DSP ALGORITHM AND SYSTEM DESIGN FOR UWB COMMUNICATION SYSTEMS

Size: px
Start display at page:

Download "DSP ALGORITHM AND SYSTEM DESIGN FOR UWB COMMUNICATION SYSTEMS"

Transcription

1 DSP AGOTH AD SYST DSG FO UWB COUCATO SYSTS YAG U (B..(Hos.), TU) A THSS SUBTTD FO TH DG OF AST OF GG DPATT OF CTCA & COPUT GG ATOA UVSTY OF SGAPO 6

2 ACKOWDGT would lke to exress my utmost arecato to my suervsors, Dr. Zhe Yuaj ad Prof. Har Krsha Gar, for be so uderstad, helful ad ecoura at all tmes rouh my research ad study atoal Uversty of Saore ad sttute of croelectrocs. am esecally rateful for e valuable udace ad costructve suestos offered by Dr. Zhe rouhout e develomet of s dssertato. would also lke to ak my freds, r. Cao zhe, ss Ya Jaa, r. Cao u, r. To Ya, ss Zhou Qaoer, ss We Xaoqa ad ss u aomao, for er kdess ad suortve work. ay aks to my famly, for always be ere for me, wheever ad wherever. astly also wsh to exted my arecato to all e oers who have, oe way or aoer, heled mak s dssertato a very reward oe. t has bee a leasure work toeer.

3 TAB OF COTTS ACKOWDGT... TAB OF COTTS... SUAY... V ST OF FGUS... V ST OF ABBVATOS... V ST OF ABBVATOS... V ST OF SYBOS... X TODUCTO.... Backroud.... Scoe of e Thess Orazato of e Thess... TATU VW.... & B-OFD.... ake ecever... 3 SYST OD & CV DSG Pulse-sha & odulato ulta Chael odel... 7

4 3.3 Bake ecever Structure PFOAC AAYSS Correlato ecevers S Crtero & Weer Soluto S Aalyss B xresso sadjustmet SUATO SUTS Theoretcal versus Smulated B ffect of µ o B Aroxmato B Comarso w Oer ake Structures COCUSO Cocluso emarks Future Works FCS ST OF PUBCATOS... 7

5 SUAY Ultra-Wdebad (UWB) commucato s recozed as oe of e most roms techoloes for ext eerato Wreless Persoal Area etwork (WPA) sce t has e otetal to rovde a low comlexty, low cost, low ower cosumto, ad hh data rate coectvty commucato systems. t reles o trasmsso of ultra-short ( aosecod scale) ulses ad avods us susodal carrers or termedate frequecy (F) rocess. Oe of e most sfcat features of UWB commucato s ts fe multa resolvablty. To maxmally exlot e chael dversty, a ake recever s usually emloyed to effectvely cature e multa eery. Covetoal ake recevers requre chael formato, clud e multa delay ad atteuato, to be rovded before comb e multa eery. However e chael formato s ukow to e recever ad dffcult to estmate ractcal UWB trasmsso systems. s ess, a ovel bld ake (Bake) structure s roosed for hh data rate low ower cosumto UWB commucato systems. The word bld here does ot mly at tra sequece ca be totally elmated, but refers to e fact at e chael formato s ot eeded for e ake system to erform effectvely. oer words, t avods e estmato of multa chael whch must be rovded for covetoal ake recever systems. The trascever comlexty s furer reduced by us Aalo-to- Dtal Coverter (ADC) work at sub-yqust saml rate. The closed form bt error rate (B) erformace aalyss s rovded as well. xtesve smulatos have bee carred out whch verfed e eoretcal aalyss. V

6 ST OF FGUS Fure. Frequecy & eery comarso for commucato systems Fure. FCC reulated sectral mask for UWB door commucato systems... 4 Fure.3 Oerato dstace ad data rate of major wreless stadards... 5 Fure.4 Comarso of UWB ad covetoal B trascever archtectures... 6 Fure. The B-OFD frequecy bad la Fure. A tycal trascever archtecture for a B-OFD system... 4 Fure.3 Waveforms for dervatves of Gaussa moocycle Fure.4 -UWB modulatos... 9 Fure.5 A eeral ake recever structure... Fure 3. Tycal Cs for UWB door chaels Fure 3. Bake recever archtecture Fure 5. Ta weht Fure 5. Ta weht dfferece ad ts hstoram Fure 5.3 Theoretcal versus smulated B for C-C Fure 5.4 ffect of µ o B aroxmato... 6 Fure 5.5 B erformace for C Fure 5.6 B erformace for C Fure 5.7 B erformace for C Fure 5.8 B erformace for C V

7 ST OF ABBVATOS A-STA ADC Aake ASC AWG B BPSK Bake CFJ CFO C C CS DC DS P FAS FCC FPGA FTT Aecy of Scece, Techoloy ad esearch Aalo-to-Dtal Coverter All ake Alcato-Secfc terated Crcut Addtve Whte Gaussa ose Bt rror ate Bary Phase Shft Key Bld ake Carrer Frequecy Jtter Carrer Frequecy Offset Chael mulse esose Chael odel Chael State formato Drect Curret Drect Sequece ffectve sotroc adated Power lectromaetc Federal Avato System Federal Commucato Commsso Feld Prorammable Gate Array Fast Fourer Trasform V

8 GPS F FTT S S PD P A B-OFD B-UWB BOA F FP S C S B OFD PA PDP PP Pake Global Posto System termedate Frequecy verse Fast Fourer Trasform sttute of croelectrocs mulse ado ter Symbol terferece east ea Square ow Probablty of Detecto ow Probablty of terceto ult-user Access ult-bad Orooal Frequecy Dvso ultlex ult-bad Ultra-Wdebad ult-bad OFD Allace atched Flter atched Frot-d Processor mum ea Square rror axmum ato Comb ea Square rror arrowbad Orooal Frequecy Dvso ultlex Pulse Amltude odulato Power Delay Profle Pulse Posto odulato Partal ake V

9 PT PSD &O F S S-V S Sake TD TFC TFK U- UWB WA WPA Pulse eetto Tme Power Sectral Desty eort & Order ado Frequecy oot ea Square Saleh-Valezuela Sal to ose ato Selectve ake Taed Delay e Tme Frequecy Code Tme Frequecy Kerel Ulcesed atoal formato frastructure Ultra-Wdebad Wreless ocal Area etwork Wreless Persoal Area etwork V

10 ST OF SYBOS A C J K Μ X m W A factor to ormalze e ulse eery Chael Caacty S cost fucto Total ta umber a brach Total brach umber Total umber of multa a chael sadjustmet Chael lo-ormal shadow coeffcet The m multa comoet a chael Chael Badwd ( t) d order Gaussa dervatve waveform m Gaussa ulse vector based o real chael tm l, k Gaussa ulse vector based o recever saml tm h ( t) Chael mulse resose r ( t) eceved sal r ( t) eceved sal for symbol s ( t) Trasmtted sal s Vector of sal after correlato w ( t) ose sal X

11 w, ose for l k k ta of l brach after correlato oerato w Vector of ose mxed w sal after correlato w Vector of ure ose after correlato K ( t) P ( f ) order Gaussa dervatve waveform tme doma order Gaussa dervatve waveform frequecy doma b Costellato for PA/Trasmtted symbol bˆ stmated symbol b Vector of tra sequece TH c User-secfc TH code for mulse c, Weht for l k k ta of l brach l k c, c c o Weht for k ta of l brach for symbol dur tra hase Ta weht vector Otmum ta weht vector o c otmum ta weht S c S ta weht d Costellato for PP e rror sal for symbol durato tra hase S e S error v, ut of e l k k ta e l brach l k v, ut of e k ta e l brach for symbol X

12 v y Decso sal for Ta ut vector symbol b Bt eery Pulse eery P P detty matrx P P b umber of fers a ake umber of fers a Sake umber of fers a Pake c umber of chs each frame f umber of relcas each trasmtted symbol Oe sde ower sectrum desty of AWG P b Bt error rate (B) Q evector matrx Ta ut autocorrelato matrx ( ) Pulse correlato fucto ( ) S Delay sectrum h T c Ch durato T d axmum excess delay of e chael T mulse durato T r Pulse reetto tme T s Symbol durato X

13 T l Chael tme delay of e l cluster σ Stadard devato of a Gaussa radom varable σ e Varace of S error σ w l,k Varace of ose after correlato γ φ S rato Process a µ Adatato ste sze dur tra hase m Chael atteuato of e Pulse osto offset PP m multa comoet k,l Chael atteuato of e k multa e l cluster m Vector of chael atteuato m Chael delay of e k,l Chael delay of e m multa comoet k multa e l cluster l ecever saml terval of l brach l,k ecever saml tme of e k ta e l brach m Vector of Chael delay lk Vector of recever saml tme S Chael S delay sread Λ Λ S Delay betwee cosecutve Bake brach Cluster arrval rate of chael multa Daoal eevalue matrx X

14 ς evalue λ ay arrval rate of chael multa a cluster λ eevalue of autocorrelato matrx of ut data Successfully collected chael eery [ ] xectato oerato δ ( ) Drac delta fucto Q ( ) Comlemetary cumulatve dstrbuto fucto tr ( ) atrx trace oerato Covoluto oerato X

15 CHAPT : TODUCTO TODUCTO Ths chater ves a bref troducto of UWB commucato systems. UWB techoloy has bee used for more a a cetury. However t attracts real atteto bo dustry ad academy oly after e Federal Commucato Commsso (FCC) released a hue ew badwd of GHz for t. UWB alcatos clude, but ot lmt to, Wreless Persoal Area etworks (WPAs), sesor etworks, ma systems, vehcular radar systems, etc. The scoe ad orazato of e whole ess are also reseted s chater.

16 CHAPT : TODUCTO. Backroud Scetsts ad eeers have kow about UWB sals sce Gulelmo arco veted rado commucatos utlz eormous badwd as formato was coveyed by sark-a devces more a a cetury ao. However e sals were more dffcult to cotrol or detect a arrowbad (B, sle-frequecy) sals at at tme. oder UWB techoloy came to e cture sce 96s, whe e troducto of UWB mulse radar systems was motvated by e hh sestvty to scatters ad low ower cosumto alcatos []-[3]. Commercally, e early UWB vestato was larely uder e aes of e U. S. Deartmet of Defese at adoted wdebad sals rmarly for very accurate localzato ad ma e cotext of secure commucatos [4], [5]. Academcally, e UWB research ad develomet were larely oeered by Prof. Scholtz ad hs rou [6]-[9], [3], [8], [9], [3], focus maly o low-rate alcatos. UWB, as e ame suests, occues a very lare badwd for sal trasmsso whle e emsso ower s well below covetoal arrowbad or wdebad systems. Fure. below llustrates s cocet. UWB ulses sread eery over severalahertz rae of frequeces, as oosed to tradtoal arrowbad, whch covers a lmted bad of about 3 klohertz. Cellular hoes oerate e wdebad, whch covers about 5 meahertz.

