Reliability and validity of motion sensor and radar for measuring shuttlecock velocity in badminton. Reliability and validity to measure velocity in badminton

In . j. acke spo s sci. ol. 4(2), 2022, 11-19. eISSN: 2695-4508 11 DOI: 10.30827/Digibug.---O iginal In es iga ion Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on. Reliabili y and alidi y o measu e eloci y in badmin on Fiabilidad y alidez del senso de mo imien o y el ada pa a medi la elocidad del olan e en bádmin on. Fiabilidad y alidez pa a medi la elocidad en bádmin on Abs ac Rada dopple and ine ial measu emen uni a e o en used o analyze he p ojec ile eloci y. The aim o he p esen s udy was o analyse he eliabili y and alidi y o a speci ically mo ion senso (named: Zepp Tennis) and a ada (Dopple - ada gun) o measu ing p ojec ile eloci y. Thi y- ou (no ice, in e media e and expe ) s oke badmin on smash in a loca ed a ge . P ojec ile eloci y om i e smashes we e ex ac ed using Zepp Tennis and Dopple - ada gun da a. Be ween ep oducibili y o measu es was de e mined by compa ing he wo sessions. Zepp Tennis and Dopple - ada gun measu es we e compa ed wi h high- equency ideo da a o es ablish alidi y. Bo h ins umen s we e highly ep oducible be ween ials a di e en eloci y (in a-class co ela ion coe icien : 0.88- 0.94 o ada and 0.78-0.89 o mo ion senso ). In addi ion, he posi ioning o he ada ( on o he p ojec ile and angula ion) and he placemen o he mo ion senso and he complexi y o he mo emen ( o ea m ex ension and p ona ion) a ec he ep oducibili y. In e ms o alidi y, ada and mo ion senso p o ides an accu a e measu e bu unde es ima e p ojec ile eloci y (-9.7% and -13.6% espec i ely). Keywo ds: Shu le un, pe o mance analysis, ecological alidi y, lunge. Resumen El ada Dopple y la unidad de medición ine cial se u ilizan a menudo pa a analiza la elocidad del p oyec il. El obje i o de es e es udio ue analiza la iabilidad y la alidez de un senso de mo imien o (denominado Zepp Tennis) y un ada (pis ola de ada Dopple ) pa a medi la elocidad del p oyec il. T ein a y cua o jugado es (no a os, in e medios y expe os) ealiza on golpes de bádmin on en un obje i o localizado. Se ex ajo la elocidad del p oyec il de cinco golpes u ilizando los da os del Zepp Tennis y de la pis ola de ada Dopple . La ep oducibilidad en e las medidas se de e minó compa ando las dos sesiones. Las medidas del Zepp Tennis y de la pis ola de ada Dopple se compa a on con los da os de ídeo de al a ecuencia pa a es ablece su alidez. Ambos ins umen os ue on al amen e ep oducibles en e las p uebas a di e en e elocidad (coe icien e de co elación in aclase: 0,88-0,94 pa a el ada y 0,78-0,89 pa a el senso de mo imien o). Además, la ubicación del ada (en en e del p oyec il y angulación), la ubicación del senso de mo imien o y la complejidad del mo imien o (ex ensión y p onación del an eb azo) a ec an a la ep oducibilidad. En é minos de alidez, el ada y el senso de mo imien o p opo cionan una medida p ecisa, pe o subes iman la elocidad del p oyec il (-9,7% y -13,6% espec i amen e). Palab as cla e: Ca e a, análisis del endimien o, alidez ecológica, zancada. Ci e his a icle as: Phomsoupha, M., Bonneau, F., & La aye, G. (2022). Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on. In e na ional Jou nal o Racke Spo s Science, 4(2), 11-19. This is an open access a icle unde he CC BY license (h ps://c ea i ecommons.o g/licenses/by/4.0/). Co esponding au ho : Michael Phomsoupha, [email p o ec ed] Michael Phomsoupha 1, 2 *, Flo ie Bonneau 3, 4 and Guillaume La aye 5 1 M2S (Labo a oi e Mou emen , Spo , San é), Uni . o Rennes, - EA 1274, Rennes, F ance 2 APCoSS - Ins i u e o Physical Educa ion and Spo s Sciences (IFEPSA), Uni e si é Ca holique de l’Oues , Ange s, F ance 3 CIAMS, Uni . Pa is-Sud, Uni e si é Pa is-Saclay, 91405 O say Cedex, F ance 4 CIAMS, Uni e si é d’O léans, 45067, O léans, F ance 5 Resea ch Cen e o Spo s Science, Sou h U al S a e Uni e si y Chelyabinsk, Russia Recei ed: 29-11-2021 Accep ed: 10-11-2022 In . j. acke spo s sci. ol. 4(2), 2022, 11-19. eISSN: 2695-4508 Michael Phomsoupha, Flo ie Bonneau & Guillaume La aye 12 INTRODUCTION The abili y o p oduce a high p ojec ile eloci y du ing a s oke o a sho is one o he main pe o mance ac o s in spo s such as able ennis, ennis, oo ball and baseball. Fu he mo e, high eloci y is co ela ed wi h pe o mance in se e al spo s (La aye e al., 2012). P ojec ile can include objec s (ball, shu lecock…) o a hle es in ligh . In Badmin on, he shu lecock eloci y e ol ed linea ly wi h skill le els (Phomsoupha & La aye, 2020). Especially in badmin on, many s udies in es iga ed he shu lecocks eloci y o ind he biomechanical p inciples ha op imize he mo ion. In badmin on games, he highes eloci y is ob ained du ing he smash s oke, which gene ally allows o inish a ally (Rambely e al., 2005). A