Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Look up keyword
Like this
1Activity
0 of .
Results for:
No results containing your search query
P. 1
Interpersonal Dynamics in Sport

Interpersonal Dynamics in Sport

Ratings: (0)|Views: 13|Likes:
Passos, P., Araújo, D., Keith Davids, K., Gouveia, L., Serpa, S. (2006). Interpersonal dynamics in sport: The role of artificial neural networks and 3-D analysisBehavior Research Methods, 38 (4), 683-691.

From abstract in article: In previous attempts to identify dynamical systems properties in patterns of play in team sports, only 2-D analysis methods have been used, implying that the plane of motion must be preselected and that movements out of the chosen plane are ignored. In the present study, we examined the usefulness of 3-D methods of analysis for establishing the presence of dynamical systems properties, such as phase transitions and symmetry-breaking processes in the team sport of rugby. Artificial neural networks (ANN s) were employed to reconstruct the 3-D performance space in a typical one-versus-one subphase of rugby. Results confirm that ANs are reliable tools for reconstructing a 3-D performance space and may be instrumental in identifying pattern formation in team sports generally.
Passos, P., Araújo, D., Keith Davids, K., Gouveia, L., Serpa, S. (2006). Interpersonal dynamics in sport: The role of artificial neural networks and 3-D analysisBehavior Research Methods, 38 (4), 683-691.

From abstract in article: In previous attempts to identify dynamical systems properties in patterns of play in team sports, only 2-D analysis methods have been used, implying that the plane of motion must be preselected and that movements out of the chosen plane are ignored. In the present study, we examined the usefulness of 3-D methods of analysis for establishing the presence of dynamical systems properties, such as phase transitions and symmetry-breaking processes in the team sport of rugby. Artificial neural networks (ANN s) were employed to reconstruct the 3-D performance space in a typical one-versus-one subphase of rugby. Results confirm that ANs are reliable tools for reconstructing a 3-D performance space and may be instrumental in identifying pattern formation in team sports generally.

More info:

Categories:Types, Research, Science
Published by: Laboratory in the Wild on Jul 09, 2012
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

07/28/2013

pdf

text

original

 
Idntifing th rasns fr which sm indiidas andtams achi sccss is n f th mst cmmn gas inth rsarch itratr n sprt prfrmanc. Th mthd-gis sd t std sprt prfrmanc ha bn stab-ishd t cmpar th actins f sccssf and nsccss-f tams and athts, and th tcms incd a rangf statistica data f discrt actins prfrmd b tamsand pars dring cmptitin (s Hghs & Franks,2004). Dspit th imprtanc f ths data, a significantcriticism f th ntatina anasis f sprt prfrmancis that it ds nt idntif th rasns fr ths discrtactins in rdr t xpain th diffrnc btwn sccss-f and nscssf tams. Th data tpica infrm s f what happns, bt nt f hw and wh it happns.T ndrstand th dnamica prcsss in sprt prfr-manc, McGarr and Pr (2004) sd a spcific tp f nra ntwrk: th Khnn fatr map. Th argdthat th nra ntwrks apprach mst b sd first trcgniz sitatins and anaz prcsss, and n thnt idntif dcisin-making prcsss. Hwr, with thatapprach, a 2-D infrmatin strctr (.g., a pair f c-rdinats) was impmntd t rprsnt a pattrn, whichdcrass th abiit f a nra ntwrk t anaz thcmpx prcsss that mrg in sprts cntsts.Additina prbms f ntwrk arning prcsssincd thir dpndnc n inpt pattrns and th factthat ths pattrns chang cntins as a rst f a par’s tactica bhairs. A thrtica ratina frmth fid f cgica pschg aws s t ndrstandthat tactica bhairs ar dpndnt n th infrmatinaaiab in spcific cntxts, and that that infrmatinis gard t ach par’s tactica bhair. Car, ths f Khnn fatr maps t std prfrmanc mightincr prbms in riab rcgnizing pattrns in sprtscntsts. Anthr mthd prpsd b McGarr and Pr(2004) is th dnamica cntrd ntwrk (DCN), a-rad tstd sccssf in diffrnt aras f sprt (Pr,2002). This mthd das with th prbm f dpndncn inpt pattrns. Th DCN has th abiit t arn cn-tins r in discrt phass, s inpt pattrns nd nt b prsntd t th ntwrk in a sing data st. McGarrand Pr (2004, p. 241) argd that th infrmatin ac-qird dring th arning f th DCN cd b sd in“cmpmnting r ndrstanding f th prcsss thattak pac in th arning f cgniti and/r mtr skis683 Cpright 2006 Pschnmic Scit, Inc.
Crrspndnc cncrning this artic shd b addrssd t P. Pas-ss, Ra Jsfa Gabis Diniz, N. 153, 2. Dt., Arnir, 2775-522 Car-cas, Prtga (-mai: p.passs@cix.pt).
Itpl dymic i pt: T l f tificil l twk d 3-D lyi
PeDro Passos
Technical University of Lisbon, Lisbon, Portugaland Lusófona University of Humanities and Technologies, Lisbon, Portugal
DuarTe araújo
 