17 CHAPT : TODUCTO arrowbad, 3 klohertz ery Wdebad, 5 meahertz Ultra-wdebad, several ahertz Frequecy Fure. Frequecy & eery comarso for commucato systems. The defto of UWB evolves w tme. The rule mak of UWB was oeed by FCC 998. The result Frst eort ad Order (&O) at ermtted deloymet of UWB devces was aouced o 4 February ad released Arl [], whch uleashed a very lare badwd of GHz to UWB trasmssos. Three tyes of UWB systems are defed s &O: ma systems, commucato ad measuremet systems, ad vehcular radar systems. Secfcally, UWB characterzes trasmsso systems w stataeous sectral occuato excess of 5 Hz or a fractoal badwd of more a %. The badwd ad sectral mask for door commucato systems assed by FCC s llustrated Fure.. t ca be see at e FCC reulated ower levels are very low (below -4.3 dbm), whch allows UWB techoloy to coexst w leacy servces such as 8.a wreless local area etwork (WA), radar systems, etc., as well as 3

18 CHAPT : TODUCTO overlay w sestve mltary ad cvla servces adjacet bads such as lobal osto system (GPS) ad federal avato system (FAS). Cellular hoes, for examle, trasmt u to 3 dbm, whch s equvalet to 7 hher ower sectral desty (PSD) a UWB trasmtters are ermtted []. Curretly e 8.5 Work Grou s utt efforts to stadardze UWB wreless rados for door multmeda trasmssos UWB P msso evel dbm GPS Bad.96.6 door mt Part 5 mt Frequecy GHz Fure. FCC reulated sectral mask for UWB door commucato systems. Based o e FCC reulato, UWB s mostly sutable for e short-rae hh-seed 4

19 CHAPT : TODUCTO alcatos amo oer wreless systems, as ca be see Fure.3. From Shao s eory [], ( S) C.44 W lo, (.) where C stadards for e chael caacty bts er secod, W reresets e chael badwd hertz ad S s e sal to ose rato, e chael trasmsso rate rows learly w chael badwd but oly loarmcally w S. oer words, e chael caacty creases much faster as a fucto of badwd a ower. Thus UWB has e otetal to offer hh data rate ( several hudred meabt er secod, bs) to emer hh-seed-demad alcatos. Dstace Coverae m m km km cellular G/.5G Cellular 3G Cellular Fure.3 Oerato dstace ad data rate of major wreless stadards. 5

20 CHAPT : TODUCTO A comarso betwee a basc UWB trascever ad a covetoal arrowbad trascever s show Fure.4. The ma dfferece betwee em s e sav of e comlex suerheterodye structure a UWB trascever. The trasmsso of UWB waveforms ca be free of se-wave carrers (called carrer-less short ulse techque [3]) ad do ot requre ay F rocess because ey ca oerate e basebad. Data ut odulator PA A F Flter BPF PF Demodulator Data Outut F F F (a) Tycal arrowbad trascever archtecture Data ut odulator F Flter A F Flter PF Demodulator Pulse Geerator PF Pulse Geerator PF (b) Tycal UWB trascever archtecture Fure.4 Comarso of UWB ad covetoal B trascever archtectures. To characterze e UWB roaato chael, may resoses have bee receved to e Call for Cotrbutos o UWB Chael odel [4]. Cosder varous roosals at otmally chose e resectve model arameters (such as mea excess delay, root mea square (S) delay, mea umber of sfcat as etc) to best ft e observed data, e 8.5.3a chael model task rou has adoted a modfed verso of e Saleh-Valezuela (S-V) model for door chaels [5]. The chael measuremets 6

21 CHAPT : TODUCTO showed a cluster of e multa arrvals whch s best catured by e S-V model. ore detals o e chael model wll be dscussed chater 3. Bascally, e very wde badwd of e trasmtted ulse allows fe resoluto of e multa comoets. Ths has bo ros ad cos. Fe delay resoluto mles e otetal for sfcat dversty as due to e lare umber of avalable as. However, e total receved eery s dstrbuted over a lare umber of as, whch meas e recever must be caable of ck u ad comb e multa eery a roer way. ormally a ake recever s emloyed for s job. summary, occuy hue badwd by trasmtt ultra short ( aosecod scale) formato-bar ulses, UWB rado has uque advataes at have lo bee arecated by e radar ad commucatos commutes: The wde badwd ad hh tme resoluto ablty of UWB sals eerally make em more robust to multa terferece ad chael fad; A drect alcato of e Shao's caacty eorem to a addtve whte Gaussa ose (AWG) chael shows at UWB systems offer a otetally hh data rate trasmsso caablty w caacty creas learly w badwd; The low trasmsso ower of e UWB sals traslates to a F sature w a low robablty of terceto ad detecto (P/PD), ad also roduces mmal terferece to roxmty systems ad mmal F heal hazards; The fe tme resoluto of UWB systems makes em ood caddates for 7

22 CHAPT : TODUCTO locato ad ra alcatos (w recso at e cetmeter level); ad ore mortatly, UWB systems have low system comlexty ad low cost, sce ey are essetally basebad systems ad ca be made early "all-dtal", w mmal F or mcrowave electrocs. 8

23 CHAPT : TODUCTO. Scoe of e Thess The whole roject s fuded ad suorted by e Aecy of Scece, Techoloy ad esearch (A-STA) UWB esearch roram, jotly collaborated betwee sttute of croelectrocs () of Saore ad atoal Uversty of Saore (US). The taret of s ess s to develo a ractcal ad effectve alorm for low comlexty recevers of mulse ado () UWB commucato systems. The eoretcal erformace aalyss s carred out detals w a closed form bt error rate (B) exresso derved. xtesve smulatos have bee doe to verfy e correctess of e dervato. The B erformace of e roosed archtecture s also comared w covetoal ake recevers to demostrate ts feasblty of mlemetatos. 9

24 CHAPT : TODUCTO.3 Orazato of e Thess chater, e backroud of UWB commucato systems s brefly troduced. Chater resets some relevat works whch have bee doe so far s area. Two ormal UWB trasmsso schemes, ad ult-bad Orooal Frequecy Dvso ultlex (B-OFD), are descrbed. Commo ake recever structure s also ve s chater. The ovel bld ake (Bake) alorm s roosed chater 3 w detaled exlaatos o e sal model, chael model ad recever structure, followed by -de erformace aalyss chater 4. Chater 5 shows e smulato results whch have verfed e eoretcal dervato, as well as comared e erformace of e roosed alorm ad covetoal meods. Cocluso remarks are ve chater 6 w suested future work.

25 CHAPT : TATU VW TATU VW Ths chater revews some of e related lterature works. Drect Sequece (DS) ad ult-bad Orooal Frequecy Dvso ultlex (B-OFD) are e two mostly emloyed UWB trasmsso techques. mulse ado () s e smlest ad most frequetly adoted form of DS-UWB, erefore t s e focus of s vestato. B-OFD has a major advatae flexble sectrum selecto, but e trascever structure s comlcated due to e resece of multle frequeces., o e oer had, oerates e basebad ad does ot requre ay F rocess, whch results much smlfed recever structure. Therefore scheme s adoted s ess. The mulse shae ad sal modulatos of -UWB are e revewed. The secod order dervatve Gaussa moocycle ad BPSK are chose our system. Commo ake recever structure s also revewed s chater. ake recevers are used to cature multa dsersed chael eery wreless commucato systems. All ake (Aake), Selectve ake (Sake) ad Partal ake (Pake) are commoly kow ake tyes. Performace ad comlexty are trade-offs cosder us dfferet tyes of ake structures. Chael state formato (CS) has to be rovded for ake recevers to work effectvely, but UWB chaels are very dffcult ad costly to be estmated real tme. Ths s why our Bake structure s roosed ad develoed.

26 CHAPT : TATU VW. & B-OFD There are dfferet ways of utlz e 7.5 GHz badwd assed by FCC UWB systems. O oe had, e sal ca be shaed so at ts eveloe exads e full frequecy sectrum. Ths s called -UWB [6]-[8], [5], also kow as ulsed UWB, ad t s oe of e smlest forms of DS-UWB. O e oer had, e hue bad ca be dvded to multle subbads, ad e sal ca be shaed so at t occues a subbad of 5-8 Hz. Ths s called multbad UWB (B-UWB) [6]-[8]. W e formato of e ult-bad OFD Allace (BOA) Jue 3, e OFD techque for each subbad was added to e tal multbad aroach to mrove e erformace due to OFD s heret robustess to multa [9], []. The B-OFD aroach allows for adatato to dfferet reulatory evromets by dyamcally tur off subbads ad dvdual OFD toes to comly w local rules of oerato o allocated sectrum. t also facltates future scalablty of e sectrum use. The bad la for e BOA roosal s show Fure.. The avalable sectrum of 7.5 GHz s dvded to 4 subbads each of 58 Hz. Subbads are roued to fve local chaels. Chael, whch cotas e frst ree bads, s madatory for all UWB devces ad rados whle e oer rema chaels are otoal. There are u to four tme-frequecy codes (TFC) er chael, us allow for a total of coets w e curret BOA roosal. addto, e roosal also allows flexblty to avod chael whe ad f Ulcesed atoal formato frastructure (U-) terferece, such as from 8.a, s reset [].

27 CHAPT : TATU VW Power Fure. The B-OFD frequecy bad la. The formato trasmtted o each subbad s modulated us OFD. OFD techque dstrbutes e data over a lare umber of carrers at are saced aart at recse frequeces. Ths sac rovdes e orooalty roerty whch revets e demodulators from see frequeces oer a er ow. The beefts of OFD are hh-sectral effcecy, reslecy to F terferece, ad lower multa dstorto. Fure. shows a tycal trascever archtecture for a B-OFD system. ex ( jπf c t) (a) Trasmtter structure. 3

28 CHAPT : TATU VW cos( πf c t) s( πf c t) (b) ecever structure. Fure. A tycal trascever archtecture for a B-OFD system. The trasmtter ad recever archtecture for a B-OFD system are very smlar to ose covetoal wreless OFD systems. The ma dfferece s at e B-OFD system uses a tme-frequecy kerel (TFK) to secfy e ceter frequecy for e trasmsso of each OFD symbol []. Deste e above metoed advataes rovded by B-OFD scheme, ere are also roblems assocated w s aroach. Comlcated trascever archtecture s a major cocer. Frstly, to eerate all e subcarrer frequeces, frequecy syeszer must be bult whch esures at system ca swtch betwee e ceter frequeces extremely fast (w a few aosecods). Secodly, local oscllators at e UWB trasmtter ad recever ed are roe to msmatch ad ve rse to multle carrer frequecy offset/jtter (CFO/CFJ) []. Uless comesated for, CFO/CFJ s kow to derade e system erformace severely. Thus B-OFD eeds for CFO sestvty aalyss, low comlexty CFO estmators, ad er-subcarrer based chael estmato modules. astly, e addtoal FFT/FFT modules e trascever systems add about 5% of e dtal 4

29 CHAPT : TATU VW basebad rocess comlexty, whch also results more ower cosumto. scheme, o e oer had, avods us susodal carrers or ay F rocess so at t reatly reduces e trascever comlexty ad overall ower cosumto. Therefore t s adoted our system w ts smler trascever archtecture. works by trasmtt basebad ulses of very short durato, tycally o e order of aosecod or sub-aosecod, ereby sread e eery of e rado sal very ly from ear DC to a few ahertzes [8]. The shae of e ulse secfes e frequecy sectrum of e trasmtted sal ad well desed ulse shae allows maxmum emtted ower uder FCC frequecy mask. A varety of ulse shaes ad er corresod frequecy sectrums have bee roosed ad dscussed [3]-[8]. Amo em, e most frequetly emloyed ulse shaes are e dervatves of Gaussa fucto. The tme doma ad frequecy doma reresetatos of e order Gaussa dervatve ( t) ad ( f ) P are ve (.) ad (.) [8], resectvely, ( 3 t ) /! / / 4k k / 4k k ( ) t / / t 4 σ σ π e, (.) k ( k )! k! ( )!! where ( )!! ( ) ( 3) K 3 fucto whch s assocated w e ulse durato., σ s e stadard devato of e Gaussa P ( f ) 4 ( ) ( π ) / ( σ ) / / 4 π f σ f e. (.)!! ( ) where reresets e maary art of a comlex umber. 5