ecen s udy showed ha du ing he di e en Olympic Games inals in men’s single, smash is used in abou 10 o 14% o he o al s okes (La aye e al., 2015). The smash can be de ined as an agg essi e o e head sho wi h a downwa d ajec o y (Phomsoupha & La aye, 2015). A his ime, he eco d o shu lecock eloci y is 493 km/h pe o med by he Malaysian Tan Boon Heong (Yonex, 2013). In such a con ex , assessing he shu lecock eloci y wi h accu acy is a main s ake o a hle e moni o ing o es ing. To eco d pe o mance, image p ocessing is applied o a ious ields o spo s, such as mo ion analysis, game analysis (La aye e al., 2015), and physical educa ion (Nagasawa e al., 2012). The in o ma ion including da a om he science s a is needed by he playe , he supe iso , and he coach o imp o e he pe o mance (Nagasawa e al., 2012; Takahashi & Kawaha a, 2011). Fu he mo e, his is in o ma ion eedback o he playe ’s pe o mance. To eco d he maximal eloci y o he p ojec ile, s udies habi ually used s anda d ideo came a (Hussain & A shad Ba i, 2011; La aye e al., 2014). Howe e , his me hod e eals wo main weaknesses. The s anda d e o o measu emen depended o he a io o he launch eloci y on he numbe o ame pe second and esolu ion o he sys em (Nagasawa e al., 2012). The momen o he peak eloci y has o be as close as possible o he momen he shu lecock qui s he acke . The con ac be ween acke s, speci ically he s ing and he ball happens a a e y sho ime (abou 5-6 ms) (Mille , 2006). Phomsoupha & La aye (2014) showed ha he shu lecock eloci y V (in m.s-1) is a loga i hmic unc ion o ime T (in ms) as ollow: V= -9.2 ln(T) +2.4, meaning ha he eloci y is di ided by wo imes jus 0.05 sec a e he end o he acke con ac . The di e ence o eloci y be ween eli e and high skilled playe s is abou 10 m.s-1 (Phomsoupha & La aye, 2014), whe eas he eloci y accu acy o a came a a 50 ps is abou 5 m.s-1. This showed ha i is impossible o assess a badmin on shu lecock eloci y wi h such de ices. The measu emen o he eloci y o p ojec ile used in spo games is becoming inc easingly common. Such speeds a e usually measu ed using ada guns (Robinson & Robinson, 2016). These de ices measu e he equency di e ence be ween he e lec ed signal and he ansmi ed signal o ela e he ela i e speed o he ball and he ada (Halliday e al., 2011). An inc ease in equency o he e lec ed signal shows he p ojec ile is app oaching and a dec ease indica es an inc ease o he dis ance. Mo eo e , ada gun de ices measu e only he adial eloci y and will always unde -es ima e he eal eloci y (F ench, 1968; Resnick, 1968). Wi h ad ances in mic oelec onics, wea ables ecen ly gained signi ican a en ion in spo s (Coyle e al., 2009). Speci ically, o acke spo s, he machine lea ning me hods de ec and classi y a basic se o sho classes such as o ehand and backhand (Pe ko ic e al., 2001). Mo eo e , ine ial measu emen uni (IMU) including accele ome e , gy oscopes, and magne ome e could be used o de ec he occu ence o sho s (Connaghan e al., 2010). Beyond academic esea ch, ew de ices ha e been made and ma ke ed o playe s and aine s. The mos p ominen a e ce ainly Babola Play and Zepp Tennis. An IMU wi h a wi eless ansmi ing de ice is a ached on he acke handle and da a is sen o a sma phone o a able o u he analysis. To ob ain a be e accu acy, some s udies used 3D mo ion analysis sys em and high speed came a o eco d he shu lecock eloci y (Domone e al., 2012; Huang e al., 2002; Jai ne & Wol , 2007; Lee, 1993; S ohmeye e al., 2009; Tsai e al., 2005, 2006, 2008; Tsai, Chang, e al., 2000; Tsai, Huang, e al., 2000). High speed came a allows ha he p ojec ile o e ol es in a plan, o a oid pa allax e o o measu emen , whe eas 3D mo ion analysis is ee o his kind o e o . This allows o ob ain an accu acy be ween 12 and 50 m.s-1, depending on he came a equency (120 and 500 ps in s udies) and i could be conside ed o he measu emen o eloci y as he gold s anda d. Howe e , hese de ices a e expensi e o coaches and pe sonal aine s and hei use is la gely con ined o Uni e si y labo a o ies and eli e spo s clubs (Balsalob e-Fe nándeza e al., 2015). Fu he mo e, some o hese ins umen s need speci ic compu e so wa e o analyse he da a. Howe e , o he bes o ou knowledge, he e a e no s udies alida ing a mo ion senso coupled wi h a sma phone applica ion o a ada o measu ing shu lecock eloci y. The aim o he p esen s udy, he e o e, was o analyse he alidi y and eliabili y o a speci ically ada (Dopple - ada gun) and a mo ion senso o IMU (named: Zepp Tennis) o measu ing p ojec ile eloci y, by compa ing wi h a ‘gold s anda d’ measu emen sys em, he Vicon high speed came a sys em. Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on 13 MATERIAL AND METHODS Pa icipan s Thi y- ou heal hy olun ee s (12 no ices, 11 in e media es and 11 expe s) ee o inju y (age = 20.1 ± 3.5 yea s; heigh = 1.75 ± 10.1 m; body mass = 69.2 ± 13.3 kg; aining expe ienced 8.3 ± 3.1 yea s) pa icipa ed in his s udy (Table 1). Thei skills we e e lec ed acco ding o hei yea o expe ience and a e labelled as ollowed: no ice (lowe han 1 yea ); in e media e (be ween 3 o 5 yea s o p ac ice) and expe s (mo e han 5 yea s o p ac ice). All pa icipan s we e physically heal hy, in good physical condi ion, and epo ed no inju ies du ing he ime o he s udy. They we e ully in o med abou he p o ocol be o e pa icipa ing in his s udy. In o med consen was ob ained p io o all es ing om all subjec s, in acco dance wi h he app o al o he local e hical commi ee and adhe ed o he la es amendmen s o he Decla a ion o Helsinki. The w i en in o med consen was ob ained om each pa icipan be o e expe imen . The sample was di ided on h ee g oups o ob ain di e en maximal shu lecock eloci y du ing a smash. The yea o p ac ice pe mi o ob ain di e en eloci y. Table 1. Age and an h opome ic cha ac e is ics o he h ee samples (mean and s anda d de ia ion). Va iables No ice In e media e Expe Age (yea s) 24.5 ± 7.6 21.1 ± 4.4 24.4 ± 8.1 Heigh (cm) 182.3 ± 7.2 179.9 ± 6.3 176.9 ± 9.7 Weigh (kg) 76.5 ± 9.8 72.8 ± 9.3 74.3 ± 1.7 T aining expe ienced (yea s) 0.2 ± 0.7 4.1 ± 1.3 10.6 ± 2.9 S udy design The pa icipan comple ed a gene al 10-min wa m- up composed o jogging, uppe body dynamic s e ches and s oke wi h he acke . Then, each pa icipan pe o med i e badmin on smash s okes in a a ge loca ed in on o him (2m x 2m). Du ing each ial, pa icipan s we e no in o med o hei pe o mance. Each smash s oke was sepa a ed by 30 sec passi e es pe iod. A shu lecock was suspended om he ceiling wi h a s ing a he playe ’s p e e ed hi ing heigh . When in con ac wi h he acke , he shu lecock is pulling away om he celling o p oduce he ajec o y o he smash. No pa icipan exp essed esidual a igue om p eceding p ocedu e. The expe imen ook place in wo sessions wi h a minimum o 2 sepa a ed days be ween each one. Badmin on smash s okes Pa icipan s pe o med badmin on smash s okes wi h he same acke (Wilson D aco Blx; heigh = 674 mm; weigh = 86 g; lexibili y = semi- igid; s ing ension = 10.5 kg). No ins uc ion was gi en o he pa icipan s on how o p oceed du ing a badmin on smash s oke. They we e only ins uc ed o s oke as ha d as possible. Equipmen Spo s ada . A Dopple - ada gun- S alke Spo sys em (Texas, Uni ed S a es) a a equency o 250 Hz and an accu acy claimed by he cons uc o o ± 0.027 m/s was used o measu e he p ojec ile eloci y. The ada pe mi ed o ob ain he maximal and he e olu ion o he eloci y du ing each ial. The expe imen e is loca ed 2 me e s behind he playe in he playe - a ge axis a app oxima ely 2m50 (Chelly & Denis, 2001). Mo ion senso . To eco d he shu lecock eloci y wi h Zepp Tennis, a moun was a ached o he handle o he acke and he senso was inse ed in o he moun . The applica ion was designed o analysing he eloci y o he acke and he eloci y po en ial o he p ojec ile. Zepp Tennis is a ailable on he Apps o e (Apple Inc., USA) and on Google play (Google Inc., USA). High speed came a e e ence. The high-speed came a eco ded by nine Vicon V8i mo ion cap u e sys em a a equency o 500 Hz (Vicon Peak, Ox o d, UK) in o de o measu e he p ojec ile eloci y. A e lec i e ma ke o 14 mm diame e was a ixed on he on o he shu lecock. The Vicon sys em was connec ed o a PC equipped wi h he so wa e o analyse and ob ain he maximal p ojec ile eloci y (Vicon Mo ion Sys em L d., UK). The ideo-based sys em is conside ed as he gold s anda d e e ence o es ablishing concu en alidi y o he eloci y. The ma e iel pe mi ed o compa e he e o o he measu e wi h he o he ma e ials. S a is ical analyses Se e al analyses we e conduc ed o de e mine he eliabili y and alidi y o badmin on smash s okes using he mo ion senso and he ada in he p esen s udy. To summa ise he da a om all pa icipan s and each ial, desc ip i e s a is ics we e ealised. All da a we e no mally dis ibu ed on he basis o Shapi o-Wilk es . i) Rela i e eliabili y is ela ed o he deg ee o which sys em main ain hei posi ion in a sample wi h epea ed measu emen s (A kinson & Ne ill, 1998). To analyse he es - e es eliabili y o bo h ins- umen s be ween ials o measu emen s, in a-class co ela ion coe icien (ICC) was pe o med. These coe icien s we e compu ed as [ICC = 1- (SEM/SD)²], whe e SEM is he s anda d e o in measu emen and SD is he mean be ween pa icipan SD o he ial ob ained by weighing he a iances on he basis o hei deg ees o eedom (Hopkins, 2000). The SEM was compu ed as [SEM = SD (be ween- ial di e ence in measu es) /√2]. To analyse he eliabili y o he mo ion senso and he ada when measu ing smash s oke o each pa icipan , he coe icien o a ia ion (CV) was used, on he basis o [CV = (SEM/mean)/100%], whe e In . j. acke spo s sci. ol. 4(2), 2022, 11-19. eISSN: 2695-4508 Michael Phomsoupha, Flo ie Bonneau & Guillaume La aye 14 he mean akes