Improving the energy efficiency for the WBSN bottleneck zone based on random linear network coding

Transcription

1 Improvig he eergy efficiecy for he WBSN boleeck zoe based o radom liear ework codig Al Shahee, H ad Takruri Rizk, H hp://dx.doi.org/ /ie wss Tile Auhors Type URL Improvig he eergy efficiecy for he WBSN boleeck zoe based o radom liear ework codig Al Shahee, H ad Takruri Rizk, H Aricle Published Dae 2017 This versio is available a: hp://usir.salford.ac.uk/43904/ USIR is a digial collecio of he research oupu of he Uiversiy of Salford. Where copyrigh permis, full ex maerial held i he reposiory is made freely available olie ad ca be read, dowloaded ad copied for o commercial privae sudy or research purposes. Please check he mauscrip for ay furher copyrigh resricios. For more iformaio, icludig our policy ad submissio procedure, please coac he Reposiory Team a: usir@salford.ac.uk.

2 This paper is a pospri of a paper submied o ad acceped for publicaio i IET Wireless Sesor Sysems ad is subjec o Isiuio of Egieerig ad Techology Copyrigh. Whe he fial versio is published, he copy of record will be available a he IET Digial Library Submissio Templae for IET Research Joural Papers Improvig he Eergy Efficiecy for he WBSN Boleeck Zoe based o Radom Liear Nework Codig Hisham Alshahee 1*, Haifa Takruri Rizk 2 1 Compuig, Sciece ad Egieerig School, Uiversiy of Salford, Macheser, UK 2 Compuig, Sciece ad Egieerig School, Uiversiy of Salford, Macheser, UK * h.s.s.alshahee@edu.salford.ac.uk Absrac: The reducio of eergy cosumpio ad he successful delivery of daa are impora for he Wireless Body Sesor Nework (WBSN). May sudies have bee proposed o improve eergy efficiecy, bu mos of hem have o focussed o he biosesor odes i he WBSN boleeck zoe. Eergy cosumpio is a criical issue i WBSNs, as he odes ha are placed ex o he sik ode cosume more eergy. All biomedical packes are aggregaed hrough hese odes formig a boleeck zoe. This paper proposes a ovel mahemaical model for body area ework (BAN) opology o explai he deployme ad coecio bewee biosesor odes, simple relay odes, ework codig relay odes ad he sik ode. Therefore, his paper is dedicaed o researchig boh he eergy savig ad delivery of daa if here is a failure i oe of he liks of he rasmissio, which relaes o he proposed Radom Liear Nework Codig (RLNC) model i he WBSN. Usig a ovel mahemaical model for a WBSN, i is appare ha eergy cosumpio is reduced ad daa delivery achieved wih he proposed mechaism. This paper deails he sages of he research work. 1. Iroducio A Wireless Body Sesor Nework (WBSN) cosiss of several biological sesors. WBANs are used i boh medical ad o-medical applicaios. I medical applicaio, hese are desiged o moior vial sigs i he huma body, ad WBSNs are used o show differe vial sigs, for isace blood glucose level, hear rae, body emperaure, body posure ad blood oxyge sauraio. I addiio, WBSNs are used o measure Elecrocardiography (ECG), blood pressure, ad Elecroecephalography (EEG). Such sesor devices fall io wo ypes: implaable medical devices ad wearable medical devices. The former are implaed iside he huma body. The laer are wor or placed o he ski or very ear o he paie. The biosesor odes sese or oherwise measure he differe healh sigs of he body; he medical daa is he rasmied o he sik ode. The sik ode is siuaed o he huma body or a a earby locaio [1]. There is some eergy wasage i he boleeck zoe by he odes ha are placed ear he sik ode, which cosume more eergy as hey are required o forward daa from odes ouside he boleeck zoe. For his reaso, he proposed desig of a mahemaical model for he body area ework aemps o explai 1

3 he deployme ad coecio of odes. Furhermore, i his paper, we cosider a echique o reduce eergy usage for he biosesor odes i he WBSN boleeck zoe. This paper coribues a ovel mahemaical model desig for a body area ework (BAN) opology which cosiders he coecio ad relaioship for he biomedical sesor odes, he simple relay odes, he ework codig relay odes ad he sik ode. I his model, ework codig echique is isalled i a sample of he relay odes o creae he ework codig relay ode iside he boleeck zoe. Moreover, applyig he ew algorihm o he odes i he boleeck zoe of he WBSNs will improve eergy cosumpio of biosesor odes ad guaraees beer daa delivery. The paper is orgaised as follows. Secio 2 reviews he relaed work. Secio 3 preses he eergy cosumpio model. Secio 4 shows pah loss model for he body. Secio 5 describes he body area ework model desig. Secio 6 preses he proposed desig for RLNC. Secio 7 discusses wireless body sesor ework performace. Fially, coclusios are draw i Secio Relaed work The ework codig (NC) cocep was firs iroduced by Ahlswede e al. [2]. NC has become oe of he popular research areas i pracical eworkig sysems. I is a echique ha iegraes differe ses of daa a relay odes i such a way ha hey ca be decoded a he desiaio. This echique ca lead o beer hroughpu of he ework [2] [3]. Based o he NC cocep, isead of simply forwardig he received packes, he iermediae ode will combie hem ad creae oe or several oupu packes [3]. NC also ehaces he reliabiliy of rasmissio i WSN by reducig he umber of los packes [4] as i has bee used for various applicaios [5]. May sudies used NC o achieve he ehaceme of eergy efficiecy. I [6], Plaz e al. showed ha deermiisic NC achieve a reducio i eergy usage. However, i [7], Fude e al. proposed a mechaism o achieve he miimizaio of eergy usage, he improveme of he hroughpu ad packe delivery raio based o icludig he duy cycle, ework codig, ad cluser head selecio. Clearly, i he muli-hop body area sesor eworks (BASNs), he auhors proposed a disace-aware relayig eergy efficie proocol o decrease he eergy usage ad prolog he ework lifeime bu he umber of paies was oly eigh [8]. Movassaghi proposed a ovel NC echique o improve eergy usage i WBSNs[9]. I [10, 11], he auhors demosrae a improveme i he eergy efficiecy based o Cooperaive commuicaio echiques for WBSNs. I addiio, i [12], he auhors propose he cooperaive MAC proocol, which is ermed CooPNC (cooperaive physical ework codig) for shor rage eworks. The characerisics of PNC is exploied i order o improve he eergy wih high reliabiliy. 2