30 CHAPT : TATU VW Fure.3 shows e waveforms of st order to 4 order dervatve Gaussa moocycles ad er corresod ower sectrums. t ca be see at creas e dervatve order of e Gaussa moocycle has e effect of shft e frequecy sectrum to occuy a hher frequecy rae so at t ca better ft FCC frequecy mask. However hher order dervatve moocycle s ot feasble from crcut ersectve as well as maematcal tractablty. s ess, a commoly used d order Gaussa moocycle s emloyed w ulse durato of T. 5 s. The mulse ca be wrtte as, a smlfed verso from (.): t σ t ( t) A 3.5 e, (.3) σ where σ s related to e ulse wd by T 7σ w T be e durato of a sle ulse whch covers 99.99% of e total ulse eery, ad A s a factor to ormalze e ulse eery to oe. 6

31 CHAPT : TATU VW st order d order 3rd order 4 order Am ltude Tme (s) (a) Tme doma dervatves of Gaussa waveform. Power Sectrum (db) st order d order 3rd order 4 order Frequecy (GHz) (b) PSD of dervatves of Gaussa waveform. Fure.3 Waveforms for dervatves of Gaussa moocycle. 7

32 CHAPT : TATU VW Three tyes of modulatos are usually emloyed -UWB systems. They are ulse osto modulato (PP), ulse amltude modulato (PA), ad bary hase shft key (BPSK) or bhase modulato [6]. PP s based o e rcle of ecod formato w two or more ostos tme, referred to e omal ulse osto, as show Fure.4 (a). The rah shows - PP, where oe bt s ecoded oe mulse. Addtoal ostos ca be used to rovde more bts er symbol. The tme delay betwee ostos s tycally a fracto of a aosecod, whle e tme delay betwee omal ostos s called ulse reetto tme (PT), T r e rah, ad t s much loer to avod terferece betwee mulses. PP was mostly adoted e early develomet of UWB rados whe eat ultra short ulses were dffcult to mlemet []. PA s based o e rcle of ecod formato w e amltude of e mulses, as show Fure.4 (b). The rah shows -PA, ad more bts ca be rovded er symbol by us more levels of dfferet amltudes. BPSK or bhase modulato, formato s ecoded w e olarty of e mulses, as ca be see Fure.4 (c). a more eeral sese, BPSK ca also be rearded as a secal case of a -PA. Oer modulato techques have also bee exlored [9], [3] ad ey are bascally derved or combed from e above metoed ree basc tyes. 8

33 CHAPT : TATU VW.5 "" "".5 "" "".5 "" "" T PP Tr -.5 PA T - BPSK T (a) PP (b) PA (c) BPSK Fure.4 -UWB modulatos. -UWB trasmts formato by sed extremely short ulses w low duty cycle, so e trasmtted ower ca be small. t requres o carrer modulato ad F ower amlfer whch results smle archtecture. ts fe tme resoluto roerty makes t robust to multa fad. To effectvely exlore e chael dversty, a ake recever s usually emloyed at e recever ed. 9

34 CHAPT : TATU VW. ake ecever a tycal wreless evromet, a electromaetc () wave traverses a multtude of as from e trasmtter to e recever, whose les are uequal eeral due to deflectos, reflectos ad scatters. The comoste sal at recever sde s e suerosto of tme shfted, amltude scaled, olarzato rotated ad hase altered coes of e trasmtted sal, ad t s called multa dstorted. The exact umber of multas deeds o e eometry of e evromet, e locato of e trascevers, e lacemet of e obstructve ad o-obstructve objects, ad e roertes of materals used costructo. eeral e chael mulse resose (C) of a multa fad chael ca be modeled as: h ( t) mδ ( t m ) m, (.4) where m ad m rereset e chael atteuato ad e roaato delay of e multa comoet ad ey are referred to as chael state formato (CS). δ ( ) s e Drac delta fucto. reresets e total umber of multa comoets ad m usually. The chael amltude resose s ormalzed such at: m m. (.5) et s ( t) rereset e trasmtted sal, e e receved sal after ass rouh e multa chael s: ( t) s( t) h( t) w( t) r

35 CHAPT : TATU VW m ( t ) w( t) s, (.6) m m where w ( t) s AWG w two sde ower desty sal eery s dsersed multa comoets.. Ths exresso dcates at e ( k ) Tf kt f dt T f y k s ˆ ) ( t ktf ( k ) Tf kt f dt T f y k γˆ r( t) s ˆ ) ( t ktf γˆ â k ( k ) Tf kt f dt T f ( ) y k s( t kt f ˆ ) γˆ Fure.5 A eeral ake recever structure. A eeral ake structure s show Fure.5. t cossts of multle fers (correlators) where each fer ca extract oe of e multa comoets rovded by e chael. The oututs of all e fers are arorately wehted ad combed to rovde e multa dversty. The equvalet matched flter (F) verso of e recever volves a matched frot-ed rocessor (FP) (matched oly to e trasmtted waveform) followed by a taed delay le (TD) ad a comber [3]. The FP resolves multa comoets whose delays dffer by at least oe ch durato, T c,

36 CHAPT : TATU VW aroxmately equal to e verse of e sread badwd. The outut of FP s assed rouh a TD fler w tas, whch are combed w maxmum rato comb (C) for e best ossble erformace. The deal S at e outut of ake recever s ve by [3]: γ m, (.7) b / m where s e bt eery of e trasmtted sal. W, (.7) recovers e b m m larest ossble S of e receved sal. The B erformace s show to be [3]: ( S ) Q( γ ) P b Q, (.8) where Q ( ) s e comlemetary cumulatve dstrbuto fucto. A deal ake s referred to as all ake (Aake), whch dcates e recever w ulmted resources ad stat adatablty so at t ca resolve all e multa comoets. To acheve s, t requres T d Tc correlators/tas, where d / T s e maxmum excess delay. There are hudreds of multa comoets a tycal UWB chael. Therefore Aake s urealstc due to e ower cosumto, des comlexty ad chael estmato [3]. t ca oly serve as a bechmark at rovdes a uer lmt to e achevable erformace. Selectve ake (Sake) ad artal ake (Pake) are two commoly emloyed comlexty-reduced ake structures whch rocess oly a subset of avalable multa

37 CHAPT : TATU VW comoets. Sake selects e b best (w larest stataeous S) out of total multa comoets ad combes em us C [33]-[35], whle Pake has fers at tract e frst as arrved at e recever [36], [44]. Sake requres fast adatablty, kowlede of stataeous values of all multas ad effcet chael estmato, ad s erefore more comlex a Pake. Systems emloy varous tyes of ake structures ad er erformaces have bee reorted [], [33]-[46]. However ake mlemetato ca be effectve oly whe e CS s avalable or ca be roerly estmated at e recever. Ths oses a major roblem UWB commucatos, sce eve f we ot to utlze oly a few (e.. fve stroest out of hudreds of) as, accurate estmato of e fve stroest chael as ad er corresod delays s requred. Chael estmato has lo bee a challee UWB systems. mulse resose estmator for UWB chaels has bee reseted [3], [47]-[5]. Comutatoal comlexty of e chael estmator [3] ad [47] creases w creas umber of multa comoets, ad s uaffordable for realstc UWB chaels. oreover, chael estmators requre e ADC to work at over-saml rate to et e correct estmato of e tme delays. W e UWB sub-aosecod-wd mulse, e ADC saml rate s e formdable rae of over GHz [] whch s very costly. s ess, a low comlexty bld ake (Bake) structure s roosed, whch does ot requre ay CS yet rovdes accetable B erformace. The saml rate requred for e recever ADC s also reatly reduced. 3

38 CHAPT 3: SYST OD & CV DSG 3 SYST OD & CV DSG The roosed system model ad recever archtecture are reset s chater. Pulse sha ad modulato of our system s dscussed. BPSK s adoted for e sake of maematcal smlfcato a multa door chael models are emloyed ad e lare umber of multas s observed. The Bake recever structure s e roosed. t cossts of mult braches of ake fers ad each brach s samled at dfferet rates. The recever model arameters are chose based o chael delay rofles. 4

39 CHAPT 3: SYST OD & CV DSG 3. Pulse-sha & odulato As dscussed chater, e d order dervatve of a Gaussa ulse, ( t) ve (.3), s adoted our system des sce t s oe of e mostly used mulses -UWB ad t ca be easly eerated for smulato. However e alorm reseted s ess s also alcable to oer ulse shaes. eeral, e trasmtted waveform of a -UWB system s ve by []: TH s( t) b / t T c T d f f c, (3.) / f where s e ulse eery at e trasmtter ed. To crease S at e recever ed, each symbol s rereseted by f relcas, each relca trasmtted er frame of durato T f >> T. T f s also kows as PT. For mult-user access (A), TH c reresets e usersecfc tme ho (TH) code. very frame s dvded to c chs w each of durato T c, ad TH c T corresods to a tme shft dur e c frame. W [, ] m k deot e -ary formato symbol trasmtted dur k symbol durato, (3.) subsumes PP ad PA. Whe d m ad b, (3.) models -ary k k k PP for UWB trasmssos, ad d k s e ulse osto offset. Whe bk mk ad d k, (3.) descrbes -ary PA. TH To smlfy e aalyss s ess, sle user evromet s cosdered ( c ) ad BPSK ( d ) modulato s emloyed our system. oreover, each symbol s / f 5

40 CHAPT 3: SYST OD & CV DSG trasmtted by oe ulse to crease e trasmsso rate ( ). The e trasmtted sequece s much smlfed from (3.) ad ca be wrtte as: f where { ±} ( t T ) s ( t) b, (3.) b s e bary data trasmtted ad s s T reresets e bt durato. The quatty Ts φ s called e rocess a ad should be much reater a to T mmze e ter symbol terferece (S). At e same tme, trasmsso rate decreases w creas φ. Thus comromse must be made betwee erformace ad caacty. Alouh e erformace aalyss s ess s carred out us BPSK modulato, e results ca be easly exteded to PP or oer modulato schemes. 6

41 CHAPT 3: SYST OD & CV DSG 3. ulta Chael odel Sce most of e evsoed UWB commercal alcatos wll be door commucatos, UWB door chaels are cosdered s ess. The 8.5.3a Task Grou roosed a door chael model whch s a modfed verso of e well kow Saleh-Valezuela (S-V) model ovember [4]. S-V door chael model was establshed 987 based o measuremets utlz low ower ultra-short ulses of wd s ad ceter frequecy.5 GHz a medum sze, two-story offce buld [5]. ulta comoets arrve at e recever rous (clusters) s model. ultle rays are observed w each cluster. The cluster ad ray arrvals are bo Posso dstrbuted. ayleh dstrbuto s assumed for e multa a matude w a exoetally decay ower rofle. The cluster heomeo has bee exermetally cofrmed characterz UWB door chaels. Thus UWB door chael models are derved from S-V model w a coule of slht modfcatos. A loormal dstrbuto raer a a ayleh dstrbuto for e multa a s adoted sce t fts e measuremet data better [4]. addto, deedet fad s assumed for each cluster as well as each ray w e cluster. The multa chael model s erefore exressed as [4]: h K ( t) X δ ( t T ) l k, (3.3) k, l l k, l where X s e lo-ormal shadow, k, l s e multa a coeffcet of e k multa of e l cluster, T l s e delay of e l cluster ad k, l s e delay of k 7