Technical University of Lisbon, Lisbon, Portugal
KeITh DavIDs
University of Otago, Dunedin, New Zealand
Luís GouveIa
University of Lisbon, Lisbon, Portugal
dsIDónIo serPa
Technical University of Lisbon, Lisbon, Portugal
I pi ttmpt t idtify dymicl ytm ppti i ptt f ply i tm pt, ly2-D lyi mtd  b d, implyig tt t pl f mti mt b plctd d ttmmt t f t c pl  igd. I t pt tdy, w xmid t fl f 3-D mtd f lyi f tbliig t pc f dymicl ytm ppti, c  ptiti d ymmty-bkig pc i t tm pt f gby. atificil l twk(ann) w mplyd t ctct t 3-D pfmc pc i  typicl -- bpf gby. rlt cfim tt ann  libl tl f ctctig  3-D pfmc pc dmy b itmtl i idtifyig ptt fmti i tm pt glly.
 Behavior Research Methods2006, 38 (4), 683-691
 
684 PASSoS, ARAÚJo, DAvIDS, GouveIA, AND SeRPAin hmans.” This prpsa raiss a paradx: Hw cansprts cntsts b intrprtd as dnamica sstms withan appid mthdgica md f fnctining basd ndiscrt cgniti apprachs?This qstin highights th nd fr a mthdg thatwi aw rsarchrs t anaz th cmpxit f dnami-ca prcsss in sprt prfrmanc whn th sk t stab-ish paramtrs in th thr pans f mtin. Th artificianra ntwrk (ANN) prpsd in this artic rprsntsan attmpt t initiat this mthdgica apprach.
Prvis Rsarc n Tam Sprt Sbpass
viwd as dnamica sstms, tam ba sprts arhigh cmpx phnmna bcas f th nmbr f d-grs f frdm ind that charactriz th ratin-ships btwn sstm cmpnnts that  r tim(.g., an attackr and a dfndr ar th tw cmpnntsf a dadic sstm). Thrfr, intrprsna crdina-tin within tams mrgs frm th cping f th cm- pnnts f th sstm (i.., th pars). Th first stp inndrstanding tam ba gams as dnamica sstms ist find simp masrs that captr th cmpxit f th sstm. T achi this ga, th paramtrs that bstdscrib intrprsna crdinatin r tim nd t bdiscrd.Sm initia attmpts ha bn mad t dp tsfr ccting sch data frm ra gam sitatins, thrbawing an anasis that gs bnd th s f cntratndnc paramtrs, which prid n intrmittntdata. Fr xamp, within th framwrk f dnamicasstms thr, Araúj, Daids, Saínhas, and Frnands(2002) cncptaizd th attackr–dfndr dad in bas-ktba as an intrprsna crdinatin sstm, sggst-ing that th attackr and th dfndr ar tw mnts f a cmpx, dnamica sstm inkd b isa (and thr)infrmatina fids.In thir std, Araúj t a. (2002) sd a mthdgt masr th (2-D) hriznta-pan trajctr f thcntr f mass f ach par. Rsts wr rprtd in aqaitati anasis sbstantiatd b psitin–tim (.g.,intrprsna distanc r tim, distanc f ach par t baskt r tim) and psitin–psitin (.g., dfndr  psitin as a fnctin f attackr psitin) crdinats.
1
 Data spprtd th iw that a smmtr-braking pr-css had ccrrd in sccssf dribbing whn thr wasan abrpt chang in th shap f th cr dscribing thintrprsna distanc f ach attackr–dfndr dad.undr th distinct task cnstraints f rgb, man simiar instancs f intrprsna crdinatin ma b bsrd— fr xamp, attackr-rss-dfndr mmnt pattrnsnar th tr in. Hr, th ga f th attackr is t crss thtr in and grnd th ba b hand. Th dfndr attmptst rmain btwn th attackr and th tr in b cn-trbaancing an mmnt f th attackr, rsting in astat f intrprsna crdinatin dring this sbphas f th gam. As pris wrk n basktba (Araúj, Daids,Bnntt, Bttn, & Chapman, 2004) sggsts, in rgb thattackr’s dcisin making shd mrg frm his r hr mmnts dring a prcss f sarching fr an aaiabrnning in tward th tr ara and aiding a tack. Dr-ing a sccssf rn b th attackr, th initia stat f rdr f th dad (.g., a psitina baanc btwn attackr anddfndr) is annihiatd r sstaind b infrmatin fids.A nw stat f rdr mrgs in this dnamica sstm dr-ing a phas transitin.As w statd arir, in xpring th mrgnc f smmtr-braking prcsss in tam ba sprts, Araújt a. (2004; Araúj t a., 2002) sd 2-D anasis tch-niqs. Hwr, 2-D anasis mthds rqir that th pan f mtin f ach participant b prsctd, and,thrfr, mmnts t f th chsn pan ar ignrd(Barttt, 1997). Mrr, in qaitati trms, Ks,Bchanan, DGzman, and Ding (1993) dmnstratdthat whn mtin is rstrictd t th hriznta pan,n th smmtric and asmmtric mds f crdina-tin ar stab r a rang f frqncis. Fr frqncisab that rang, hwr, n cmparab stab pattrnis aaiab in th hriznta pan f mtin. Th sstmcnsisting f an stabishd, acti st f cmpnnts thatha bn sf-rganizd fr particar mmnt pattrnsis n ngr ab t spprt sch bhair. T achi sta- biit, pris qiscnt sstm dgrs f frdm(i.., ths in th rtica dimnsin) ar spntansrcritd. With th aaiabiit f ths additina dgrsf frdm, th sstm can attain stabiit nc mr. Itis ntwrth that th nw mmnt pattrn ma sti b tpgica qiant t th pris n, bt ad-ditina dgrs f frdm ar rqird t prfrm thtask. Cnsqnt, athgh Araúj t a. (2004; Araújt a., 2002) cd bsr phas transitins in intrpr-sna crdinatin dnamics n in th hriznta pan,smmtr-braking prcsss might as b bsrab inth rtica pan.In mthdgica trms, accrding t Barttt (1997),th majr adantags f 3-D mtin anasis r 2-Danasis incd th facts that it (1) can shw th cmptspatia mtin f participants; (2) aws thr angs f isin—fr xamp, th trajctris f attackrs and d-fndrs can b isaizd frm ab; and (3) nabs thrcnstrctin f simatd iws f prfrmanc thr than ths sn b th camras. Thrfr, with 3-D ana-sis, pars’ dispacmnt dring sbphass f gamscan b bsrd frm an chsn ang, and anasis f smmtr-braking prcsss is nt imitd as it is withth s f n a hriznta-pan anasis.
Biding and Dvping a 3-D Md
T anaz th trajctr f bth pars in a dad, wsd a 3-D imag masrmnt tchniq cad
video- grammetry
(Grn, 1997). Th ga f this tchniq is trcnstrct 3-D spac b transfrming th pairs f crdi-nats f ach camra int thr crdinat pans (
 x
,
 y
, and
 z
). Th rcnstrctin f 3-D spac is trmd th
resolution stereo problem
, and th tchniq that ffrs th mst ad-antags is th ANN apprach (Mmn & Khan, 2001).Th adantags f adpting ANN t s th rs-tin str prbm incd th fwing: (1) on candispns with th knwdg f intrinsic camra param-
 