in o conside a ion all pa icipan s and bo h ials (A kinson & Ne ill, 1998; Hopkins, 2000). To de ec sys ema ic bias be ween ials, S uden ’s - es was pe o med (A kinson & Ne ill, 1998). ii) Da a se used concu en alidi y has no mal dis ibu ion. The di e ence be ween he ma e ials was examined using a one-way analysis o a iance. When a signi ican F- alue was ound (p < 0.05), he Bon e oni pos -hoc was applied (Cohen, 1988). In complemen , he bi a ia e Pea son p oduc momen co ela ion coe icien (R) was used. iii) Bland-Al man plo s we e c ea ed, which a e known o gi e a good ep esen a ion o he ag eemen be ween he h ee ins umen s (Bland & Al man, 1986). To quan i y he s a is ical dispe sion, a Whi e’s es was used (Whi e, 1980) o ob ain he le el o he e oscedas ici y. Concu en alidi y was assessed by compa ing he mean o ials pe o med a he maximal pe o mance be ween wo sys ems. Simila s a is ical measu es o hose used o assess eliabili y we e employed o concu en alidi y. Mo e p ecisely, we compu ed S uden ’s - es o pai ed samples (sys ema ic bias), ICC alues ( ela i e alidi y), be ween-sys em di e ences in means (absolu e alidi y in aw uni s and %) and CVs (absolu e alidi y in %). To ob ain a be e esul , he e o size and a maximal e o o 5% is conside ed o be accep able o a p ac ical applica ion compa ed o high-speed came as. On he basis o commonly used h esholds, he ela i e eliabili y and alidi y measu es we e conside ed poo , ai and good when he co esponding ICC alues we e <0.4; 0.4-0.75; >0.75 (Po ney & Wa kins, 2009). The absolu e eliabili y and alidi y o measu es we e conside ed adequa e when he co esponding CV alues we e equal o o lowe han 10% (S okes, 1985). All calcula ions we e pe o med using S a is ica 10 so wa e (S a So Inc., Tulsa, OK), Mic oso Excel 2010 (Mic oso Co p., Redmon , WA, USA) and so wa e R (www. -p ojec .o g). The in a-session e o is ee o me hodological e o s and may be conside ed as “in insic a ia ion” and se ed as an app op ia e baseline o compa isons, emaining independen o o he e o sou ces. In a- session eliabili y o p ojec ile pe o mance is c i ically impo an o ensu e ha obse ed di e ences be ween es ing ials, a e no due o sys ema ic bias, such as lea ning e ec , a igue, o andom e o due o possible biological o mechanical a ia ion. This a iabili y is usually caused by he emo ional s a e o he pa icipan s be ween he ials and hei le el o adap a ion wi h he measu ing sys em. RESULTS Tes - e es eliabili y The eloci y pa ame e s o each p ojec ile and he mean be ween- ial di e ence a e epo ed in Table 2 o he h ee sys ems. The be ween- ial di e ence in p ojec ile eloci y ac oss he le el (no ice, in e media e and expe ) we e 5.4 ± 3.7 m.s-1, 5.5 ± 2.8 m.s-1 and 4.8 ± 3.4 m.s-1 o he ada ; 6.6 ± 6.4, 4.9 ± 4.2 m.s-1 and 5.1 ± 3.4 m.s-1 o he mo ion senso and 5.2 ± 4.9 m.s-1, 4.9 ± 2.9 m.s-1 and 5.5 ± 3.0 m.s-1 o he high-speed came as. The ICC and CV alues speci ic o he eliabili y a he di e en es speeds a e epo ed in Table 2 o he h ee sys ems. The means o he ICCs was 0.907 ± 0.027 ( ange 0.88-0.94), 0.840 ± 0.054 ( ange 0.78-0.89) and 0.940 ± 0.018 ( ange 0.92-0.96) o he ada , he mo ion senso and he high-speed came as, espec i ely. Thei co esponding mean CV alues we e 5.8 ± 0.7 ( ange 5.3-6.6), 7.7 ± 1.8 ( ange 6.1-9.6) and 4.3 ± 1.0 ( ange 3.5-5.4). O e all, all h ee sys ems demons a ed a good ela i e and adequa e absolu e eliabili y o p ojec ile eloci y (Table 3). Table 2. P ojec ile eloci y s ide pa ame e s calcula ed using he ada , he mo ion senso and he high-speed came a sys ems. Rada Mo ion senso High speed came a Pa ame e (uni ) Session 1 Session 2 ∆ [%] Session 1 Session 2 ∆ (%) Session 1 Session 2 ∆ (%) Veloci y (m.s-1) No ice 34.9 ± 8.9 33.3 ± 7.8 4.6 33.6 ± 10.1 33.4 ± 8.4 0.8 37.5 ± 9.1 37.2 ± 7.5 0.9 In e media e 45.1 ± 6.2 46.2 ± 7.9 -2.6 47.9 ± 4.5 46.6 ± 7.7 2.7 48.1 ± 6.1 49.1 ± 8.9 -1.9 Expe 57.1 ± 9.9 57.9 ± 8.7 -1.5 53.8 ± 8.3 54.4 ± 7.3 -1.3 56.7 ± 9.5 60.4 ± 10.8 -1.2 No e: The mean ± SD o each ial and es eloci y, and he di e ence be ween ials (∆, in %) a e epo ed. Table 3. The ela i e (in a-class co ela ion coe icien , ICC) and absolu e (coe icien o a ia ion, CV) ep oducibili y o p ojec ile eloci y s ide pa ame e s calcula ed using he ada , he mo ion senso and he high-speed came a sys em. Rada Mo ion senso High speed came a Pa ame e (uni ) ICC CV (%) ICC CV (%) ICC CV (%) Veloci y (m.s-1) No ice 0.937 6.6 0.893 9.6 0.940 5.4 In e media e 0.884 5.5 0.785 7.5 0.921 4.1 Expe 0.901 5.3 0.842 6.1 0.957 3.5 No e: ICC < 0.75 and CV > 10% a e i alicised and ep esen ai ela i e ep oducibili y and less han adequa e absolu e ep oducibili y o measu es, espec i ely. Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on 15 Also, he Pea son p oduc momen co ela ion coe icien showed almos pe ec co ela ion be ween he ada and he high-speed came a measu emen s o eloci y ( = 0.917; p < 0.001); and good co ela ion be ween he mo ion senso and he high-speed came a measu emen s ( =0.682; p < 0.001) ( igu e 1 e 2) Co ela ion be ween ada and high-speed came a y = 0.9447x + 6.703 High-speed came a eloci y (in m/s) 0 020 20 40 40 60 60 80 80 100 100 120 Rada eloci y (in m/s) R2 linea = 0.9171 Figu e 1. Concu en alidi y be ween ada and high-speed came a. Co ela ion be ween mo ion senso and high-speed came a y = 0.8944x + 10.273 High-speed came a eloci y (in m/s) 0 020 20 40 40 60 60 80 80 100 120 Mo ion senso eloci y ( in m/s ) R2 linea = 0.6817 Figu e 2. Concu en alidi y be ween mo ion senso and high-speed came a. Validi y The mean o bo h ials is p esen ed in Table 3 o each sys em, as a e he di e ences be ween ada and mo ion senso and he e e ence sys em (high- speed came as). Rada and mo ion senso eco ded signi ican ly sho e eloci y compa ed o he high- speed came as (Table 4). In con as , le el pa ame e s eco ded using he ada and he mo ion senso showed no signi ican di e ences (all le els, P > 0.05). The ICC and CV alues desc ibing he concu en alidi y o p ojec ile eloci y calcula ed using he high-speed came a, agains he wo e e ence sys ems a e epo ed in Table 5. The absolu e (ICC) concu en alidi y o he ada and he mo ion senso was o e all ai o p ojec ile eloci y agains he high-speed came as (0.757 ± 0.207 and 0.866 ± 0.029 espec i ely). The co esponding ela i e (CV) concu en alidi y measu es we e highe han adequa e (12.5 ± 1.99 and 12.1 ± 0.46%). Rada and mo ion senso alues we e signi ican ly lowe han hose ob ained wi h he high-speed came a (p < 0.05) ( igu e 3 and 4). Table 4. P ojec ile eloci y s ide pa ame e s calcula ed using he ada , he mo ion senso and he high-speed came a sys ems. Pa ame e (uni s) Rada Mo ion senso High speed came a Rada s high speed came a Mo ion senso s high speed came a Veloci y (m.s1)∆ (%) No ice 32.8 ± 8.6 32.2 ± 9.4 37.4 ± 8.3 -14.1* -16.2** In e media e 44.3 ± 7.1 44.1 ± 6.1 48.6 ± 7.1 -9.8 -10.3* Expe 56.2 ± 9.4 52.8 ± 7.8 60.1 ± 10.1 -6.9 -13.6* No es: The mean ± SD o bo h ials combined a each es speed o each sys em, and he di e ences be ween he high-speed came a and he o he wo sys ems (∆ in %) a e epo ed. *P < 0.05; ** P< 0.01, signi ican di e ence be ween he high-speed came a and he ada o he mo ion senso using pai ed - es s. Table 5. The ela i e (in a-class co ela ion coe icien , ICC) and absolu e (coe icien o a ia ion, CV) concu en alidi y o p ojec ile eloci y calcula ed using he high-speed came a sys ems agains he ada and he mo ion senso . Rada s High speed Came a Mo ion senso s High speed Came a Pa ame e (uni s) ICC CV (%) ICC CV (%) Veloci y (m.s-1) No ice 0.911 24.8*** 0.627 26.5*** In e media e 0.841 16.7*** 0.303 14.9*** Expe 0.825 16.8*** 0.295 17.2*** No e: ICC < 0.75 and CV > 10% a e i alicised and ep esen ai ela i e ep oducibili y and less han adequa e absolu e ep oducibili y o measu es, espec i ely. *** P< 0.001, signi ican di e ence be ween he high-speed came a and he ada o he mo ion senso using pai ed - es s. DISCUSSION The pu pose o his s udy was o analyse he concu en alidi y and eliabili y o a ada (Dopple - ada gun) and a mo ion senso (Zepp Tennis). The ada and mo ion senso we e eliable, bu ada and mo ion senso unde es ima ed eloci y compa ed o high-speed came as. Hence, ada and mo ion senso can be conside ed as a eliable sys em o compu ing p ojec ile eloci y du ing a badmin on smash s oke anging om no ices o expe s. Howe e , mo ion senso did no demons a e good concu en alidi y o each le el measu es and only o no ice o ada , wa an ing cau ion agains he compa isons o esul s be ween he high-speed came as and ada . In e media e and expe le el ob ained om he ada p o ed o be highly eliable and alid compa ed o ou e ence sys ems. In . j. acke spo s sci. ol. 4(2), 2022, 11-19. eISSN: 2695-4508 Michael Phomsoupha, Flo ie Bonneau & Guillaume La aye 16 -15-1.96 SD +1.96 SD Veloci y ada / high-speed came a Mean o he wo ins umen s (m/s) Di e ences be ween he wo ins umen s ( m/s ) Biais -10 -5 0 5 10 Figu e 3. Bland-Al man plo s o ada and high-speed came a eloci y da a. The cen al line ep esen s he absolu e a e age di e ence be ween ins umen s, while he uppe and he lowe lines ep esen ±1.96 SD. -15 -1.96 SD +1.96 SD Veloci y mo ion senso / high-speed came a Mean o he wo ins umen s (m/s) Di e ences be ween he wo ins umen s ( m/s ) Biais -10 -25 -20 -5 0 5 10 15 20 25 Figu e 4. Bland-Al man plo s o mo ion senso and high-speed came a eloci y da a. The cen al line ep esen s he absolu e a e age di e ence be ween ins umen s, while he uppe and he lowe lines ep esen ±1.96 SD. The wo sys ems demons a ed a good homoscedas ici y wi h no signi ican di e ences wi h he Whi e’s es on he quan i ica ion o he s a is ical dispe sion o ada and