42 CHAPT 3: SYST OD & CV DSG multa comoet relatve to e l cluster arrval tme. Cluster ad ray arrvals are deedet Posso dstrbuted w rate Λ ad λ, resectvely, ad λ > Λ. The total umber of as s defed as e umber of multa arrvals w exected ower w db from at of e stroest arrval a. The detaled UWB multa chael characterstcs ad arameters for chael model to chael model 4 (C-C4) are rovded [4]. Tycal realzatos of e chael mulse resose eerated for C-C4 are show Fure 3.. t ca be see at e multa comoets for each tye of e chael are eormous. For C3 ad C4, ere are eve more multas reset ad e dsersed chael le s also loer a at for C ad C. Whe do erformace aalyss chater 4, e smlfed C model ve (.4) s adoted stead of (3.3), ad e chael s vewed as cosst of dscrete tme mulse resoses. 8

43 CHAPT 3: SYST OD & CV DSG Amltude Tme (s) Amltude Tme (s) (a) C (b) C.5.. Amltude Amltude Tme (s) Tme (s) (c) C3 (4) C4 Fure 3. Tycal Cs for UWB door chaels. 9

44 CHAPT 3: SYST OD & CV DSG 3.3 Bake ecever Structure ( t) v,k v, K v, z z c,k c, K z c, s( t) h( t) w( t) r( t) ( t) v,k v, K v, z z z c,k c, K c, d ( t) v, v K, K v, z z z c, K c, K c, y e bˆ Fure 3. Bake recever archtecture. The roosed Bake recever structure s show Fure 3.. t cossts of braches w K tas each brach. The receved sal r ( t) s delayed by a small tme each cosecutve brach to esure at e recever does ot samle at e same tme stace. After at e sal s correlated w a locally eerated ulse ( t) ad e samled at dfferet rates every brach. ach brach reresets a Bake brach w a uque saml terval {,,, } saml rate, < < K <, where e fastest l K s well below e yqust rate. The samled data are roerly wehted by e ta coeffcets {, c,, c, c, c, K, c, K c } c,,, K,,, K,, K K ad summed u to form e decso sal y. The decso rule s: 3

45 CHAPT 3: SYST OD & CV DSG Dur e tra hase, e TD coeffcet set { } least mea square (S) meod [5]: ˆ f y > b. (3.4) f y < c, s udated adatvely us e l k e d y, (3.5) c c µ e v, (3.6) l, k l, k l, k where d reresets e desred sal, e reresets e error, l k c, reresets e adatato value of ta c,, l {,, K, }, l k k {,, K, K}, ad l k v, s ut of e k ta e After { } l k l brach for symbol. c, covere to er otmum values, e alorm wll exam e atter of e ta values. There are certa tas at do ot samle at ay correct tme stace of e multa, ad ese ta values covere to very small umbers ear zero. Thus ey ca be dscarded sce clud em oly adds e ose comoet ad cosumes more hardware. O e oer had, whe e ta haes to samle at e correct tme stace of ay multa, e ta value wll covere to e chael atteuato value corresod to at delay tme, ad e successfully samled eery wll be couted to e decso sal. ore detals o s wll be revealed e ext chaters. From e roosed structure ad alorm, t ca be calculated at e total saml terval for each receved data bt s {,,, } K. The choce of, K, K ad are closely related to e chael rofle. eferr to e ower 3

46 CHAPT 3: SYST OD & CV DSG delay rofle (PDP) of C-C4, t s foud at e chael multa comoets ad eery reduce exoetally w tme delay [4]. Thus t makes sese to samle at a reduced rate for cosecutve braches,.e., < < K <. At e same tme, e total saml terval should cover most of e dsersed chael eery. The S delay sread, defed as e secod cetered momet of e delay sectrum of a stochastc chael, s commoly used to descrbe a dsersed chael PSD: σ ( ) S, (3.7) where S S h h ( ) ( ) d d ad ( ) S s e delay sectrum. The S delay sread for h C-C4 s 5 s, 8 s, 4 s, ad 6 s [4], resectvely. As a suesto, t may be set:,,, K,.< ξ <. 5, (3.8) ξ ad ( ) ( K ) ξ S,.5 < ξ < 4. (3.9) Accord to our vestato ad smulato, e above suested rae of ξ ad ξ ca acheve ood B erformace by catur eouh multa eery. 3

47 CHAPT 4: PFOAC AAYSS 4 PFOAC AAYSS Thorouh erformace aalyss of e roosed recever s carred out s chater. Frstly correlato recevers are dscussed because ake structure cossts of a umber of correlators. To determe e otmum ta wehts, mum ea Square rror (S) crtero s adoted ad Weer soluto s derved based o s. However, e ta wehts obtaed ractcally are adated us S techque, where ere s a addtoal error term assocated w e adatato ste sze. The outut S s e derved cosder e S error, from whch e fal B exresso s obtaed. 33

48 CHAPT 4: PFOAC AAYSS 4. Correlato ecevers The ake structure s used to roduce a wehted sum of e oututs from a set of correlato recevers. So e erformace aalyss bes w e sal ass rouh e correlato recevers. eferr to Fure 3., some vectors are defed whch wll aear s chater: Tra sequece: T b [ b bk b ] ; (4.) Chael atteuatos: ; (4.) m [ ] T K Chael delays: ; (4.3) m [ ] T K ecever saml tmes: T [,, K, K ], (4.4) lk where, l,k, k,k, reresets e saml tme of e brach; l, k K k ta e l Gaussa ulse vector based o real chael tm: ( t T ) ( t T ) ( t T ) T K ] ; (4.5) m [ s s s 34

49 CHAPT 4: PFOAC AAYSS Gaussa ulse vector based o recever saml tm: ( t T ) ( t T ) ( t T ) T ] ; (4.6) l, k [ s, s, K s, K Ta weht: T c [ c, c, K c, K c, c, K c, K K c, K ]. (4.7) From (.4), (.6) ad (3.), e receved sal r ( t) ca be exressed as: ( t) s( t) h( t) w( t) r b s m m m ( t T ) δ ( t ) w( t) ( t T ) w( t) b m s m. (4.8) m Due to e low duty cycle roerty of -UWB trasmssos, t s reasoable to assume at ere s o S reset. For smlfcato of aalyss, e waveform for detecto of e symbol ca be wrtte as: r ( t) b ( t T ) w( t) b T m w( t) m s m m m. (4.9) After e ulse correlator ad ADC saml, e k ta ut of e l brach s: v l, k Ts T l, k Ts l, k Ts T l, k Ts l, k r ( t) ( t T ) b s T m m l, k dt ( t T ) s l, k dt 35

50 CHAPT 4: PFOAC AAYSS ( t T ) Ts T l, k w t Ts s dt l, k l, k ( l, k m ) wl k, (4.) b l, k, where Ts T l, k ( ) ( t T ) ( t T )dt l, k m, (4.) Ts l, k s m s l, k ad w, s e addtve ose after correlato oerato. ts ower ca be calculated as: l k σ w ' [ w ] wl, k l, k l, k Ts T l, k w t s l k l, k Ts T l, k ( t T ) dt w( t' ) ( t' T ) T s Ts s dt', l, k l, k [ ] ( t T ) Ts T l, k w t Ts s dt l, k l, k ( t T ) Ts T l, k Ts s dt l, k l, k. (4.) Thus.e., w, s Gaussa dstrbuted w zero mea ad two sde ower desty l k, l, k w. Sce e ulse eery s ormalzed,.e., ( ),,, w l, k. t s obvous at ( ) l, k m whe l k m > T,, as e ulse durato of ( t) s 36

51 CHAPT 4: PFOAC AAYSS T. Assume ere are fers at correctly samle e multa comoets out of e K total fers, e (4.) ca be wrtte as: v wl, k b ( ) w whe l, k l, k whe l, k l, k m m < T > T, (4.3) K, s a subset of vector m where { },, rereset e correctly samled chael mulse resose. et rereset e successfully collected chael eery rato:, (4.4) e e mssed chael eery rato s. Sort v l, k order of ad costellate to form e ut ta vector: where: [ w ( ) w b ( ) w K b ( ) v b w b s w w K w K w ] T K, (4.5) [ ( ) ( ) K ( ) ] T s, (4.6) [ w w ] T w, (4.7) K w [ w w w ] T w. (4.8) K K K The e detecto sal after weht comb s, exressed vector form: 37

52 CHAPT 4: PFOAC AAYSS y c T v. (4.9) 38

53 CHAPT 4: PFOAC AAYSS 4. S Crtero & Weer Soluto To determe e otmum weht vector c o, S crtero s emloyed ad c o ca be exressed as: co ar m J, (4.) c where J [ d y ], (4.) reresets e mea square error (S) cost fucto. The well kow Weer soluto ves [5]: c o T ( [ vv ]) [ d v]. (4.) T To solve (4.), we exam [ vv ] ad [ d v] searately e follow. From (4.5), ad ote at sal ad ose (w zero mea) are ucorrelated: b T s w T T T [ vv ] [ b s w ] w K K w S K, (4.3) where 39

54 CHAPT 4: PFOAC AAYSS 4 O S, (4.4), (4.5) K K K, (4.6) ad P P deotes e P P detty matrx. Therefore, [ ] K T S vv. (4.7) The desred sal for e data bt s ve by: T f b dt t b d. (4.8) From (4.5) ad (4.8), [ ] K K s w w s v b b d, (4.9) where K deotes e colum vector w K zero elemets. Substtute (4.7) ad (4.9) to (4.): K K o s S c s S. (4.3) From (4.4), t s observed at S s a rak- symmetrc matrx. Us eevalue decomosto [5], t ca be wrtte as:

55 CHAPT 4: PFOAC AAYSS T S QΛ SQ, (4.3) where ( ) ( ) s e matrx cota all e eevectors of Q, (4.3) O ( ) Λ S S ad, (4.33) s a daoal matrx w equvalet eevalues of S : ς,,, K, (4.34) as e daoal elemets. ecall e matrx verso lemma [5]: ad let ( A BCD) A A B[ C DA B] DA, (4.35) A, B Q, C Λ S, T D Q, e: T S ( QΛ Q ) S T T [ Q Q] Q QΛ S. (4.36) quatos (4.37)-(4.39) below derve e exresso for S ste by ste: Λ T S Q Q ( ) ( ) ( ) O 4

56 CHAPT 4: PFOAC AAYSS 4 O O O. (4.37) T T S Q Q Q QΛ O O O O O

57 CHAPT 4: PFOAC AAYSS 43 O. (4.38) Substtute (4.5) ad (4.38) back to (4.36): Ι Ι Ι S O 4 O S 4. (4.39) To obta e otmum value of ta weht from (4.3), s S 4 S

58 CHAPT 4: PFOAC AAYSS 44 4 O 4 4 γ, (4.4)

59 CHAPT 4: PFOAC AAYSS where γ s e ut S defed as: γ, (4.4) Comar (4.3) ad (4.4), e otmum ta wehts ca be obtaed as: c o ( ) γ,, K,. (4.4),, K, K 45