ARTIFICIAl NeuRAl NeTWoRK AND 3-D ANAlySIS 685trs dmandd b inar mds (.g., fca distanc,imag cntr, imag sca, distrt cfficint ns), asw as xtrinsic camra paramtrs (.g., camra psitinratd t rigin, camra rintatin). (2) With th ANN,it is pssib t wrk with an kind f camra. (3) Thris n nd fr camras t b in a fixd catin in ratint th rigin. (4) Th axs amng th camras d nt ndt b aignd in a fixd psitin. (5) unik inar md-s, th ANN das atmatica with th distrtin ns prbm (Grn, 1997).An ANN is an infrmatin-prcssing sstm with para distribtin and a tndnc t str xprimntadata fr ftr s (Hakin, 1994; Smith, 2001; Strgi& Sigans, 1996). Th ANN’s fnctins ar basd ntw main facts: (1) that knwdg is acqird thrgh aarning prcss and (2) that th intrnrn cnnctinstrngthns with snaptic wights that ar sd t chan-n infrmatin (Hakin, 1994). In r std, w sd anANN f th fd-frward cass bcas it has a tpgthat aws infrmatin t fw frm inpt t tpt.Fina, sing ANN nhancs th chrnc btwnth thrtica and mthdgica apprachs w ar f-wing. As an xamp, th ANN can da with th ari-abiit f camra paramtrs, whras inar mthds can-nt. Frm a thrtica prspcti, dnamica sstmsthr assms that bth inar and nninar prcsssxist in bigica mmnt sstms (s, .g., Ks,1995); frm a mthdgica pint f iw, sing inar mthds biass th anasis tward inarit, ths cratinga fndamnta cntradictin.T dmnstrat hw ANNs might b a sf mthd-gica adanc in th std f pattrn frmatin andintrprsna dnamics in tam sprts, which ar iwdas cmpx sstms, w appid th mthd fr anazing phas transitins in rgb. Rgb is a prfrmanc cn-txt charactrizd b intns phsica cntact. Accrdingt th rs f th gam, dfndrs ar awd t tackattackrs, tak thm t th grnd, and rcr pssssinf th ba fr thir tam. Mrr, it is nt prmissib t pass th ba frward with th hands, and th bas shapmaks it diffict t cntr it with th ft. Ths task cnstraints charactriz rgb as a sprt with a diffrntfnctina rganizatin than that f th sprt f baskt- ba, instigatd in pris wrk.Th rgb caching itratr prids dtaid d-scriptins f ptima was t tack an ppnnt (s, .g.,Biscmb & Drwtt, 1998). Ths dscriptins xpainwhat dfndrs mst d whn cntact with attackrs c-crs. Hwr, cntribtrs t th pdaggica itratrtpica fai t addrss hw dfndrs can dri attackrst a cntact pint. Prhaps th xcptin is Grnwd(2003), wh statd that fr a sccssf tack, dfndrsmst “start fast bt arri in baanc” (p. 166). W canspcat frm his wrds that if dfndrs “arri in ba-anc,” th dad frmd btwn an attackr and a dfndr wd maintain its initia stat f rdr, and that if dfnd-rs cannt “arri in baanc,” a smmtr-braking pr-css is mr ik t ccr. Hwr, th trm “arriingin baanc” is rathr nbs fr ndrstanding th mr-gnc f stabiitis and instabiitis in th intrprsnadnamics f tam ba gams. A particar intrstingqstin cncrns th paramtrs in th attackr–dfndr dads that nab thm t maintain r chang thir fnc-tina rganizatin.Ths, r std had thr gas, a f which w achidsing sbphass f th sprt f rgb: (1) t idntif phastransitins in a tam ba sprt with task cnstraints diffr-nt frm ths f basktba, (2) t prsnt a 3-D anasisf intrprsna dnamics f attackr–dfndr dads,and (3) t idntif paramtrs t masr dnamica ss-tms prprtis in ths dads.
MeThoD
Participants
eight ma rgb pars (11–12 ars f ag) participatd in thstd (man ars f rgb practic
 