4 I WBAN, he researchers applied NC a he relays o icrease he delivery of daa, which also reduced he eergy cosumpio [13]. I a similar way, i [14], he auhors used a reduced cooperaive NC o reduce he packe loss ad icrease he recovered daa, bu he odes were fixed. Furhermore, i [15], researchers proposed he decode ad forward NC echique o miimise he umber of rasmissios per ode for WBANs while icreasig eergy efficiecy. I was assumed he addiio of he relay odes o he ework would coiue uil a leas oe relay ode was i he lie of sigh for he biosesor odes ad relay odes. However, he icreased umber of relay odes has a impac o he mobiliy of he paie [16]. I [17] he auhors proposed Disribued Queuig Body Area Nework (DQBAN) MAC proocol depedig o fuzzylogic rules for he biomedical sesor ework. DQBAN proocol uilises a ovel cross-layer fuzzy logic, which is implemeed i he biosesor ode i order o reduce eergy cosumpio, improve he required reliabiliy, ad produce a beer qualiy of service i healhcare applicaios. I [18], D Adreagiovai & Nardi proposed he mahemaical opimizaio model for BAN o solve he uceraiy raffic i he BAN by usig relay odes. The previous works improved he eergy efficiecy i WBSNs by addig relay odes. However, mos of he previous works did o cosider he ework codig relay odes i he boleeck zoe. I his paper, we propose a echique o miimize eergy usage for biosesor odes i he WBSNs boleeck zoe. Furhermore, i he proposed echique, we use a special case of ework codig family, called RLNC (Radom Liear Nework Codig) o ecode he biomedical packes o improve he eergy efficiecy. 3. The Eergy Cosumpio Model The eergy cosumpio for commuicaio i he body sesor ework are cosidered i his paper which are he rasmissio eergy ad he recepio eergy as (1) ad (2) respecively [19]. E x (k bio, d) = E TXelec. k bio + E amp (). k bio. d (1) E rx (k bio ) = E RXelec. k bio (2) I he former formula, he E x represes he rasmissio eergy, E TXelec represes he dissipaed radio eergy o ru he circui for rasmissio, k bio represes he umber of rasmied biomedical bis, E amp represes he eergy cosumpio for he rasmier amplifier, fially is he pah loss coefficie. The laer formula for recepio eergy, E rx represes he recepio eergy ad E RXelec represes he dissipaed radio eergy of recepio circui. Toal eergy cosumpio of rasmiig ad receivig for he ode is give as: E xr (k bio, d) = E x (k bio, d) + E rx (k bio ) (3) 3

5 Nordic RF2401 has low power cosumpio, i is operaed i GHz, ad commoly used WSNs [19]. Table 1 shows he values of he specific parameer for Nordic RF2401 [16]. Table 1 Shows he values of he specific parameer for Nordic RF2401 [16] Parameer RF2401 Parameer RF2401 E TXelec 16.7 J/bi E amp(3.38) 1.97e-9 J/bi E RXelec 36.1 J/bi E amp(5.9) 7.99e-6 J/bi 4. Pah Loss model for he body This model is a fucio of he disace bewee he rasmiig ad he receivig aea [20].I is measured by (4).There are wo ypes of he propagaio model i he WBSN: he lie of sigh (LOS) ad he o-lie of sigh (NLOS) propagaio. The former applies o propagaio alog he fro of he orso. I was ivesigaed i [21], however i did o cosider he commuicaio bewee he orso ad back. The laer propagaio is a higher pah loss aroud he orso [22]. The semi-empirical formula is used for boh models as follows: P db = P 0,dB log ( d d 0 ) (4) I he above formula, P0,dB represes he pah loss a a referece disace d0 ad represes he pah loss expoe, i is equal o 2 i free space[14]. Table 2 demosraes he wo differe propagaio models of pah loss accordig o (4). The pah loss coefficie () for a lie of sigh ad o-lie of sigh are cosidered as 3.38, ad 5.9 respecively [16]. Table 2 The pah loss model: The values of parameers Parameer Value LOS[21] Value NLOS[22] d0 10 cm 10 cm P0,dB 35.7 db 48.8 db σ 6.2 db 5.0 db