o mo ion senso (bo h de ices, p < 0.05). The use o se e al s a is ical pa ame e s is ecommended o quan i ying he eliabili y o measu es (A kinson & Ne ill, 1998). In his s udy, all s a is ical indica o s implied high eliabili y o eloci y de i ed om he h ee di e en sys ems. The ine dis inc ion in p ojec ile eloci y and echnique be ween ials is no mal a ia ions expec ed in any es ing si ua ion. Hence he impo ance o es ablishing he ep oducibili y o measu es de e mine which di e ences exceed ypical a ia ions in pe o mance (Gind e e al., 2016). O he han measu emen noise, indi idual a ia ions in s oke biomechanics in he a m mo emen con ibu ed o he impe ec eliabili y o measu es analysed in he h ee sys ems. Kinema ics mo emen s o he a m a e di e en and depending o he le el ha p oduced di e en shu lecock eloci y (Phomsoupha & La aye, 2020) and con ibu ed o he impe ec eliabili y o measu es in all h ee sys ems. The di e ences in s oke echnique be ween each ial a e no mal a ia ions expec ed in any es ing si ua ion. The impo ance o es ablishing he ep oducibili y o es measu es o de e mine which di e ences exceed ypical a ia ions in pe o mance (Gind e e al., 2016). Fo ada speci ically, he be ween- ial ICCs we e all abo e 0.880; and CVs below 10%. These esul s end o highligh ha ada eco ded he maximal speed du ing each s oke. Ou indica o s o eliabili y o ada de ice pa ame e s end o show ha he p ojec ile gene a ed highe maximal eloci y han Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on 17 acke head (Rambely e al., 2005). In addi ion, hese esul s end o highligh ha he posi ioning o he ada (i.e. in on o he p ojec ile and iden ic angula ion) subs an ially a ec he ep oducibili y o ada pa ame e s (Robinson & Robinson, 2016). Fo mo ion senso speci ically, he be ween- ial ICCs we e all abo e 0.780; and CVs below 10%. These esul s show ha his de ice has a good eliabili y. Mo eo e , nei he he posi ioning o he de ice (i.e handle o w is ) no he di ec ion o complexi y o he mo emen (i.e. o ea m ex ension and adio-ulna p ona ion) subs an ially a ec he ep oducibili y o accele ome e de i ed pa ame e s. The con i ma ion o he ep oducibili y o eloci y pa ame e s du ing smash s oke using ada and mo ion senso allowed p ac i ione s o be con iden in hei abili y o eco d hese speeds o e ime using hei de ice. The concu en alidi y o p ojec ile eloci y be ween ada / high-speed came as and mo ion senso / high-speed came as was ai . On a e age, ada unde es ima ed p ojec ile eloci y by -9.7% and mo ion senso by -13.6% be ween high-speed came as. These esul s showed a di e en way be ween-sys em di e ence in cap u e eloci y and ea men me hods. On one hand, he mo ion senso consis s o an ex apola ion o he eloci y by he in eg a ion o mula o he accele a ion on h ee dimensions and he addi ion o he le e a m. The wo majo d awbacks o he mo ion senso a e ha he e is a possibili y ha ew p ojec ile eloci ies will no be eco ded and he lack o conside a ion o he acke de lec ion. The dynamics ob ained wi h a de lexion coupled wi h a high accele a ion o he w is con ibu e o acke head eloci y (Phomsoupha e al., 2015). G ea e lexibili y inc eases he capaci y o he acke o s o e and elease mo e s ain ene gy and o inc ease he p ojec ile eloci y. In addi ion, he mo ion senso seems unable o cap u e a p ojec ile eloci y highe han 325 km/h (≈ 95 m/s). This ma e ial is able o measu e p ojec ile eloci y accu a ely o all popula ions, including ained a hle es bu his is no possible o expe s and high speed. The e is no equi emen and any expe ience o use and o analyse he da a om mo ion senso . On he o he hand, he ada consis s o bo h a ecei e and a ansmi e . I sends a adio wa e ha is e lec ed o by any objec ha is in he pa h. To calcula e he speed, he ada ge s he echo and uses he p inciple o Dopple shi . Howe e , he majo d awback is he il on he sagi al plane which could eco d he acke ins ead o he p ojec ile. To ensu e be e o ming esul s, he ada gun should be posi ioned nea he pa icipan (1 me e a shu lecock heigh du ing he s oke) and he expe imen e ha e o be ca e ul wi h he ecommenda ion o he manu ac u e abou he ield o angle accu acy. Du ing he expe imen a ion, a ound 20% o he p ojec ile eloci y was no epo ed. Clinicians and scien is mus be awa e o hese be ween-sys em de ia ions, pa icula ly when compa ing esul s om di e en s udies, labo a o y o clinics, and acqui ing new equipmen o he pu pose o quan i ying p ojec ile eloci y. The da a o p ojec ile we e homoscedas ici y be ween ada and high-speed came as and be ween mo ion senso and high-speed came as. Conside ing he low eliabili y wi h a low alidi y o he ada and he mo ion senso o measu e shu lecock eloci y, co ec ion ac o s o alid absolu e alues and o acili a e c oss-s udy compa isons o esul s may be p oposed. Linea eg ession analyses on ou da a sugges s using