l acquisiti on hardw are block d i ag ramF i g 3 C ircuit pri nc i p l e b l o ck diagra m o f sen s o r i nput unitF ig 4 P i c t ure o f t he senso r nodethe frequency range of the physiolog ical si gnal is fro m 5to100H z .U si ng a hard w are SPI i nterface ,t he collected data are trans m itted by the w ireless sensor net wor k w it h the Z i g Bee protocol by a CC2420RF chip .A s t he gate way to a per sonal co mputer (PC ),the sink node co mm un icates w it h the PC thr ough a RS 232i nterface .1 2 Soft w are chart of data co llect i ng and trans m itting parts 1 2 1 Topo l ogy confi gurati on of theWSN T he WSN is expected to realize t he dyna m ic m anage m ent o f nodes ,suc h as the j o i nt and leave of the nodes ,serv ice d iscovery ,and dev ice auto configurat i on .The topology con tr o l can adapti vely organize certa i n numbers o f nodes to a net by certain m echan is m s .W e establish a star net wor k i n the intelli gent classr oo m.The sink node init i alizes the net work ,broadcasts its net work address and sca n c hanne,l and receives responses fr o m the ter m inal nodes ,whil e the ter m i nal nodes atte mpt to seek and joi n the net after i n itialization [4].1 2 2 Task sc he duling m echanis m I n our syste m,w e realize tasks such as data acqu isit i on ,storage ,processi ng and tr ans m ission ,usi ng the sm i ple F I F O (first in first out)task scheduli ng .The data str ucture used74Q i u W en ji ao ,and Zhang Yongku ii n the F I FO is a c irc ular que ue ;eac h ele m ent i n it i s a func t i on po i nter ,po i nti ng to the head address of the releva nt task .A ny task m ust enter t he task queue by a task subm it t i ng f unction and c ha nge the rear poi nter ,wh ile it can only run the task runn i ng functi onwhen t he front poi nter points to its function pointer .The tasks i n the queue w ill be exec uted successi vel y .A task can appear as one kind of operat i on ,wh ic h can only be interrupted by events such as cloc k i nter ruption ,A /D interrupti ons and so on ,but not by any othertasks [14].1 2 3 D esi gn of data packet The space nodes i n the classr oo m can collect vari ous phys i o l og ical data of st ude nts and env ir onm e nta l data .I n our sys te m,we use si x A /D channels f or the co llection of the si g nals of the ECG,2 axis accelerat i on ,pressure ,te mperature and hu m i d ity ,respectively .W e asse mble the gathered data (H ere it i s the pay l oad of the data pac ket)and so me othercontrol i nfor m at i on i nto a data packet [15 16].T he fra m e contr o l fi eld w ith a 2 byte leng t h defi nes the control infor m ati on or fra m e for m at .The data seque nce nu mber i de ntifies the data packets in an orderly fashion .The address conta i ns t he PAN I D ,destinat i on node s address and source node s address .T he payload is the data of si gna ls w e co llect fro m the students and the environment of the classroo m.W e put these data into a n array ,storing ECG,x axis accelerati on ,y ax is acceleration ,z ax is acceler at i on and pressure in seque nce for 5tm i es of circular sa mpling ,w ith te mperat ure and hu m idity data at t he e nd of the array . T he soft w are design of our proposed syste m is based on the Z ig Bee protoco.l H o wever ,our collected data has a h i gh data rate c o m pared to the ho me control syste m [25].W e take package trans m issi on m ethods to m i prove net wor k perf or m ance .F ig 5and F ig 6sho w the flo w c harts of the ter m i na l