60 CHAPT 4: PFOAC AAYSS 4.3 S Aalyss The otmum ta wehts obtaed Secto 4. are derved us Weer soluto. ractcal mlemetatos, e ta wehts are adated recursvely us S techque. Thus ere s a addtoal error term assocated w e adatato ste sze µ for each ta weht as follows: c c e, (4.43) S o S where e varace of e error s ve by [53]: σ σ µ J m e S e µλ, (4.44) whe e ta value coveres. The mea value of e error s zero whe e ta value coveres. Here Weer soluto s acheved [53]: J m s e mmum value of e cost fucto (4.) whe e otmum T o T [ vv ] c o J d m c, (4.45) ad λ s e eevalue assocated w e autocorrelato matrx of e ut data ve by (4.3). From (4.3), (4.5), (4.6), (4.3) ad (4.33), e eevalue of (4.3) ca be derved as:,, K, λ. (4.46),, K, K 46

61 CHAPT 4: PFOAC AAYSS 47 By substtut (4.3), (4.8), (4.3) ad (4.4) to (4.45), e mmum value of e cost fucto ca be calculated as: s S S s S K m T b J. [ ] s S s T [ ] [ ] T K K K K γ. (4.47) From (4.44), (4.46) ad (4.47), e varace of e S error term, for,,, K, ca be exressed as: m J e µ γ µ µλ µ σ. (4.48) Smlarly, for K,,, K, e S error term s:

62 CHAPT 4: PFOAC AAYSS 48 m J e µ γ µ µλ µ σ. (4.49) The decso sal, y, as show (4.9), ca be exressed as e sum of two terms: K S S K S v c v c v c y v c T. (4.5) From (4.5), (4.4), ad (4.43), S S w b e v c γ S S e b w w e b γ γ. (4.5) Smlarly, t ca be show: K S K S K S w e w e v c. (4.5) Substtut (4.5) ad (4.5) back to (4.5), e decso sal s show to be: K S S v c v c y K S S S w e e b w w e b γ γ

63 CHAPT 4: PFOAC AAYSS 49 K S S w e e b w b γ γ. (4.53) The frst term e above equato reresets e sal orto, whle e rest ree terms are e ose duced by e chael as well as S adatato rocess.

64 CHAPT 4: PFOAC AAYSS B xresso The erformace of e roosed recever s evaluated by e B t ca acheve. From (.8), e B s determed by e recever outut S. eferr to (4.53), e outut S ca be obtaed by tak e rato of e sal ower over e exected ose ower e decso sal: K S S o w e e b w b S γ γ. (4.54) Assume e S error s ucorrelated w e ose [5],.e., [ ] [ ] [ ] K j K w e w e j S j S K K,,,,,,. (4.55) The e deomator of (4.54) s smly e sum of e exectatos of e ree dvdual square terms, sce e ose ad S error bo have zero mea. The ree exected square terms are evaluated oe by oe e follow. γ γ w w

65 CHAPT 4: PFOAC AAYSS 5 [ ] γ σ γ w w γ. (4.56) e S e b σ. (4.57) Substtute (4.48) to (4.57): e b S µ γ µ. (4.58) [ ] [ ] K e e w K w e K S K S w e w e σ σ σ σ σ. (4.59) Substtut (4.48) ad (4.49) to (4.59), t ca be show: K w e K S µ γ µ µ γ µ. (4.6) The deomator of (4.54) s e sum of (4.56), (4.58) ad (4.6):

66 CHAPT 4: PFOAC AAYSS 5 K S S w e e b w γ K µ γ µ µ γ µ µ γ µ γ γ γ µγ γ γ µγ γ γ γ µ γ K. (4.6) Substtute (4.6) to (4.54), e outut S ca be exressed as: γ γ µγ γ γ µγ γ γ γ µ γ γ o K S γ µγ γ γ µγ γ γ γ µ K. (4.6)

67 CHAPT 4: PFOAC AAYSS f e ste sze µ s very small ad aroaches to, e frst term e deomator of (4.6) ca be ored ad e S term s much smlfed: S o γ. (4.63) γ Therefore e B ca be easly obtaed from e outut S as (.8): ( S ) Q( γ ) Pb Q o. (4.64) 53

68 CHAPT 4: PFOAC AAYSS 4.5 sadjustmet The msadjustmet of e cost fucto due to e S adatato s ve by [5]: ( ) µ T tr( [ vv ]) J Μ ex, (4.65) J where tr ( ) deotes e trace oerato o e matrx ad ( ) S. m J reresets e excess ex From (4.3), (4.4)-(4.6), tr T ( [ vv ]) tr( S ) tr( ) tr( ) K ( ) ( K ) K. (4.66) Substtut (4.66) to (4.65), e uer boud of e msadjustmet ca be show to be: µ K µ K Μ. (4.67) γ low ower alcatos (small ), e msadjustmet ca be cotrolled to be very small w small eouh adatato ste sze µ. 54

69 CHAPT 5: SUATO SUTS 5 SUATO SUTS s chater, extesve smulatos are carred out to verfy e eoretcal results derved chater 4 as well as to comare e erformace of e roosed Bake w oer reular ake structures. The smulato setu s as follows: e d order dervatve Gaussa moocycle as (.3) ad chael models C-C4 as 8.5.3a [4] are emloyed. BPSK modulato s adoted w e trasmsso rate at Hz. The ulse wd should be arrow eouh to mmze e S ad ft to e FCC sectrum, whle ulses w too arrow wd are dffcult to eerate mlemetato ad requre extremely hh saml rate. Cosder e tradeoff, e ulse wd s set to.5 s. Therefore e badwd of UWB sal s aroxmately 4 GHz. As s s a mulse rado scheme where o carrer s used, e oerat frequecy bad would be to 4 GHz. The emsso ower ca be adjusted to meet e FCC frequecy mask. ote Carlo realzatos w, ulses each tral are used for e B calculato. For C-C, 5 ake braches are used w 8 tas each brach as we have foud s sett ca acheve comettve erformace to oer ake structures ad e mrovemet s ot sfcat by add more braches or tas. eferr to secto 3.3 ad equatos (3.8) ad (3.9), e saml rates for every brach are set to.5 s,.65 s,.85 s,. s ad.4 s, resectvely; s set to be.65 s. For C3-C4, 6 ake braches are used w 8 tas each brach sce ese two chaels have more multa comoets a C-C. 55

70 CHAPT 5: SUATO SUTS 5. Theoretcal versus Smulated B S alorm s used for ta estmato. Fure 5. catures e weht ( absolute value) for every ta after alorm coveres C whe S s 5 db. t s ot easy to fd out e reshold value to dscard e ms-samled tas s rah aloe. order to clearly obta e roer reshold, some techques are emloyed as follows. Frstly e st order dervatve of e tab weht s calculated ad lotted Fure 5. (a). The ese data are u-samled by a factor of 4 sce e oral data ool s too small (cotas oly 39 data) for ts dstrbuto atter to be foud. ext e hstoram of e u-samled data set s lotted Fure 5. (b). t ca be easly see from Fure 5. (b) at e cut-off ot should be set at.. Back to Fure 5. (a), e reshold value. corresods to ta. The dfferece of ta value after ta are very small, ad t ca be see from Fure 5. at e ta wehts after ta (ta to ta 4) are very small, whch meas e successfully catured eery rato was aroud /4.5. So for C e eoretcal B calculato was set to.5. Based o s estmato of, e eoretcal B erformace curve s lotted accord to (4.64), whch matched very well to e smulated B curve, as show Fure

71 CHAPT 5: SUATO SUTS.4. Dscard Ta Ta Weht Ta umber Fure 5. Ta weht. Weht Dfferece Ta umber (a) Ta weht dfferece. P robablty D esty Sfcat Ta Weht Dfferece Ta Weht Dfferece (b) Hstoram after usaml. Fure 5. Ta weht dfferece ad ts hstoram. 57

72 CHAPT 5: SUATO SUTS - B C Theoretcal C Smulated C Theoretcal C Smulated C3 Theoretcal C3 Smulated C4 Theoretcal C4 Smulated S (db) Fure 5.3 Theoretcal versus smulated B for C-C4. The curves for C-C4 Fure 5.3 are obtaed e same maer. ractce, ere s o eed to clude all e tas e Bake comb. After tal adatato ad estmato by tra data set, roer reshold s obtaed ad small-value-tas are dscarded. Therefore e recever comlexty ca be furer smlfed. s smulato, e tal set-u was 4 fers for C-C ad 48 fers for C3-C4. After ta coverece, e rema useful fers were less a, erefore reatly smlfed e recever comlexty. 58

73 CHAPT 5: SUATO SUTS 5. ffect of µ o B Aroxmato secto 4.4, e S exresso s much smlfed from (4.6) to (4.63) whe we assume e ste sze µ s very small,.e., t s assumed: F γ K µ, (5.) γ ( ) γ µγ γ γ µγ w small eouh µ. Ths secto exames e effect of µ o e valdty of e aroxmato. s smulato,, ad K were set to.5,, 5 ad 8, resectvely, whch are cosstet w e revous secto. To esure e stablty of e S alorm, e ste sze µ s bouded by < µ < [5] where s e autocorrelato matrx ve by (4.3). From (4.66), 3tr e trace of e matrx s also assocated w e ower of ut sal ad ut S. For a fxed ulse eery, e smaller e ut S, e ber e ose ower, ad e ber e trace. So µ should be uer bouded by e smallest ut S whch s db our smulato. t s also foud from (5.) at F s roortoally related w. However ber results ber tr ad smaller uer boud of µ, erefore smaller F. To rove s, two smulatos were carred out w set to 3 ad 3, resectvely. The B s obtaed us (.8) ad (4.6). Whe 3, t s calculated at µ. t ca be see from Fure 5.4 (a) at whe µ s less a 3 max.6.6 3, e B erformace s almost ot affected by e varato of µ. oer 59

74 CHAPT 5: SUATO SUTS words t s reasoable to dscard F, e frst term of e deomator (4.6), f 3 µ <.6. Whe 3, t ca be calculated at µ. The 5 max.6 smulato results s show Fure 5.4 (b) ad e same arumet ales. Therefore e aroxmato from (4.6) to (4.63) s very arorate. - - B Ste Sze mu (a) B aroxmato whe 3. So db So db So 4 db So 6 db So 8 db - So db So db So 4 db So 6 db So 8 db - B X Ste Sze mu (b) B aroxmato whe 3. Fure 5.4 ffect of µ o B aroxmato. 6

75 CHAPT 5: SUATO SUTS 5.3 B Comarso w Oer ake Structures The B erformace of Bake ad covetoal Aake, ad Sake w 5 ad fers of C-C4 are show Fure 5.5 to Fure 5.8. t s see at e Bake acheves satsfactory erformace (w less a 5dB deradato for C-C ad less a db deradato for C3-C4 from Aake, resectvely) wout emloy chael estmato. ote e saml rate for each brach of Bake s at sub-yqust sal rate. our setu, e ulse durato s.5 s ad e saml rate for e frst brach s.5 s, whch s at oe quarter of e yqust rate (.5 s). Saml rates for subsequet braches are eve less as ere are less eery we ca collect at lare delay tmes. Ths reatly relaxes e ADC requremets ad sfcatly reduces e ower cosumto of e recever ractcal mlemetatos. Aake Sake- Sake-5 Bake - B S (db) Fure 5.5 B erformace for C. 6