he ollowing equa ion o ob ain eloci y (x ) om he high-speed came a ha a e compa able o hose om he ada (0.945x + 6.703) and he mo ion senso (0.894x + 10.273) when indi iduals eloci y anging om no ices o expe s ( igu e 1 and 2). When analysing he eliabili y o he mo ion senso and ada o measu ing he p ojec ile eloci y o each pa icipan , he esul s showed alues ha we e close o he ones ob ained wi h he high-speed came a, despi e di e ences be ween de ices in sampling equency. Fu he mo e, he ada and he mo ion senso da a showed in Bland-Al man plo s (Figu e 3 and 4) ha se e al o he p ojec ile eloci ies we e close o he mean o he high-speed came a. This is ep esen ing a low le el o conco dance eloci ies be ween mo ion senso and high-speed came as (Bland & Al man, 1986). The high ICC showed ha mo ion senso is no eliabili y and he esul s should be empe ed and could be inc eased accu acy wi h he linea eg ession. The e a e no p e ious s udies ha compa ed di e en echnology o measu ing p ojec ile eloci y wi h high-speed came a da a. Howe e , some s udies used he high-speed came a, which seem o be he bes way o eco d and analyse p ojec ile eloci y. This allows o ob ain speci ic alues abou he highes eloci y du ing he s oke and he ime o equi e i . Fo he momen , he mos accu a e sys ems o measu ing p ojec ile eloci y a e p o essional and labo a o y high speed came as. This ype o came a pe mi s o eco d a 500 o 1000 Hz compa ed o 60 Hz o comme cial came a. The isk wi h a s anda d came a was he maximal eloci y could be no eco ded du ing he impac . Thus, expe imen al da a could miss he highe alues. Ne e heless, he ad ancemen o new echnologies will pe mi o in eg a e highe eco ding equencies in he u u e on s anda d came a (Balsalob e-Fe nándeza e al., 2015). Thus, he o ien a ion has a g ea impac o he pe o mance da a. Howe e , an expe ience in he use o he ada was equi ed in o de o eco d he co ec eloci y. This is he i s s udy ha alida es a mo ion senso and ada o measu ing he p ojec ile eloci y. In . j. acke spo s sci. ol. 4(2), 2022, 11-19. eISSN: 2695-4508 Michael Phomsoupha, Flo ie Bonneau & Guillaume La aye 18 CONCLUSION The abili y o e alua e and moni o p ojec ile eloci y abili y is impo an in a eas o alen iden i ica ion and spo ing pe o mance. The esul s o he p esen s udy showed ha p ojec ile eloci y can be e alua ed using wo ins umen s. Mo ion senso could be o ien ed o he acke spo s ( ennis, squash…) and gol . Rada was also mo e e icien o he h owing p ojec ile spo s (baseball, oo ball, olleyball…). These indings could help coaches and aine s who wish o moni o he p ojec ile eloci y abili y o hei a hle es o clien s in a alid and economic way wi h some ideas o he limi a ion o each ma e ial. CONFLICT OF INTEREST All au ho s ha e decla ed he e is no any po en ial con lic o in e es s conce ning his a icle. REFERENCES A kinson, G., & Ne ill, A. M. (1998). S a is ical me hods o assessing measu emen e o ( eliabili y) in a iables ele an o spo s medicine. Spo s Medicine, 26(4), 217-238. Balsalob e-Fe nándeza, C., Glais e b, M., & Lockeyb, R. A. (2015). The alidi y and eliabili y o an iPhone app o measu ing e ical jump pe o mance. Jou nal o Spo s Science, 33(15), 1574-1579. Bland, J. M., & Al man, D. G. (1986). S a is ical me hods o assessing ag eemen be ween wo me hods o clinical measu emen . The Lance , 1(8476), 307-310. Chelly, S. M., & Denis, C. (2001). Leg powe and hopping s i ness: ela ionship wi h sp in unning pe o mance. Medicine and Science in Spo s and Exe cise, 33(2), 326- 333. Cohen, J. (1988). S a is ical powe analysis o he beha io sciences (Law ence E). Connaghan, D., Conai e, C., Kelly, P., & O’Conno , N. (2010). Recogni ion o ennis s okes using key pos u es. Signals and Sys ems Con e ence, 245-248. Coyle, S., Mo is, D., Lau, K., Diamond, D., & Moyna, N. (2009). Tex ilebased wea able senso s o assis ing spo s pe o mance. Wea able and Implan able Body Senso Ne wo ks, 307-311. Domone, S., Whea , J., Choppin, S., Hamil on, N., & Helle , B. (2012). Wa ele based de-noising o non s a iona y kinema ic signals. 30 h Annual Con e ence o Biomechanics in Spo s, 44, 376-379. F ench, A. (1968). Special Rela i i y. Gind e, C., Lussiana, T., Hebe -Losie , K., & Mo in, J. (2016). Reliabili y and alidi y o he Myo es ® o measu ing unning s ide kinema ics. Jou nal o Spo s Science, 34(7), 664-670. Halliday, D., Resnick, R., & Walke , J. (2011). Fundamen als o Physics (9 h edn). Hopkins, W. G. (2000). Measu es o eliabili y in spo s medicine and science. Spo s Medicine, 30, 1-15. Huang, K. S., Shaw-Shiun, C., & Tsai, C. L. (2002). Kinema ic analysis o h ee di e en badmin on backhand o e head s okes. In e na ional Symposium on Biomechanics in Spo s, 200-202. Hussain, I., & A shad Ba i, M. (2011). Kinema ical analysis o o ehand and backhand smash in badmin on. Inno a i e Sys