5
4.0,
SD
 
5
0.5). Th dcisint std sch ng pars was mad t prnt th cnfndingffcts f xprtis and diffrnt dgrs f arning, sinc w b-id that ths ariabs wd ha biasd th data (cf. Zann &Ks, 1994). W fwd th prcdrs tind b Bats, Dfk,and Dais (1992) fr nsring a crrspndnc btwn th nmbr f trias (task rptitins) and th nmbr f participants: Fr sampsizs f 10, 5, and 3 trias, th samp sizs f participants shd b 5, 10, and 20, rspcti. Bcas f th natr f r task, participant intractins wr anazd as nts f intrprsnacrdinatin in attackr–dfndr dads. Thrfr, w randmacatd a nmbr t ach rgb par f r samp. each par actd as bth attackr and dfndr and participatd with r thr  par in dads. In rdr t prnt pssib fatig ffcts n pr-frmanc, it was dcidd that ach dad wd prfrm tw trias.With a pssib nmbr cmbinatins, w bsrd th rspnssf 32 diffrnt dads.
Task 
W dsignd a task that was rprsntati f a sbphas f rgbining th minimm nmbr f pars: th cmmn n-rss-n sitatin nar th tr in. In this sbphas, a pair f rgb pa-rs with ppsing fnctins and gas frms a dad in which thattackr aims t pac th ba in th tr ara with his hand and thdfndr aims t stp th attackr’s prgrssin tward th tr inand tak him t th grnd within th rs f th gam. W stab-ishd th dpth and width f th fid fwing th prcdrs f Biscmb and Drwtt (1998), wh rcmmndd th fid dimn-sins fr n-rss-n sbphas practic tasks. This adic hpds t cnstrct th task cnstraints f th xprimnta wrk in r std (s Figr 1).Accrding t thir cachs, a th sctd participants knw hwt tack (i.., hw t intrcpt an attackr and stp his prgrssintward th tr in). Dfndrs attmptd t cntact attackrs withth shdrs and arms, and th hd thir gs r trnks t takthm t th grnd (Biscmb & Drwtt, 1998).
Instrmnts
Par mtin was captrd b tw digita id camras. Fr imag tratmnt, w sd th sftwar TACTo 7.0, digitizd at 25frams/sc (Frnands & Caixinha, 2003). Th ba sd was siz 5,as rqird b th Prtgs Fdratin f Rgb fr th ag grpf th participants. Tw rds, ach 1.8 m in ngth, wr sd t bid a cb-shapd rfrnc strctr, which was md in rdr t cr th ntir data cctin m. This rfrnc strctrga s th pints sd fr th training and aidatin prcdrsfr th ANN.

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->