6 5. The Body Area Nework (BAN) model desig A model of a BAN is usually represeed by a direced graph G (V, A) based o graph heory[23].there is a se of verices V ha icludes oe eleme for each wireless device (biosesor ode, simple relay ode, ework codig relay ode or sik ode) of he ework. The A represes a umber of liks (arcs). However, i [18], he auhors explaied he relaioship bewee biosesor odes, relay odes ad sik ode oly wihou ework codig relay odes. Wih respec o he se of verices, he se V is he uio of four disjoied ses of verices as follows: he se B of verices refers o biosesor odes, he se Rr of verices refers o he deployme of relay odes, he se Rc of verices refers o ework codig relay ode, ad he se S of verices refers o sik odes (assumpio he umber of sik ode i his model is oe). V = B b R r R c S (7) Each ode is a device siuaed wihi a rage depedig upo he power of he rasmiig device. Each biosesor geeraes he daa which is roued from a source ode b o a desiaio ode s (sik ode) i he ework is represeed by he graph G (V, A). Moreover, he capaciy of relay ode i a WBSN is (0 < cap r 250 kbps) for each relay r R. The rasmissio lik from he bio-medical sesor ode o he sik ode hrough a simple relay ode ad a ework code relay ode ca be expressed as A = A B S A B Rr A B Rc A Rr Rr A Rr S A Rr Rc A Rc S (8) 5.1 Eergy cosumpio assumpios of he desiged model The followig are he coeciviy parameers for WBAN of he desiged model represeig he relaioship bewee biosesor odes, simple relay odes, ework codig relay odes ad he sik ode. If he biosesor ode b ca esablish a lik wih he simple relay ode r, i ca be expressed as a br ij = { 1 a lik o arc(i, j) A B Rr 0 Oherwise (9) Someimes, i WBSN opology, he biosesor ode is coeced o more ha he simple relay ode. I his case a br ij = hr where hr is he umber of relay odes ha receive he packes from biosesor b. If he biosesor ode b ca esablish a lik wih he NC ode, i is represeed as a bc ij = { 1 Lik o arc (i, j) A B Rc 0 Oherwise (10) If he biosesor ode b ca esablish a lik wih he sik ode, i follows as a bs ij = { 1 a lik o arc (i, j) A B S 0 oherwise (11) 5

7 If he simple relay ode r ca esablish a lik wih he sik ode s, i ca be expressed as e ij rs = { 1 lik o arc (i, j) A Rr S 0 oherwise (12) If he simple relay ode r ca esablish a lik wih he NC ode, i is represeed as e ij rc = { 1 lik o arc(i, j) A Rr Rc 0 Oherwise If he simple relay ode r ca esablish a lik wih aoher simple relay ode l, i ca be represeed as e ij rl = { 1 lik o arc arc (i, j) A Rr Rr 0 Oherwise If he NC relay ode ca esablish a lik wih he sik ode s, i is expressed as e ij cs = { 1 lik o arc (i, j) A Rc S 0 Oherwise There are wo biary decisio variables: he former relaes o he daa geeraed from he biosesor ode while he laer is a decisio variable of he isallaio of he ework codig echique i he simple relay ode o creae he NC relay ode i he ework [24]. Biary geeraed daa variable is x ij bs {0,1} b B, s S, (i, j) A, ad he biosesor ode geeraed daa rasmied o he sik odeca be expressed as x bs 1 lik o arc (i, j) A ij = { 0 Oherwise The biary NC relay ode deployme variable z c {0,1} c Rc is represeed as (13) (14) (15) (16) z c = { 1 if isall NC i relay ode 0 oherwise (17) The oal rasmissio ad recepio eergy for all wireless odes i he WBSNs represes he oal eergy cosumpio. The calculaio of he oal eergy cosumpio o rasmi medical daa from all biosesor odes o he relay odes is give as E TXbr = k bs bio x bs ij a ij b B,r R,s S br (E TXelec + E amp ( br )) D br br (18) The simple relay odes receive he medical daa from he biosesor odes. The oal eergy cosumpio for recepio is compued by (19). E RXbr = k bs bio x bs ij a br ij E RXelec b B,r R,s S (19) 6

8 Table 4 deails all erms used i his model. The simple relay odes cosume eergy o forward he medical packes o aoher relay ode l as follows as = f rl s (E TXelec + E amp ( rl )) E TXrl r,l R,s S D rl rl + E RXelec (1) The oal eergy cosumpio o relay medical daa from simple relay odes o he sik ode is give = f rs s (E TXelec + E amp ( rs )) E TXrs r R,s S 5.2 Eergy cosumpio wih Nework Codig D rs rs + E RXelec (21) Wih respec o he coecio bewee ework codig relay odes ad bio-medical sesor odes wih simple relay odes. The calculaio of he oal eergy cosumpio o rasmi medical daa from all biosesor odes o he ework codig relay odes is give as E TXbc = k bs bio x bs ij a ij b B,c NC,s S bc (E TXelec + E amp ( bc )) D bc bc (22) The NC relay odes receive he medical daa from he biosesor odes. The oal eergy cosumpio for recepio is compued by (23). E RXbc = k bs bio x bs ij a bc ij E RXelec b B,c NC,s S (23) The simple relay odes cosume eergy o forward he medical packes o he ework codig relay odes which ca be expressed as by (25) E TXrc = f rc r R,c NC,s S s (E TXelec + E amp ( rc )) D rc rc + E RXelec (24) The oal eergy cosumpio o relay medical daa from he NC relay odes o he sik ode is give E TXcs = f cs c NC,s S s (E TXelec + E amp ( cs )) D cs cs + E RXelec (25) 7