76 CHAPT 5: SUATO SUTS - Aake Sake - Sake-5 Bake B S (db) Fure 5.6 B erformace for C. - Aake Sake- Sake-5 Bake B S (db) Fure 5.7 B erformace for C3. 6

77 CHAPT 5: SUATO SUTS - Aake Sake- Sake-5 Bake B S (db) Fure 5.8 B erformace for C4. 63

78 CHAPT 6: COCUSO 6 COCUSO 6. Cocluso emarks s ess a ovel Bake recever structure s roosed for UWB trasmsso systems. The erformace aalyss of e roosed alorm has bee carred out w closed form B exresso followed by smulatos whch rove e valdty of e eoretcal aalyss ad e feasblty of e mlemetato. The frst two chaters ve a bref troducto of UWB hstory ad des cosderatos. Two ma UWB trasmsso schemes, ad OFD, are reseted ad er advataes ad drawbacks are comared. -UWB s emloyed s ess due to ts relatvely smle recever archtecture. The rcle of ake recevers for wreless alcatos s also dscussed s chater. Aake, Sake ad Pake are descrbed ad comared. The major drawback assocated w covetoal ake structure s at t requres chael formato clud multa delays ad atteuatos to be kow or correctly estmated at e recever ed, whch s ot ractcal real lfe UWB alcatos. The Bake recever archtecture s reseted chater 3. Frstly system model s dscussed. BPSK modulato s emloyed our system to smlfy e follow maematcal aalyss. ulta chael models C-C4 for UWB door chaels are adoted. Bake recever structure s e roosed w e brach ad ta umber determato crtero. Chater 4 focuses o erformace aalyss. t bes w correlato recever sce ake 64

79 CHAPT 6: COCUSO structures are bascally a combato of correlato recevers. S crtero s used to determe e otmum ta wehts ad Weer soluto s obtaed. S meod s emloyed dur adatato hase, ad ere s a error term assocated w e adatato ste sze. Ths s also dscussed s chater. Outut S s e calculated ad closed form B s obtaed based o t. xtesve smulatos are carred out ad e sfcat results are ve chater 5. Theoretcal aalyss s roved to be vald by comar w e smulated B. There s a comlcated term assocated w ste sze whch s ored revously to smlfy e S exresso. To verfy e arorateess of dro s term, e effect of e ste sze o e aroxmato s examed. Fally, e erformace of e roosed structure s also comared w Aake ad Sake. t s roved at e roosed structure requres o chael state formato ad acheves a much reduced saml rate w sfcat erformace loss comar w covetoally ake recevers. 65

CHAPTER-4 WIDE BAND PASS FILTER DESIGN 4.1 INTRODUCTION

CHAPTER-4 WIDE BAND PASS FILTER DESIGN 4.1 INTRODUCTION CHAPTER-4 WIDE BAND PASS FILTER DESIGN 4. INTRODUCTION The bad pass flters suested last chapter are hav the FBW less tha the 2%. I cotrast of that ths chapter deals wth the des of wde bad pass flter whch

More information

Geometric Distribution as a Randomization Device: Implemented to the Kuk s Model

Geometric Distribution as a Randomization Device: Implemented to the Kuk s Model It. J. Cotem. Math. Sceces, Vol. 8, 03, o. 5, 43-48 HIKARI Ltd, www.m-hkar.com Geometrc Dstrbuto as a Radomzato Devce: Imlemeted to the Kuk s Model Sarjder Sgh Deartmet of Mathematcs Texas A&M Uversty-Kgsvlle

More information

Less Complex Channel Estimation Technique for MIMO OFDM system

Less Complex Channel Estimation Technique for MIMO OFDM system Iteratoal Research Joural of Eeer ad Techoloy IRJET) e-iss: 2395-56 Volume: 3 Issue: 4 Ar-216 wwwrjetet -ISS: 2395-72 Less Comlex Chael Estmato Techque for MIMO OFDM system 1 Sadeekumar S Gamt, 2 Prof

More information

Module 6. Channel Coding. Version 2 ECE IIT, Kharagpur

Module 6. Channel Coding. Version 2 ECE IIT, Kharagpur Module 6 Chael Codg Lesso 36 Coded Modulato Schemes After readg ths lesso, you wll lear about Trells Code Modulato; Set parttog TCM; Decodg TCM; The modulated waveform a covetoal ucoded carrer modulato

More information

K-Map 1. In contrast, Karnaugh map (K-map) method provides a straightforward procedure for simplifying Boolean functions.

K-Map 1. In contrast, Karnaugh map (K-map) method provides a straightforward procedure for simplifying Boolean functions. K-Map Lesso Objectves: Eve though Boolea expressos ca be smplfed by algebrac mapulato, such a approach lacks clear regular rules for each succeedg step ad t s dffcult to determe whether the smplest expresso

More information

Short Note: Merging Secondary Variables for Geophysical Data Integration

Short Note: Merging Secondary Variables for Geophysical Data Integration Short Note: Mergg Secodary Varables for Geophyscal Data Itegrato Steve Lyster ad Clayto V. Deutsch Departmet of Cvl & Evrometal Egeerg Uversty of Alberta Abstract Multple secodary data from geophyscal

More information

Improved Multi-Channel Blind De-Convolution Algorithm for Linear Convolved Mixture DS/CDMA Signals

Improved Multi-Channel Blind De-Convolution Algorithm for Linear Convolved Mixture DS/CDMA Signals Joural of Commucatos Vol. 9, No., October 4 Imroved ult-chael Bld De-Covoluto Algorthm for ear Covolved xture DS/CDA Sgals Hao Cheg, Na Yu, ad Ju u Telecommucato Egeerg aboratory, Chegdu Uversty Natoal

More information

Sixth Edition. Chapter 7 Point Estimation of Parameters and Sampling Distributions Mean Squared Error of an 7-2 Sampling Distributions and

Sixth Edition. Chapter 7 Point Estimation of Parameters and Sampling Distributions Mean Squared Error of an 7-2 Sampling Distributions and 3//06 Appled Statstcs ad Probablty for Egeers Sth Edto Douglas C. Motgomery George C. Ruger Chapter 7 Pot Estmato of Parameters ad Samplg Dstrbutos Copyrght 04 Joh Wley & Sos, Ic. All rghts reserved. 7

More information

Successive Interference Cancellation for Optical CDMA Systems: Fundamental Principles

Successive Interference Cancellation for Optical CDMA Systems: Fundamental Principles Successve Iterferece Cacellato for Otcal CDMA Systems: Fudametal Prcles TAWFIG ELTAIF a*, HOSSAM M. H. SHALABY b, SAHBUDIN SHAARI a, MOHAMMAD M. N. HAMARSHEH c a Photocs Techology Laboratory (PTL), Isttute

More information

DISTRIBUTION VOLTAGE MONITORING AND CONTROL UTILIZING SMART METERS

DISTRIBUTION VOLTAGE MONITORING AND CONTROL UTILIZING SMART METERS 4 th Iteratoal Coferece o Electrcty Dstrbuto Glasgow, -5 Jue 07 DISTRIBUTION VOLTAGE MONITORING AND CONTROL UTILIZING SMART METERS Yoshhto. KINOSHITA Kazuor. IWABUCHI Yasuyuk. MIYAZAKI Toshba Japa Toshba

More information

5. Random Processes. 5-3 Deterministic and Nondeterministic Random Processes

5. Random Processes. 5-3 Deterministic and Nondeterministic Random Processes 5. Radom Processes 5- Itroducto 5- Cotuous ad Dscrete Radom Processes 5-3 Determstc ad Nodetermstc Radom Processes 5-4 Statoary ad Nostatoary Radom Processes 5-5 rgodc ad Noergodc Radom Processes 5-6 Measuremet

More information

BER ANALYSIS OF V-BLAST MIMO SYSTEMS UNDER VARIOUS CHANNEL MODULATION TECHNIQUES IN MOBILE RADIO CHANNELS

BER ANALYSIS OF V-BLAST MIMO SYSTEMS UNDER VARIOUS CHANNEL MODULATION TECHNIQUES IN MOBILE RADIO CHANNELS 202 Iteratoal Coferece o Computer Techology ad Scece (ICCTS 202) IPCSIT vol. 47 (202) (202) IACSIT Press, Sgapore DOI: 0.7763/IPCSIT.202.V47.24 BER ANALYSIS OF V-BLAST MIMO SYSTEMS UNDER VARIOUS CANNEL

More information

A CONTROL CHART FOR HEAVY TAILED DISTRIBUTIONS. K. Thaga. Department of Statistics University of Botswana, Botswana

A CONTROL CHART FOR HEAVY TAILED DISTRIBUTIONS. K. Thaga. Department of Statistics University of Botswana, Botswana A CONTROL CHART FOR HEAVY TAILED DISTRIBUTIONS K. Thaga Departmet of Statstcs Uversty of Botswaa, Botswaa thagak@mopp.ub.bw ABSTRACT Stadard cotrol charts wth cotrol lmts determed by the mea ad stadard

More information

A Parallel-Layered Belief-Propagation Decoder for Non-layered LDPC Codes

A Parallel-Layered Belief-Propagation Decoder for Non-layered LDPC Codes 400 JOURNAL OF COMMUNICATIONS, VOL. 5, NO. 5, MAY 2010 A Parallel-Layered Belef-Proagato Decoder for No-layered LDPC Codes Ku Guo Yog He ad Shusha Qao Asc ad System Deartmet, Isttute of Mcroelectrocs of

More information

New Difference Estimator in Two-phase Sampling using Arbitrary Probabilities

New Difference Estimator in Two-phase Sampling using Arbitrary Probabilities IN 1684-8403 Joural o tatstcs olume 15, 008,. 7-16 Abstract New Derece Estmator Two-hase amlg usg Arbtrar Probabltes Asa Kamal 1 ad Muhammad Qaser hahbaz A ew derece estmator has bee costructed two-hase

More information

A NOVEL APPROACH FOR MITIGATION OF HARMONICS AND INTERHARMONICS IN VARIABLE FREQUENCY DRIVES

A NOVEL APPROACH FOR MITIGATION OF HARMONICS AND INTERHARMONICS IN VARIABLE FREQUENCY DRIVES Joural of Theoretcal ad Aled formato Techology 2005-204 JATT & LLS. All rghts reserved. SSN: 992-8645 www.jatt.org E-SSN: 87-395 A NOVEL APPROACH FOR MTGATON OF HARMONCS AND NTERHARMONCS N VARABLE FREQUENCY

More information

Deinterleaving of Interfering Radars Signals in Identification Friend or Foe Systems

Deinterleaving of Interfering Radars Signals in Identification Friend or Foe Systems 8 Telecommucatos forum TEFOR Serba, Belgrade, ovember -5, Deterleavg of Iterferg Radars Sgals Idetfcato Fred or Foe Systems Youes Ahmad amal Mohamedpour Moe Ahmad Abstract I a dese moder electroc warfare

More information

Performance Comparison of Two Inner Coding Structures in Concatenated Codes for Frequency-Hopping Spread Spectrum Multiple-Access Communications

Performance Comparison of Two Inner Coding Structures in Concatenated Codes for Frequency-Hopping Spread Spectrum Multiple-Access Communications Iteratoal Joural o Recet ad Iovato Treds Computg ad Commucato IN: 31-8169 Volume: 3 Issue: 741-745 erformace Comparso of Two Ier Codg tructures Cocateated Codes for Frequecy-Hoppg pread pectrum Multple-Access