node and the sink node ,respective l y.Fig 5 F l ow cha rt o f t he ter m i na lnodeFig 6 F l ow cha rt o f t he si nk node1 3 G raphical user i nterface desi gnT he ter m i na l display and the si gna l pr ocess are sho wn i nF i g 7and F i g 8.A real tm i em onitori ng soft ware m i ple m en ted on the PC is created usi ng the LabV I E W progr a m.On the PC,t he ter m inal display a nd si gnal pr ocess can sho w the data of electrocardi o si gna ls ,pressure signals and accelera tion signals which are all m easured by the body se nsors and tra ns m itte d by the w ireless net wor k system.Fig 7 A n exa m ple o f m easuredsignalsF i g 8 The ter m i na l d is p lay and the si g na l pro ce ssI nteractions bet w een states of the autono m ic nervous syste m and cognit i ve perf or m ance have l ong been a tradit i on as a topic of psyc ho l og ical researc h .C lassic concepts fr o m mo tivational psyc ho l ogy suggested a n i nverted u shaped associa tion bet w een unspecif i c activation and m e nta l function i ng [5 6].A ccor ding to this ,the best f unctional conditi ons are expected at m i drange arousa,l and bot h over ar ousal75Physiol og i cal s i gnal acqu isiti on syste m based on w i reless se nsor net w orksand under arousal are acco m panied by decli nes i n perf or m ance.Cardi ovascular psyc hophysi ology has also contri buted to this li ne of researc h,the respecti ve models relating chan ges i n cardi ovascular activity to facilitat i on or inhi biti on of i nfor mat i on processi ng[7]or ener getic m obilizati on of the or ganis m when faced w ith a situati on requiri ng behav i oral ad j ust m e nt[8].H o wever,alt hough th i s certa i nly constitutes a benefici a l approac h,e m piricalwork in th i s fiel d re m ains rel at i vel y sparse[9].2 R esu lts and D iscussi o nT he attent i onal capac ity is assessed usi ng a classic letter cance ll at i on test[23].T asks of this type address the cogni t i ve co m ponents of selective and sustained attention that are undoubted l y of vast i m portance in everyday life[26 27]. In the test,subjects have to select and m ark as m any tar get sti m uli as possi b le i n a g i ven a m ount o f ti m e,and, hence,it also creates a certa i n load on the speed o f infor m ation processing.A utono m ic para m eters are recor ded under resting cond itions duri ng the exec ut i on of the task. O ne m ay assum e that features o f autono m ic control as sessed duri ng cognitive processi ng sho w the closest li nk to perf or m ance.On the other hand,on task m easures are i n fluenced by factors such as m enta l effort,e mo tiona l stress or sub jectively experienced task d ifficulty.I n contrast, base li ne m easures are free of these confounding variab les. T he fo ll ow ing predict i ons are m ade:1)T aking an i nverted u shaped assoc i at i on bet w een unspec ific sy m pathetic a rousal a nd m ental perf or m ance into account,a nd assum i ng an e xperm i ental situati on i n which sy mpathetic over act i v ity is unli kely to occ ur,an i nverse re l ationshi p bet ween RP I and atte ntional perfor mance may be expecte d;2)O n account of Por ges model[19],w e predict a positive correlati on bet ween the resting RSA and perfor mance;3)Our find i ngs on the baroreflex funct i on suggest t hat i ndividualsw ith an i ncreased BR S should e xhi b it poorer perfor mance[28];4)G i ve n the in verse associat i on bet ween oscillati ons in theM F ba nd a nd the m ental effort l oad,and supposing