9 Table 3 Explaaio of all erms uses i he model for WBSN erm Descripio The raffic geeraed by he biosesor odes b owards he sik ode S The disace bewee he biosesor odes ad he simple relay odes The oal raffic rasmied by he simple relay ode o eighbourig ode (aoher relay ode). The disace bewee he simple relay odes ad he eighbourig ode (aoher relay ode). Traffic from he simple relay ode o he sik ode The disace bewee he simple relay odes ad he sik ode. The disace bewee he biosesor odes ad he NC relay odes s The oal raffic rasmied by he simple relay ode o he NC relay ode oward he sik ode The disace bewee he simple relay odes ad he NC relay ode. s The oal raffic received from he NC relay ode o he simple relay ode oward he sik ode. The disace bewee he NC relay odes ad he sik ode. f bs D br br s f rl D rl rl s f rs D rs rs D bc bc f rc D rc rc f cr D cs cs bs The calculaio of he raffic flow i WBSNs is as follows: Le k bio represe he umber of rasmied biomedical bis hrough WBSN ad received by he simple relay ode, NC relay ode ad sik ode. The oal raffic geeraed by he biosesor ode oward he sik ode is give below (26). bs k bio b B x bs br ij a ij r, l R, s S (26) Therefore, all raffic is desied owards he sik ode S as give below bs k bio b B Where he erm b B x bs ij a br ij + (f s lr f s rl ) bs k bio l R f rs = 0 r R, s S (27) x bs ij a br ij represe he oal geeraed raffic by he biosesor odes owards he sik ode S. The erm l R represes he oal raffic received by he simple relay ode from f lr s eighbourig odes, l R is he oal raffic rasmied by he simple relay ode o eighbourig odes f rl s ad l R f rs is he rasmissio of raffic owards he sik ode s, hose are expressed as f s bs rl k bio b B e ij rl f lr s r, l R, s S (28) f rl s f lr s = 0 r, l R, s S (29) bs k bio b B e rl bs ij k bio b B e lr ij = 0 r, l R, s S (30) b B,s S bs k bio x bs ij a br s ij + f lr l R,s S cap r rr (31) 8

10 Where cap r represes he capaciy of he relay ode i WBSN. The raffic mus o exceed he capaciy of he ode. The oal raffic rasmied owards he sik ode s as show i (32) which represes he origial flow (aive daa) i BAN. f rs bs k bio b S e ij rs s = f aive_rs r R, s S (32) Where he value e rs ij = 1 r R, s S (33) Meaig ha, here is a coecio bewee he simple relay ode ad he sik ode. The raffic is expressed as f rs bs = k bio r R, s S (34) O he oher had, he oal raffic flow received i he NC odes is rasmied from he biosesor odes ad he simple relay odes. Afer ha, NC relay ode ecodes he received biomedical daa ad direcly seds o he sik ode S. The oal raffic i he NC relay ode is give as k bio b B bs x ij bs a bc s ij z c + f rc l R r, R, c NC, s S (35) s where f cr bs k bio b B e ij rc z c r, R, c NC, s S (36) Where he erms a bc ij, x bs ij, z c ad e rc ij are equal o 1, NC relay ode receives k bs bio from a biosesor ode ad receives k bs bio is received, i ca be expressed as from a simple relay ode. Where D daa_c represes he oal raffic which bs G i k bio b B bs from biosesor ode + k bio b B from relay ode (37) To ecode biomedical packes, he NC relay ode is chose as a sequece coefficie q= (q1, q2,.., q) from Galois Field GF(2s), his is called a ecodig vecor. The sigle oupu ecoded packe is calculaed as he sum of producs of each of he aive packes ha are received a a ode G i (i=1, 2, 3, 4, ) wih a radom coefficie qi. The oupu ecoded packe is described below as Y = q i G i q i GF(2 s ) (38) The igress flow o he sik ode S from NC relay ode is expressed as i=1 D sedig_from_nc = [q i + Y ] q i GF(2 s ) (39) 9