More information

An Improved DV-Hop Localization Algorithm Based on the Node Deployment in Wireless Sensor Networks

An Improved DV-Hop Localization Algorithm Based on the Node Deployment in Wireless Sensor Networks Iteratoal Joural of Smart Home Vol. 9, No. 0, (05), pp. 97-04 http://dx.do.org/0.457/jsh.05.9.0. A Improved DV-Hop Localzato Algorthm Based o the Node Deploymet Wreless Sesor Networks Jam Zhag, Ng Guo

More information

DYNAMIC BROADCAST SCHEDULING IN ASYMMETRIC COMMUNICATION SYSTEMS: PUSH AND PULL DATA BASED ON SCHEDULING INDEX AND OPTIMAL CUT-OFF POINT YUFEI GUO

DYNAMIC BROADCAST SCHEDULING IN ASYMMETRIC COMMUNICATION SYSTEMS: PUSH AND PULL DATA BASED ON SCHEDULING INDEX AND OPTIMAL CUT-OFF POINT YUFEI GUO DYNAMIC BROADCAST SCHEDULING IN ASYMMETRIC COMMUNICATION SYSTEMS: PUSH AND PULL DATA BASED ON SCHEDULING INDEX AND OPTIMAL CUT-OFF POINT by YUFEI GUO Preseted to the Faculty of the Graduate School of The

More information

Lecture6: Lossless Compression Techniques. Conditional Human Codes

Lecture6: Lossless Compression Techniques. Conditional Human Codes ecture6: ossless Compresso Techques Codtoal uma Codes -Cosder statoary dscrete arov process, = { s, s, s } wth codtoal pmfs P P s s wth,, tates o Ps/so.9.5.5 Ps/s.5.8.5 Ps/s.5.5.6 -The margal probabltes

More information

Descriptive Statistics

Descriptive Statistics Math 3 Lecture I Descrptve tatstcs Descrptve statstcs are graphcal or umercal methods utlsed to summarze data such a way that mportat features of the sample ca be depcted. tatstcs: tatstcs s cocered wth

More information

LECTURE 4 QUANTITATIVE MODELS FOR FACILITY LOCATION: SERVICE FACILITY ON A LINE OR ON A PLANE

LECTURE 4 QUANTITATIVE MODELS FOR FACILITY LOCATION: SERVICE FACILITY ON A LINE OR ON A PLANE LECTUE 4 QUANTITATIVE MODELS FO FACILITY LOCATION: SEVICE FACILITY ON A LINE O ON A PLANE Learg objectve 1. To demostrate the quattatve approach to locate faclt o a le ad o a plae 6.10 Locatg Faclt o a

More information

A Collaboration-based Distributed TDMA Scheduling Algorithm for Data Collection in Wireless Sensor Networks

A Collaboration-based Distributed TDMA Scheduling Algorithm for Data Collection in Wireless Sensor Networks JOUNL OF NEWOKS, VOL. 9, NO. 9, SEPEME 24 239 ollaborato-based Dstrbuted DM Schedulg lgorthm for Data ollecto Wreless Sesor Networks o Zeg ollege of omuter Scece ad echology, Zhejag Uversty, Hagzhou, ha

More information

Thermometer-to-binary Encoder with Bubble Error Correction (BEC) Circuit for Flash Analog-to-Digital Converter (FADC)

Thermometer-to-binary Encoder with Bubble Error Correction (BEC) Circuit for Flash Analog-to-Digital Converter (FADC) Thermometer-to-bary Ecoder wth Bubble Error Correcto (BEC) Crcut for Flash Aalog-to-Dgtal Coverter (FADC) Bu Va Heu, Seughyu Beak, Seughwa Cho +, Jogkook Seo ±, Takyeog Ted. Jeog,* Dept. of Electroc Egeerg,

More information

On the Techniques for Constructing Even-order Magic Squares using Basic Latin Squares

On the Techniques for Constructing Even-order Magic Squares using Basic Latin Squares Iteratoal Joural of Scetfc ad Research Publcatos, Volume, Issue 9, September 0 ISSN 50-353 O the Techques for Costructg Eve-order Magc Squares usg Basc Lat Squares Tomba I. Departmet of Mathematcs, Mapur

More information

A HIGH ACCURACY HIGH THROUGHPUT JITTER TEST SOLUTION ON ATE FOR 3GBPS AND 6GBPS SERIAL-ATA

A HIGH ACCURACY HIGH THROUGHPUT JITTER TEST SOLUTION ON ATE FOR 3GBPS AND 6GBPS SERIAL-ATA A HIGH ACCURACY HIGH THROUGHPUT JITTER TEST SOLUTION ON ATE FOR 3GBPS AND 6GBPS SERIAL-ATA Yogqua Fa, Y Ca ad Zeljko Zlc LSI Corporato 0 Amerca Parkway NE, Alletow, Pesylvaa 809 Emal: y.ca@ls.com Departmet

More information

Robot Path Planning Based on Random Coding Particle Swarm Optimization

Robot Path Planning Based on Random Coding Particle Swarm Optimization (IJACSA) Iteratoal Joural of Advaced Computer Scece ad Applcatos, Vol. 6, No. 4, 5 Robot Path Pla Based o Radom Cod Partcle Swarm Optmzato Ku Su Collee of Electroc ad Electrcal Eeer Shaha Uversty of Eeer

More information

Information Theory and Coding

Information Theory and Coding Iformato heory ad Codg Itroducto What s t all aout? Refereces: C..hao, A Mathematcal heory of Commucato, he Bell ystem echcal Joural, Vol. 7, pp. 379 43, 63 656, July, Octoer, 948. C..hao Commucato the

More information

Zigbee wireless sensor network localization evaluation scheme with weighted centroid method

Zigbee wireless sensor network localization evaluation scheme with weighted centroid method Zgbee wreless sesor etwork localzato evaluato scheme wth weghted cetrod method Loesy Thammavog 1, Khamphog Khogsomboo 1, Thaadol Tegthog 2 ad Sathapor Promwog 2,* 1 Departmet of Electrocs ad Telecommucato

More information

Comparison of Measurement and Prediction of ITU-R Recommendation P.1546

Comparison of Measurement and Prediction of ITU-R Recommendation P.1546 Comparso of Measuremet ad Predcto of ITU-R Recommedato P.546 Chag-Hoo Lee *, Nam-Ryul Jeo *, Seog-Cheol Km *, Jug-m Lm * Isttute of New Meda ad Commucatos, Seoul Natoal Uversty, Cosoldated Mateace Depot,

More information

Transformer-Coupled Loopback Test for Differential Mixed-Signal Specifications

Transformer-Coupled Loopback Test for Differential Mixed-Signal Specifications Trasformer-Couled Looback Test for Dfferetal Mxed-Sgal Secfcatos Byougo Km, Zea Fu ad Jacob A. Abraam Comuter Egeerg Researc Ceter Natoal Semcoductor Cororato Te Uversty of Texas at Aust Sata Clara, CA

More information

Iterative Enhanced Directional Based Demosaicking Algorithm

Iterative Enhanced Directional Based Demosaicking Algorithm AJayachadra et al / Iteratoal Joural o Computer Scece ad Eeer IJCSE Iterate Ehaced Drectoal Based Demosack Alorthm AJayachadra # DrDhaasekara* Prof P aa # # PS Collee of Eeer ad Techoloy TrueelIda * Prcpal

More information

A Spectrally Efficient Frequency Division Multiplexing Based Communications System M. R. D. Rodrigues and I. Darwazeh

A Spectrally Efficient Frequency Division Multiplexing Based Communications System M. R. D. Rodrigues and I. Darwazeh IOWo'03, 8th Iteratoal OFDM-Workshop, Proceedgs, Hamburg, DE, Sep 24-25, 2003 (Prepublcato draft) A Spectrally Effcet Frequecy Dvso Multplexg Based Commucatos System M. R. D. Rodrgues ad I. Darwazeh Laboratory

More information

A Fast Path Planning Algorithm for a Mobile Robot

A Fast Path Planning Algorithm for a Mobile Robot Proceeds of the World Coress o Eeer 018 Vol I WCE 018, July 4-6, 018, Lodo, U.K. A Fast Path Pla Alorthm for a Moble Robot Patece I. Adamu, IAENG, Member, Hlary I. Okabue, Pelum E. Outude ad Abodu A. Opaua

More information

Frequency Assignment for IEEE Wireless Networks

Frequency Assignment for IEEE Wireless Networks Frequecy Assgmet for IEEE 8 Wreless Networks K K Leug Bell Labs, Lucet Techologes Murray Hll, NJ 7974 k@bell-labscom Byoug-Jo J Km AT&T Labs Research Mddletow, NJ 7748 macsbug@researchattcom Abstract The

More information

Filtered Multitone (FMT) Modulation for Broadband Fixed Wireless Systems

Filtered Multitone (FMT) Modulation for Broadband Fixed Wireless Systems Fltered ulttoe FT odulato for Broadbad Fxed Wreless Systems A dssertato submtted to the Uversty of Cambrdge for the degree of aster of Phlosophy Igaco Bereguer, Hughes Hall August LABORATORY FOR COUNICATIONS

More information

An ANOVA-Based GPS Multipath Detection Algorithm Using Multi-Channel Software Receivers

An ANOVA-Based GPS Multipath Detection Algorithm Using Multi-Channel Software Receivers A ANOVA-Based GPS Multpath Detecto Algorthm Usg Mult-Chael Software Recevers M.T. Breema, Y.T. Morto, ad Q. Zhou Dept. of Electrcal ad Computer Egeerg Mam Uversty Oxford, OH 4556 Abstract: We preset a

More information

Consecutiveness Loss-Aware Routing and Spectrum Assignment Algorithm in Flexible Bandwidth Optical Networks

Consecutiveness Loss-Aware Routing and Spectrum Assignment Algorithm in Flexible Bandwidth Optical Networks 0 7 Iteratoal ICST Coferece o Commucatos ad Networkg Cha (CHINACOM) Cosecutveess Loss-Aware Routg ad Spectrum Assgmet Algorm Flexble Badwd Optcal Networks Xbo Wag Yogl Zhao Jawe Zhag Xaosog Yu Je Zhag

More information

Variable Lateral Silicon Controlled Rectifier as an ESD Protection

Variable Lateral Silicon Controlled Rectifier as an ESD Protection Varable Lateral Slco Cotrolled Rectfer as a ESD Protecto PETR BETAK, VLADISLAV MUSIL Deartmet of Mcroelectrocs, FEEC Bro Uversty of Techology Údolí 53 CZECH REPUBLIC etr.betak@hd.feec.vutbr.cz Abstract:

More information

Applied Statistics and Probability for Engineers, 6 th edition December 31, 2013 CHAPTER 6. Section 6-1

Applied Statistics and Probability for Engineers, 6 th edition December 31, 2013 CHAPTER 6. Section 6-1 Appled Statstcs ad Probablty for Egeers, 6 th edto December 31, 013 CHAPTER 6 Secto 6-1 6-1. No, usually ot. For eample, f the sample s {, 3} the mea s.5 whch s ot a observato the sample. 6-3. No, usually

More information

Switching Angle Design for Pulse Width Modulation AC Voltage Controller Using Genetic Algorithm and Distributed Artificial Neural Network