ems Design and Enginee ing, 2(7), 20- 26. Jai ne , T., & Wol , G. (2007). Analysis o badmin on smash wi h a mobile measu e de ice based on accele ome y. XXV ISBA Symposium Vol 20, 282-284. La aye, G., Debanne, T., & Choukou, M. A. (2012). Is he ball eloci y dependen on expe ise? A mul idimensional s udy in handball. In e na ional Jou nal o Pe o mance Analysis in Spo , 12(3), 629-642. La aye, G., Phomsoupha, M., & Do , F. (2015). Changes in he game cha ac e is ics o a badmin on ma ch: a longi udinal s udy h ough he Olympic Game inals analysis in men’s singles. Jou nal o Spo s Science and Medicine, 14, 584-590. La aye, G., Wagne , P., & Tombleson, T. (2014). Coun e mo emen jump heigh : Gende and spo - speci ic di e ences in he o ce- ime a iables. The Jou nal o S eng h & Condi ioning Resea ch, 28, 1096- 1105. Lee, B. K. (1993). The e ec s o he kinema ic link p inciple on pe o mance. 11 In e na ional Symposium on Biomechanics in Spo s, 239-242. Mille , S. (2006). Mode n ennis acke s, balls, and su aces. B i ish Jou nal o Spo s Medicine, 40(5), 401-405. Nagasawa, M., Ha o i, Y., Kaku a, M., Hayashi, T., & Sekine, Y. (2012). Smash mo ion analysis o badmin on om image. P oceedings o he IEEJ Image Elec onics and Visual Compu ing Wo kshop, 1-8. Pe ko ic, M., Jonke , W., & Zi ko ic, Z. (2001). Recognizing S okes in Tennis Videos Using Hidden Ma ko Models. IASTED In e na ional Con e ence on Visualiza ion, Imaging and Image P ocessing, 512-516. Phomsoupha, M., & La aye, G. (2014). Shu lecock eloci y du ing a smash s oke in badmin on e ol es linea ly wi h skill le el. Compu e Me hods in Biomechanics and Biomedical Enginee ing, 17(Suppl 1), 140-141. Phomsoupha, M., & La aye, G. (2015). The science o badmin on: game cha ac e is ics, an h opome y, physiology, isual i ness and biomechanics. Spo s Medicine, 45(4), 473-495. Phomsoupha, M., & La aye, G. (2020). A mul iple epea ed sp in abili y es wi h ou changes o di ec ion o badmin on playe s (pa 2): p edic ing skill le el wi h an h opome y, s eng h, shu lecock Reliabili y and alidi y o mo ion senso and ada o measu ing shu lecock eloci y in badmin on 19 and displacemen eloci y. Jou nal o S eng h and Condi ioning Resea ch, 34(1), 203-211. Phomsoupha, M., La aye, G., Cohen, C., & Clane , C. (2015). How o use he elas ici y o a badmin on acke o inc ease i s speed by 80%? Compu e Me hods in Biomechanics and Biomedical Enginee ing, 18. Po ney, L. G., & Wa kins, M. P. (2009). Founda ions o clinical esea ch: applica ions o p ac ice (Uppe Sadd). N.J: Pea son/P en ice Hall. Rambely, A. S., Osman, N. A. A., Usman, J., & Wan Abas, W. A. B. (2005). The con ibu ion o uppe limb join s in he de elopmen o acke eloci y in he badmin on smash. 23 In e na ional Symposium on Biomechanics in Spo s, 422-426. Resnick, R. (1968). In oduc ion o Special Rela i i y. Robinson, G., & Robinson, I. (2016). Rada speed gun ue eloci y measu emen s o spo s-balls in ligh : applica ion o ennis. Physica Sc ip a, 91(2), 1-19. S okes, M. (1985). Reliabili y and epea abili y o me hods o measu ing muscle in physio he apy. Physio he apy Theo y and P ac ice, 1(2), 71-76. S ohmeye , H. S., A ms ong, C., Li insky, Y., Nooney, R., Moo e, J., & Smi h, K. (2009). In e segmen al coo dina ion di e ences be ween beginning pe o me s execu ing a badmin on smash o accu acy o eloci y. 27 In e na ional Con e ence on Biomechanics in Spo s. Takahashi, H., & Kawaha a, T. (2011). Condi ioning in mul i- suppo p ojec . The Jou nal o Japanese Socie y o Clinical Spo s Medicine, 19(2), 195. Tsai, C. L., Chang, S. S., & Huang, C. (2000). Biomechanical analysis o di e ences in he badmin on smash and jump smash be ween aiwan eli e and collegia e playe s. ISBS’98 - P oceedings II, 259-262. Tsai, C. L., Hsueh, Y. C., Pan, K. M., & Chang, S. S. (2008). Biomechanical analysis o di e en badmin on o ehand o e head s okes o Taiwan eli e emal playe s. 26 In e na ional Symposium on Biomechanics in Spo s, Con e ence, Coaching and Spo s Pe o mance, 719-722. Tsai, C. L., Huang, C., Lin, D. C., Cheng, C. C., & Lai, C. M. (2000). Biomechanical analysis o he uppe ex emi y in h ee di e en badmin on o e head s okes. 18 h In e na ional Symposium on Biomechanics in Spo s, 35-38. Tsai, C. L., Yang, C. C., Lin, M. S., & Huang, K. S. (2005). The su ace EMG ac i i y analysis be ween badmin on smash and jump smash. In e na ional Symposium on Biomechanics in Spo s, 483-486. Tsai, C. L., Yang, C. C., Lin, M. S., Huang, K. S., & Chang, S. S. (2006). The su ace EMG ac i i y o he uppe limb muscles o badmin on o ehand and backhand smashes. 24 In e na ional Symposium on Biomechanics in Spo s, 3-6. Whi e, H. (1980). A he e oskedas ici y-consis en co a iance ma ix es ima o and a di ec es o he e oskedas ici y. Econome ica, 48(4), 817-838. Yonex. (2013). Nano ay Z speed. h ps://www.yonex.com/ nano ay/# op