11 Where D sedig_from_nc represes he raffic i he sik ode which is received from NC relay ode ad q i represes he radom coefficie based o Galois Field. Decodig i he sik ode: The sik ode receives daa from boh he simple relay ode ad he NC relay ode which represes he aive daa ad ecodig daa respecively. Wih respec o Gaussia elimiaio, he sik ode decodes he received packes o recover he aive packes [25] as follows D aive_pkt = q i D ecodig_c_i q i GF(2 s ) (40) i=1 The oal eergy cosumpio i he ime (for isace he duraio is [0,]) for he ework is defied as: oal E whole_ework = {[(E TXbr + E RXbr + E TXrl + E TXrs + E TXbc + E RXbc + E TXrc (41) + E TXcs )]} 6. The proposed desig for RLNC The sysem model is composed of he biosesor odes, simple relay odes, ework codig relay odes ad he sik ode as i Fig. 1. The biosesor odes se is posiioed a specific pois o he huma body. Each biosesor geeraes biomedical daa which is rasmied o he sik ode hrough he se of relay odes ad NC relay odes i he ework. These odes represe he bridges bewee he biosesor odes ad he sik ode which improve eergy efficiecy. The se of simple relay odes (R) raspor he packes colleced by he biosesor odes o he sik ode. Moreover hey forward he daa aggregaed from he biosesor odes o aoher relay ode or o he NC relay ode owards he sik ode. The odes are aroud he sik ode, ad his area is called he boleeck zoe. The defiiio of he boleeck zoe is a area wihi a cerai radius from he sik ode, where he radius is represeed by he rasmissio rage of he sesor odes [24]. 10

12 Fig. 1. The proposed desig model i he WBSN [24] Each biosesor ode rasmis duplicae biomedical packes, oe o he simple relay ode ad he oher o he NC relay ode. The processig of packes a he ode sie (simple relay ode ad NC relay ode) has bee give i Fig. 2, showig he algorihm which implemes he forwardig of he medical packes ad he ecodig algorihm a he simple relay ode ad he NC relay ode respecively. Each ode i he boleeck zoe receives a queue (RQueue) io which received packes are placed, ad he ode checks he packe o see wheher or o he packe is a aive. 6.1 Algorihm for packes processig As far as he algorihm for he processig of packes is cocered, as show i Fig. 2, each ode has a receive queue (RQueue), which icludes biomedical packes, ad he ode deposis biomedical packes i he queue (RQueue). I he firs secio of he algorihm, if he ode acs as a simple relay ode, i checks he packe which is received. If i is already forwarded oward he sik ode he i should be removed from he queue ad isered io he forward packe se; oherwise, he simple relay ode rasmis he packe oward he sik ode. I he secod secio of he algorihm, as show i Fig. 2, if he ode represes NC relay ode ad he biomedical packe is a aive packe, he NC relay ode is chose as a sequece coefficie based o he Galois Field GF (2 8 ). The NC ode ecodes he medical packes by applyig a radom coefficies o he packes. Afer successfully creaig he ecoded packes, he NC ode rasmis he ecoded packes ad coefficies o he sik ode. 11

13 6.2 Algorihm for packes decodig The decodig procedure for biomedical packes a he sik ode is as follows: he sik ode receives he aive packes ad ecoded packes from he simple relay ode ad he NC relay ode, respecively. The decodig procedure is show i Fig. 3 where he sik ode receives he aive packes from he simple relay ode. Addiioally, i receives he ecoded packes wih radom coefficie from he NC relay ode ad performs he decodig procedure for he ecoded packes. The sik ode applies Gaussia elimiaio o he ecoded packes ad coefficie i order o rerieve all source packes. Fig. 2. The algorihm for he processig of packes Fig. 3. The algorihm for he decodig he packes i he sik ode 12

14 7. Wireless body sesor ework performace I his sceario he WBSN opology is used as depiced i Fig.4a because i represes a geeral case. The WBSN sceario icludes 13 biosesor odes which are placed o he huma body, for isace elecroecephalogram (K sesor) ad elecrocardiogram (D sesor). Wih respec o his sceario, here are some biosesor odes which sese ad measure vial sigs of he huma body such as Pulse rae, Temperaure, Moio sesor, ad Blood pressure. The defiiio of he boleeck zoe is a area wihi radius (0.6 m) from he sik ode, where he radius represes he rasmissio rage of he sesor odes. The disace bewee he biosesor odes ad sik ode for he sigle-hop echique, ad bewee biosesor odes ad closes ode i he muli-hop echique is show i Table 5. The WBSN is show o he lef had side of Fig. 4a ad he opology explais he boleeck zoe which is show o he righ-had side of he figure. Wih respec o he ework codig approach, here are simple relay odes ad ework codig relay odes added i he boleeck o reduce he eergy cosumpio for he biosesor odes as show i Fig. 4b.The iformaio i Table 1 for Nordic RF2401 is used o calculae eergy cosumpio for all approaches. 13

15 Fig. 4. WBSN opology ad explai he boleeck zoe a WBSN opology wih 13 biosesor ode ad explai he area of he boleeck zoe b Tree opology for WBSN wih simple relay odes(r) ad ework codig relay odes (NC) added i he boleeck zoe. Table 4 The disace (meers) bewee biosesor ode ad sik ode for he sigle hop, ad bewee he biosesor ad he eares ode for he muli-hop [2] Sesor A B C D E F G H I J K L M Sigle -Hop Muli -Hop The LOS ad NLOS Performace I WBSNs, he eergy cosumpio is affeced by propagaio pah loss. The sigle hop approach uilises he lie of sigh (LOS) propagaio model i all rasmissios, ad uses he pah loss coefficie () of LOS, which equals However, i he muli hop approach, he NLOS value is uilised for he rasmissio i WBSN. I addiio, he pah loss coefficie () of NLOS is equal o 5.9. The pah loss coefficie alog LOS chael is lower ha alog NLOS chael, which affecs he eergy usage i WBSN. O he oher had, all rasmissios i he relay ework approach use he pah loss coefficie () of NLOS, which is equal o 5.9, excep for he odes ha are placed ex o sik ode, hose uilise he pah loss coefficie () of LOS, which equals Similarly, i he ework codig approach, for isace, he biosesor odes B ad C are direcly coeced o he sik ode, which use he LOS. However, he biosesor odes such as A, F, D, ad H coec wih he sik ode hrough he simple relay ode ad he NC relay ode, which uses he NLOS where he pah loss coefficie of he LOS ad NLOS are equal o 3.38 ad 5.9, respecively. I all approaches, he eergy usage for he rasmi amplifier i equaio (1) equals o 1.97e-9 J/bi for =3.38 ad 7.99e-6 J/bi for =