better perf or m ance i n the case of hi gher effort,t heM F po wer assessed duri ng task exe cut i on shou l d correlate ne gati vely w ith perfor mance;5)Con sideri ng Porges hypot hesis[19],we expect higher cardi ovas cular reactivity in indi v i duals w ith the hi gher rest i ng RS A;6)Even though the available database is so me what controver sia,l the li kel y assoc i ation bet ween mental effort and auto no m ic reactivity suggests str onger react i v ity to be related to better perfor m ance.3 Concl u sionIn this paper,w e develop a real tm i e,physiological data acquisiti on syste m f or educat i on assess m ent,wh i ch co m bi nes LabV I E W and w ireless sensor net wor ks,and w e fulfill the f unction of collecting,trans m itti ng,displ ay i ng and process i ng physiolog ical signals.In f uture w ork,we plan to add m ore f unctions to the ter m i nal display and si gnal process suc h as the e x traction of the special freque ncy spectru m s, and furt her psyc ho l og ical researches on the physi o l og ical si g nals fro m more subjects.F i nally,our subsequent consi dera t i on is to set up a more efficient m et hod to obtain physi o l ogi cal signals for assessing e motional states and mental reacti ons duri ng lear ning i n the classroo m.References[1]R a i nv ille P,Becha ra A,N aqv iN,et a.l Ba sic em o ti on s areasso ciated w ith d isti nc t pa tterns o f cardio re s p ira t o ry acti v ity [J].Int J P sychophy sio l,2006,61(1):5 18.[2]D uschek S,M uckentha ler M,W erner N,e t a.l R e l a tionshi p s be t w een fea t ures o f aut onom ic card i o va scu l ar contro l and cogn iti v e perfo r m ance[J].B i o l o g ica l P sycho l o gy,2009,81(2):110 117.[3]S te mm ler G.T he autonom i c diffe renti a ti on o f em o ti ons rev isited:conv erg en t and d iscri m i na te va lida tion[J].P sycho physi o l o gy,1989,26(6):617 631.[4]Izard C E,M alate sta C Z.P erspec tives on e m o ti ona l deve lop m ent[M].N ew Y o rk:W il ey,1987:494 554.[5]H eeb D O.D r i v es and the CN S[J].P sycho l Rev,1955,62(4):243 254.[6]Y erkesR M,D o ds o n J D.T he re l a tion o f streng th o f sti m ulus t o rapi d ity o f habit fo r m ati o n[J].Journa l o f Com pa ra tive N euro lo gy and P sycho log y,1908,18:459 482.[7]L acey J I,L acey B C.Som e au t onom ic cen tra l nerv oussy stem i nterrela tionshi p s[C]//Physi o lo g ica l C o rre l a te s o fE m o tion.N ew Y o rk:A cade m ic P ress,1970:205 227.[8]O brist P A.C ardiova scula r psychophy sio log y:a per s pective[M].N ew Y o rk:P lenu m Press,1981.[9]H ansen A L,John sen B H,Thayer J F.V aga l i nfluence onw o rking m e m o ry and a tten tion[J].Interna tiona l Journa l o f P sychophy sio l o g y,2003,48(3):263 274.[10]L ev y M N,P appano A J.C a rdi o va scu l a r phy sio log y[M].Phil adelphia:M o sby E lsev ier,2007.[11]V an R oon,M u l de r L J M,A lthausM,e t a.l Introduc i ng abaro reflex m ode l fo r st udy ing card i o vascular effects o f m en tal w orkload[J].P sychophysio l o gy,2004,41(6):961 981.[12]D em bow s ky K,Se ll er H.F rom the hea rt to the bra i n:thep sycho physi o lo gy o f c ircu l a tion bra i n i nteraction[M].Frankfur:t Europ ischer V erlag der W issenschaf t en,1995:35 60.[13]R au H,E lber tT.P sy chophy si o logy o f arter i a l baro recepto rsand t he eti o l og y o f hypertensi on[J].B i o l P sycho l,2001,57(1/2/3):179 201.