16 Each biosesor ode cosumes eergy based o he propagaio model ad he disace bewee he sesor ode ad he sik ode. For example, EMG (A) ad he ECG (ode D) cosume more eergy ha sesor B as hey sed biomedical packes oward he sik ode hrough a simple relay ode ad a NC relay ode, whereas he body emperaure sesor (B) is rasmied direcly from he biomedical packes o he sik ode. 7.2 Eergy cosumpio resuls The eergy cosumpio for he WBSN boleeck zoe is compued based o he ework codig approach comparig sigle hop [26], muli- hop [26], ad relay ework approaches. I he sigle-hop approach, he biosesor odes i he boleeck zoe cosume more eergy based o he disace whe compared wih oher approaches; he biosesor odes A, F, ad H show greaer eergy cosumpio, as show i Fig. 5. However, i he muli-hop approach, he biosesor odes relay he packes via he iermediae ode owards he sik ode. The odes A, ad D have higher eergy cosumpio i he muli-hop bu he ode C has he same value of eergy i mos approaches because i is coeced oly wih he sik ode, as illusraed i Fig. 5. Wih respec o he relay ework approach, addig o he umber of relay odes which forward he packes o he sik ode improves he eergy efficiecy ad decreases he eergy usage of biosesor odes i he boleeck zoe. Moreover, here is a leas oe possible relay ode i lie of sigh. The eergy cosumpio for all biosesor odes is lower compared wih sigle-hop ad muli-hop approach. I he ework codig approach, simple relay odes ad NC relay odes are added o he boleeck zoe o reduce he eergy cosumpio for he biosesor odes i his area. I ca be observed ha eergy usage for he odes B ad D is lower whe compared wih oher approaches excep ha he values of eergy cosumpio for A, F ad H are slighly higher ha i he relay ework approach because eergy cosumpio of hese odes are calculaed based o o-lie of sigh as illusraed i Fig. 5. Deailed resuls are show i Table 5. 15

17 Fig. 5. Compariso of eergy cosumpio for biosesor odes i he boleeck zoe based o he sigle hop, muli-hop, relay ework ad Nework codig. Table 5 The eergy cosumpio for he odes i he all approaches Node Eergy usage i Sigle Hop (μj/bi) Eergy usage i Muli Hop (μj/bi) Eergy usage i Relay ework (μj/bi) Eergy usage i Nework Codig (μj/bi) A B C D F H Moreover, each biosesor ode seds duplicaed packes, oe hrough simple relay ode ad he secod hrough NC relay ode. I he rasmissio rage of 0.3 cm, he NC relay ode receives packes from differe odes ad ecoded packes are he se o he sik ode. The sik ode decodes he received packes ad rerieves aive packes eve if here is a failure i oe of he rasmissio liks. Wih respec o ree opology for body area eworks as preseed i Fig. 6a i he firs case, he biosesor ode A (EMG sesor) seds biomedical packes hrough wo pahs, firs hrough he simple relay ode, which forwards biomedical packes o he sik ode, ad secod hrough he ework codig relay ode, which creaes he ecoded packes ad he rasmis hem wih coefficies o he sik ode. If a failure occurs i oe of he liks (oe pah) hrough he rasmissio, as show i Fig. 6a (ree opology), where he dashed red lie represes he failed lik for ode A, successful rasmissio of he medical packes is achieved hrough a aleraive lik. I his case, i ca be see ha he eergy cosumpio for ode A based o he NC wih oe lik failure saves eergy ad achieves he delivery of daa, as illusraed i Fig. 6b. 16

18 The bar char show i Fig. 6c compares he eergy cosumpio for ode A, calculaed based o he sigle hop ( µj/bi), muli- hop ( µj/bi), relay ework ( µj/bi), NC ( µj/bi,) ad NC wih oe lik failure ( µj/bi). We ca see sigifica differeces i he eergy cosumpio value for ode A i he sigle hop ad he NC wih oe lik failure. The eergy usage for ode A is higher i he sigle hop. Moreover, he eergy cosumpio for ode A i he relay ework scheme is slighly higher ha he NC wih oe lik failure. O he oher had, he amou of eergy is lower i he laer scheme. Accordig o he bar graph show i Fig. 6d, which compares he eergy usage for ode D for all approaches, he eergy cosumpio for ode D based o he sigle hop, muli- hop, relay ework, NC, ad NC wih oe lik failure, are equal o µJ/bi, µJ/bi, µJ/bi, µj/bi ad µJ/bi, respecively, which is low i he NC wih oe lik failure. 17

19 Fig. 6. Case 1: he eergy cosumpio for he all odes i boleeck zoe wih eergy cosumpio for he odes A ad D a Tree opology for WBSN wih a failure lik for he ode A ad D. b Compariso of eergy cosumpio for biosesor odes for all approaches ad NC wih a failure lik i A ad D odes. c Compariso eergy cosumpio for ode A i all approaches. d Compariso eergy cosumpio for ode D i all approaches. I he secod case, if here is a failure of liks for odes F ad H, as show i Fig. 7a, he eergy cosumpio is calculaed for biosesor odes i he boleeck zoe based o he sigle hop, muli-hop, relay ework, ework codig, ad NC wih a lik failure i F ad H odes, which are compared i Fig. 7b. The eergy usage for ode F is calculaed depedig o he sigle hop, muli- hop, relay ework, NC, ad NC wih oe lik failure, which are equal o µJ/bi, µJ/bi, µJ/bi, µj/bi ad µj/bi, respecively. Moreover, he eergy cosumpio for ode H based o he sigle hop, mulihop, relay ework, NC, ad NC wih oe lik failure are equal o µJ/bi, µJ/bi, µJ/bi, µJ/bi ad µJ/bi, respecively. The resuls i his case show more deail i Fig. 7b cocerig he compariso of he eergy usage for odes F ad H i all schemes, which are show i Fig. 7c ad Fig. 18

20 7d respecively. Fig. 7c ad Fig. 7d also show ha, he eergy cosumpio for odes F ad H usig NC wih a lik failure are lower ha for oher approaches. c 19

21 Fig. 7. Case 2: he eergy cosumpio for he all odes i boleeck zoe wih eergy cosumpio for he odes H ad F. a Tree opology for WBSN wih a failure lik for he ode H ad F. b Compariso of eergy cosumpio for biosesor odes for all approaches ad NC wih a failure lik i F ad H odes. c Compariso eergy cosumpio for ode F i all approaches. d Compariso eergy cosumpio for ode H i all approaches. Also, i he oher cases, he comparisos of he eergy cosumpio for biosesor odes i he boleeck zoe are show i Fig. 8a ad Fig. 8b for all approaches ad NC wih lik failure i he odes A ad H, ad odes F ad D, respecively. The eergy usage for odes A ad H are equal o µj/bi ad µj/bi, respecively i NC wih lik failure i odes A ad H. I addiio, odes F ad D eergy cosumpio equals o µj/bi ad µj/bi respecively based o NC wih lik failure i odes F ad D. The eergy saved is calculaed based o he differece bewee he proposed approach ad oher approaches, as show i Fig. 8c. Wih regard o NC, he resuls of he oal eergy saved are µj/bi, µj/bi, ad µj/bi for sigle-hop, muli-hop, ad relay ework respecively, as show i Fig.8c. 20

22 Fig. 8. Eergy cosumpio ad savig i he boleeck zoe a Compariso of eergy cosumpio for biosesor odes for all approaches ad NC wih a failure lik i A ad H odes. b Compariso of eergy cosumpio for biosesor odes for all approaches ad NC wih a failure lik i F ad D odes. c Eergy savig for biosesor odes based o he differece bewee ework codig ad sigle hop, muli-hop ad relay ework. 8. Coclusios I his paper we address he problem of high eergy usage of biosesor odes caused by he boleeck zoe i WBSNs. The desig of a ovel mahemaical model is proposed for body area ework opology based o he graph heory, he coecio ad relaioship bewee he biosesor odes, simple relay odes, ework codig relay odes ad he sik ode are explaied i his model. Our resuls show ha he proposed RLNC model improves he eergy efficiecy for biosesor odes i he boleeck zoe ad guaraees daa delivery i he case of lik failure. The mahemaical model is herefore developed for he ehaceme of 21