[14]Z hang W e,i Ji ng Shenq,i G ui L i n,e t a.l Phy si o l og i ca l dataacqu isition sy ste m fo r educa ti on a ssess m ent usi ng w i re l e ss senso r net w o rk[C]//P ro ceed i ng s o f the5th Interna tiona l Co nference o n Info r m a ti on T echno l o gy and Applica ti on.Shenzhen,Ch i na,2008:471 473.[15]D i ng Fe,i Song G uangm i ng,Song A i g uo,Hom e contro ls y ste m based on Z i gB ee w ire l e ss senso r net w o rks[J].Jour na l o f Sou t hea st U n i ver sity:Eng lis h Edition,2008,24(4):420 423.[16]Hong J H,K i m J M,C ha E J,et a.l A w ire less3 channelECG trans m issi on sy ste m usi ng PDA pho ne[C]//P roceed ings o f Interna tiona l Con ference on Converg ence Info r m a ti on Techno l o gy.Seou,l K o rea,2007:462 465.[17]C erny M,Penhaker M.T he hom ecare and circad i an rhy t hm[C]//P ro ceed i ng s o f t he5th Interna tiona l C onference onInfo rma tion T echno log y and A pplica ti o n.Shenzhen,Ch i na,2008:245 248.[18]C ar m en L opez C asado.N et w o rk arch itecture fo r g l o ba l b i om edical m on it o r i ng s e rv ice[C]//P ro ceeding s o f IEEE Eng i neer i ng in M ed ici ne and Bio log y27t h Annua l Co nference.Shangha,i C hina.2005:2433 2436.[19]Po rg es S W.Hear t rate va riab ilit y and dece l e ra tion as i ndexe s o f reac tion ti m e[J].Journa l o f Exper i m enta l P sycho l o gy,1972,92(1):103 110.[20]Po rter F L,Po rg es SW,M arshallR E.N e w bo rn pa i n cries76Q i u W en ji ao,and Zhang Yongku iand v ag a l tone :para lle l changes i n re s ponse t o c i rcu m c isi on[J].Ch ild D ev,1988,59(2):495 505.[21]D e G ang iG A,L aur i e R S ,C ast e ll an J ,et a.l T rea t m ent o fsenso ry ,e m o tiona,l and a ttenti ona l proble m s i n regu l a t o ry d is o rdered i n fants [J].In fants and Young Chil dren,1991(3):9 19.[22]D uschek S .E ssen ti a l hy po ten si on is accom panied by defic itsi n a tten ti on and w o rking m e m o ry[J].Beha vio ra lM edicine,2005,30(4):149 158.[23]B acks R W ,Se ljo s K A.M etabo lic and cardio re s p ira t o rym easures o fm en t a l e ffo r:t t he e ffects o f leve l o f diff i culty i n w o rki ng m e m o ry task [J].Int J P sycho physi o l,1994,16(1):57 68.[24]G endo ll a G H E ,W right R A.M o tiva tion i n soc i a l setti ng s :studies of effo r t related cardiov ascul a r arousa l[C ]//So cia lm o ti va ti o n:Consci o us a nd N onconsc i ou s P ro ce sses.L o ndo n :C a m br i dge U n i v ersity Press ,2005:71 90.[25]L ee M i H ee ,Y ang G yunghye ,L ee H yo ung K ,i et a.l D eve lop m ent stress m on it o r i ng sy st em ba sed on personal d i g ital a ssistant(PDA )[C ]//P roceedi ng s o f t he 26t h Annua l Inter na tiona l C onfer ence o f the IEEE E M BS.San F rancisco,CA,U SA,2004:2364 2367.[26]J o hnson A,Proc t o r R W.A ttention:theo ry and practice[M ].Tho usand O aks ,CA,U SA:Sag e ,2004.[27]Po s ner M I ,R afa l R D.Co gn itive theo ries of a ttenti on andt he rehab ilit a ti on o f atten tional defic its[C ]//N eu ro psycho lo gica l rehabilita ti on.Edinburgh :Churchill L i v i ng ston ,1987:182 201.[28]Y a su m asu T,R ey es de l Pa s o G A,T akahara K,et a.l R educed baro reflex cardiac sen sitiv ity pred icts i ncreased co gni tive perfo r m ance[J].P sychophysi o l o gy,2006,43(1):41 45.一种基于无线传感器网络的生理信号采集系统邱文教 张永魁(东南大学儿童发展与学习科学教育部重点实验室,南京210096)摘要:提出了一种应用于课堂教育的无线生理采集系统,以了解学生在上课过程中的生理变化情况.系统中,反映学生的注意力和情绪状态的生理电信号可通过测量ECG 来获得.生物电信号采用200H z 的采样速率,并由Z i g Bee 协议进行无线传输.同时,节点嵌入了蓝牙技术以适应高采样速率和高带宽的传输要求.本系统可以同时对30个学生进行监测,并且给出了系统在教室执行监测任务的实验结果.最后探讨了无线传感器网络在教育领域的应用.关键词:无线传感器网络;生理信号;教育中图分类号:TN9277Physiol og i cal s i gnal acqu isiti on syste m based on w i reless se nsor net w orks。