23 wireless body eworks. Moreover, eergy savig for biosesor odes i he boleeck zoe is achieved hrough applyig he RLNC scheme. 9. Refereces [1] Ifikhar M., Al Elaiwi N., ad Aksoy M. S.,.: 'Performace Aalysis of Prioriy Queuig Model for Low Power Wireless Body Area Neworks (WBANs), ' Procedia Compu. Sci., 2014,34, pp [2] Ahlswede, R., Cai, N., Li, S.-Y. R., e al.: 'Nework Iformaio Flow, ' IEEE Tras. If. Theory, 2000, 46, (4), pp [3] Li, S. Y. R., Yeug, R. W., ad Cai, N.,.: 'Liear ework codig, ' IEEE Tras. If. Theory, 2003, 49, (2), pp [4] Shahida A. A., Fisal N., Ismail N. N., e al.: ' Daa Recovery i Wireless Sesor Neworks usig Nework Codig', Jural Tekologi,2015,73,(3), pp [5] Magli, E., Wag, M., Frossard, P., e al.:' Nework codig mees mulimedia: A review', IEEE Tras. Mulimed., 2013, 15, (5), pp [6] Plaz, D., Woldegebreal, D. H.,ad Karl, H.:' Radom Nework Codig i Wireless Sesor Neworks: Eergy Efficiecy via Cross-Layer Approach,' 2008 IEEE 10h I. Symp. Spread Specr. Tech. Appl., Ialy, Auges 2008,pp [7] Fude, M. V., Gaikwad, M. A., ad Higaikar, P. A. W..:'Review of Lifeime Ehaceme of Wireless Sesor Neworks,' IORD Joural of Sciece & Techology, 2015, 2, (2), pp [8] Tauqir,A., Javaid,N., Akram, S., e al.:' Disace aware relayig eergy-efficie: DARE o moior paies i muli-hop body area sesor eworks,' Proc h I. Cof. Broadbad, Wirel. Compu. Commu. Appl. BWCCA 2013, Frace, Ocober 2013,pp [9] Movassaghi,S., Shirvaimoghaddam,M., ad Abolhasa, M.: 'A Cooperaive Nework Codig Approach o Reliable Wireless Body Area Neworks wih Demodulae-ad-Forward,' Proc h I. Wirel. Commu. Mob. Compu. Cof. IWCMC2013,Ialy,July 2013, pp [10] Deepak, K.S. ad Babu, A.V.:' Improvig eergy efficiecy of icremeal relay based cooperaive commuicaios i wireless body area eworks', Ieraioal Joural of Commuicaio Sysems, 2015, 28, (1), pp [11] Yousaf, S., Javaid, N., Qasim, U., e al.:' Towards Reliable ad Eergy-Efficie IcremealCooperaive Commuicaio for Wireless Body Area Neworks', 2016, 16, (3), pp. 284 [12] Miliois, V, Aloso, L., ad Verikoukis, C.: ' CooPNC: A cooperaive mulicas proocol exploiig physical layer ework codig', Ad hoc eworks, 2014, 14, pp [13]Dharshii,P.,M.,P., ad Tamilarasi, M.,.:'Adapive Reliable Cooperaive Daa Trasmissio Techique for Wireless Body Area Nework,' Proc. I. Cofe. Iforma. Commu. Embedd. Syse. ICICES2014, Idia,Februry 2014, pp [14] Arrobo, G., E., ad Gili, R., D.: ' Improvig he reliabiliy of wireless body area eworks,' Au. I. Cof. IEEE Eg. Med. Biol. Soc.,USA,30 Auges-1 Sepmber 2011,pp [15] Movassaghi,S., Shirvaimoghaddam,M., Abolhasa,M., e al.:'a eergy efficie ework codig approach for Wireless Body Area Neworks,' 38h Au. IEEE Cof. Local Compu. Neworks, Ausralia, Ocober 2013,pp [16] Ehyaie,A., Hashemi,M., ad Khadivi,P.:' Usig relay ework o icrease life ime i wireless body area sesor eworks,' IEEE I. Symp. a World Wirel., Mobil., Mulimed. Neworks Work, Jue 2009, pp [17] Begoya, O., Aloso, L., ad Verikoukis C.: 'Highly reliable eergy-savig MAC for wireless body sesor eworks i healhcare sysems', IEEE Joural o Seleced Areas i Commuicaios,2009, 27,(4),pp [18] D Adreagiovai,F., ad Nardi,A.:'Towards he fas ad robus opimal desig of wireless body area eworks,' Appl. Sof Compu., 2015,37, pp [19] Braem, B.,Laré, B.,Moerma, I., e al.:' The eed for cooperaio ad relayig i shor-rage high pah loss sesor eworks,' Proc I. Cof. Ses. Techol. Appl. SENSORCOMM 2007, Spai,Ocober 2007, pp [20] Bagash, J. I.,Kha, A.,W., ad Abdullah, A. H.:'Daa-Ceric Rouig for Ira Wireless Body Sesor Neworks,' Joural of Medical Sysems,2015,39,(91), pp

24 [21] Reuses, E., Joseph,W., Vermeere, G., e al.:' O-Body Measuremes ad Characerizaio of Wireless Commuicaio Chael for Arm ad Torso of Huma, Proc. 4h I. Work. Wearable Impla. Body Ses. Neworks, Germay, March 2007, pp [22] For,A., Ryckaer,J., Desse,C., e al.:' Ulra-widebad chael model for commuicaio aroud he huma body, IEEE J. Sel. Areas Commu.,2006,24, (4), pp [23] Body,J.A., ad Mury,U.S.R.: 'Graph heory wih applicaios'(macmilla Press Ld,1976). [24] Alshahee,H., ad Takruri Rizk,H.:'Improvig he Eergy Efficiecy for Biosesor Nodes i he WBSN Boleeck Zoe Based o a Radom Liear Nework Codig,' Proc. 11h I. Symposi. Medic. Iform. Commu. Tech. Poeugal, Februry 2017, pp [25] Pfleschiger, M. Navarro, ad Ibars, C.: 'Eergy-efficie daa collecio i WSN wih ework codig,' 2011 IEEE GLOBECOM Work. GC Wkshps,USA, December 2011, pp [26] Reuses,E., Joseph,W., Lare, B., e al.: 'Characerizaio of O-Body Commuicaio Chael ad Eergy Efficie Topology Desig for Wireless Body Area Neworks,'. IEEE Tras. If. Techol. Biomed., 2009, 13, (6), pp