Switching Angle Design for Pulse Width Modulation AC Voltage Controller Using Genetic Algorithm and Distributed Artificial Neural Network Swtchg Agle Desg for Pulse Wdth Modulato AC Voltage Cotroller Usg Geetc Algorthm ad Dstrbuted Artfcal Neural Network Pattarapor Jtta, Somyot Katwadvla ad Atthapol Ngaoptakkul Abstract. Ths paper proposes

More information

A New Mathematical Model for a Redundancy Allocation Problem with Mixing Components Redundant and Choice of Redundancy Strategies

A New Mathematical Model for a Redundancy Allocation Problem with Mixing Components Redundant and Choice of Redundancy Strategies Appled Mathematcal Sceces, Vol, 2007, o 45, 222-2230 A New Mathematcal Model for a Redudacy Allocato Problem wth Mxg Compoets Redudat ad Choce of Redudacy Strateges R Tavakkol-Moghaddam Departmet of Idustral

More information

Provded by the author(s) ad Uversty College Dubl Lbrary accordace wth publsher polces. Please cte the publshed verso whe avalable. Ttle Gaussa Pulse-Based Two-Threshold Parallel Scalg Toe Reservato for

More information

Silicon Limit Electrical Characteristics of Power Devices and ICs

Silicon Limit Electrical Characteristics of Power Devices and ICs Slco Lmt Electrcal Characterstcs of Power evces ad Cs Ako Nakagawa, Yusuke Kawaguch ad Kazutosh Nakamura Toshba Cororato, Semcoductor Comay 58-, Horkawa-Cho, Sawa-Ku, Kawasak, -85, Jaa; E-mal:ako.akagawa@toshba.co.j

More information

Low Complexity LMMSE Channel Estimation on GPP

Low Complexity LMMSE Channel Estimation on GPP 0 7th Iteratoal ICT Coferece o Commucatos ad etworkg Cha (CIACOM) Low Complexty LMME Chael Estmato o GPP We h, Tao Peg, Rogrog Qa Key Lab. Of Uversal Wreless Commucato, Mstry of Educato, BUPT Bejg 00876,

More information

A Novel Bandwidth Optimization Manager for Vehicle Controller Area Network, CAN, System

A Novel Bandwidth Optimization Manager for Vehicle Controller Area Network, CAN, System A Novel Badwdth Optmzato Maager for Vehcle Cotroller Area Network, CAN, System Y WANG, Z-y YOU, L-hu HUI Guzhou Normal Uversty, Guyag, Guzhou 550001, Cha Abstract Ths paper cosders the badwdth lmtato of

More information

Optimal Power Allocation in Zero-forcing MIMO-OFDM Downlink with Multiuser Diversity

Optimal Power Allocation in Zero-forcing MIMO-OFDM Downlink with Multiuser Diversity Optmal Power Allocato Zero-forcg IO-OFD Dowl wth ultuser Dversty Peter W. C. Cha ad oger S.. Cheg Abstract hs paper cosders the optmal power allocato for a multuser IO-OFD dowl system usg zero-forcg multplexg

More information

Distributed Online Matching Algorithm For Multi-Path Planning of Mobile Robots

Distributed Online Matching Algorithm For Multi-Path Planning of Mobile Robots Proect Paper for 6.854 embers: Seugkook Yu (yusk@mt.edu) Sooho Park (dreameo@mt.edu) Dstrbuted Ole atchg Algorthm For ult-path Plag of oble Robots 1. Itroducto Curretly, we are workg o moble robots whch

More information

COVERAGE ESTIMATION FOR MOBILE CELLULAR NETWORKS FROM SIGNAL STRENGTH MEASUREMENTS. Kanagalu R. Manoj, BE and MS DISSERTATION

COVERAGE ESTIMATION FOR MOBILE CELLULAR NETWORKS FROM SIGNAL STRENGTH MEASUREMENTS. Kanagalu R. Manoj, BE and MS DISSERTATION COVERAGE ESTIMATION FOR MOBILE CELLULAR NETWORKS FROM SIGNAL STRENGTH MEASUREMENTS by Kaagalu R. Maoj, BE ad MS DISSERTATION Preseted to the Faculty of The Uversty of Texas at Dallas Partal Fulfllmet of

More information

Chaotic Maps: A Tool to Enhance the Performance of OFDM Systems

Chaotic Maps: A Tool to Enhance the Performance of OFDM Systems Chaotc Maps: A Tool to Ehace the Performace of OFDM Systems E. M. El-Bakary, O. Zahra, S. A. El-Doll, ad F. E. Abd El-Same Departmet of Electrocs ad Electrcal Commucatos Egeerg, Faculty of Electroc Egeerg,

More information

From Exploring to Optimal Path Planning: Considering Error of Navigation in Multi-Agent Mobile Robot Domain

From Exploring to Optimal Path Planning: Considering Error of Navigation in Multi-Agent Mobile Robot Domain Acta Polytechca Hugarca Vol., No. 6, 4 From Exlorg to Otmal Path Plag: Cosderg Error of Navgato Mult-Aget Moble Robot Doma Istvá Nagy Óbuda Uversty, Bák Doát Faculty of Mechacal ad Safety Egeerg Isttute

More information

An ID-based Proxy Authentication Protocol Supporting Public Key Infrastructure

An ID-based Proxy Authentication Protocol Supporting Public Key Infrastructure A ID-based Proxy Authetcato Protocol Supportg Publc Key Ifrastructure Shuh-Pyg Sheh, Shh-I Huag ad Fu-She Ho Departmet of Computer Scece ad Iformato Egeerg, ABSTRACT The advatage of the ID-based authetcato

More information

Automatic Modulation Classification with Genetic Backpropagation Neural Network

Automatic Modulation Classification with Genetic Backpropagation Neural Network Automatc odulato Classfcato wth Geetc Backpropagato eural etwork Qal Zhou, Hu Lu, Lwe Ja, Kefe ao Electroc ad Iformato Egeerg Behag Uversty Beg 009, P.R.Cha Emal:xa5@buaa.edu.c, mluhu@buaa.edu.c,lwe_008@63.com,

More information

A New Aggregation Policy for RSS Services

A New Aggregation Policy for RSS Services A New Aggregato Polcy for RSS Servces Youg Geu Ha Sag Ho Lee Jae Hw Km Yaggo Km 2 School of Computg, Soogsl Uversty Seoul, Korea {youggeu,shlee99,oassdle}@gmal.com 2 Dept. of Computer ad Iformato Sceces,

More information

OPTIMAL BUS DISPATCHING POLICY UNDER VARIABLE DEMAND OVER TIME AND ROUTE LENGTH

OPTIMAL BUS DISPATCHING POLICY UNDER VARIABLE DEMAND OVER TIME AND ROUTE LENGTH OPTIMAL BUS DISPATCHING POLICY UNDER VARIABLE DEMAND OVER TIME AND ROUTE LENGTH Amal S. Kumarage, Professor of Cvl Egeerg, Uversty of Moratuwa, Sr Laka H.A.C. Perera, Cetral Egeerg Cosultacy Bureau, Sr

More information

PERFORMANCE ANALYSIS OF SUBOPTIMAL SOFT DECISION DS/BPSK RECEIVERS IN PULSED NOISE AND CW JAMMING UTILIZING JAMMER STATE INFORMATION

PERFORMANCE ANALYSIS OF SUBOPTIMAL SOFT DECISION DS/BPSK RECEIVERS IN PULSED NOISE AND CW JAMMING UTILIZING JAMMER STATE INFORMATION PRFORMC YI OF UBOPTIM OFT DCIIO DBPK RCIVR I PUD OI D CW MMIG UTIIZIG MMR TT IFORMTIO UHI UTTI Departmet of lectrcal ad Iformato geerg Uversty of Oulu OUU 4 UHI UTTI PRFORMC YI OF UBOPTIM OFT DCIIO DBPK

More information

Adaptive QoS Control for Real-time Video Communication over Wireless Channels

Adaptive QoS Control for Real-time Video Communication over Wireless Channels Adaptve QoS Cotrol for Real-tme Vdeo Commucato over Wreless Chaels Dapeg Wu Y. Thomas Hou Wewu Zhu Zhha He Ya-Q Zhag Abstract Robust trasmsso of real-tme vdeo over wreless chaels s a challegg problem.

More information

The optimization of emergency resource-mobilization based on harmony search algorithm

The optimization of emergency resource-mobilization based on harmony search algorithm Avalable ole www.ocpr.com Joural of Chemcal ad Pharmaceutcal Research, 04, 6(7):483-487 Research Artcle ISS : 0975-7384 CODE(USA) : JCPRC5 The optmzato of emergecy resource-moblzato based o harmoy search

More information

OPTIMAL DESIGN OF BROADBAND WIRELESS MESH NETWORKS

OPTIMAL DESIGN OF BROADBAND WIRELESS MESH NETWORKS OTIMAL DESIGN OF BOADBAND WIELESS MESH NETWOKS A. Bela, A. S. Haf, a M. Gereau Networ esearch Lab, Uversty of Motreal {abela, ahaf}@ro.umotreal.ca, mchel.gereau@crrelt.ca Abstract- Desg/lag of WMNs s the

More information

Speculative Completion for the Design of High-Performance Asynchronous Dynamic Adders

Speculative Completion for the Design of High-Performance Asynchronous Dynamic Adders I: 1997 IEEE Iteratoal Symposum o Advaced Research Asychroous Crcuts ad Systems ( Asyc97 Symposum), Edhove, The Netherlads Speculatve Completo for the Desg of Hgh-Performace Asychroous Dyamc Adders Steve

More information

Weighted Centroid Localization in Zigbee-based Sensor Networks

Weighted Centroid Localization in Zigbee-based Sensor Networks Weghted Cetrod Localzato Zgbee-based Sesor Networks Ja Blumethal**, Ralf Grossma, Frak Golatowsk*, Drk Tmmerma** * CELISCA, Ceter for Lfe Scece Automato Fredrch-Barewtz-Str. 8 Uversty of Rostock ** Faculty

More information

Application of SQRD Algorithm for Efficient MIMO-OFDM Systems

Application of SQRD Algorithm for Efficient MIMO-OFDM Systems Applcato of SQD Algorthm for Effcet MIMO-OFDM Systems Clemes Mchalke, rshkesh Vekatarama, V. Sha, W. ave, ad G. Fettwes. Vodafoe Char Moble Commucatos Systems, U Dresde, Germay.. Ida Isttute of echology,

More information

QR-based Channel Estimation for Orthogonal Frequency Division Multiplexing Systems

QR-based Channel Estimation for Orthogonal Frequency Division Multiplexing Systems ELOMIA Idoesa Joural o Electrcal Eeer Vol. o. 8 Auust 4 pp. 67 63 OI:.59/teloma.v8.568 67 QR-based Cael Estmato or Ortooal Frequec vso Multple Sstems Pelo Ja ou e* B Le epartmet o Electroc Scece ad ecolo

More information

Improved Training of Predictive ANN with Gradient Techniques

Improved Training of Predictive ANN with Gradient Techniques Improved Tra of Predctve ANN wth Gradet Techques Dharmder Kumar, Member, IAENG, Saeeta Gupta, Member, IAENG ad Parvee Sehal, Member, IAENG Abstract Ths paper compares the mmzato of error tra of predcto

More information

The Balanced Cross-Layer Design Routing Algorithm in Wireless Sensor Networks Using Fuzzy Logic

The Balanced Cross-Layer Design Routing Algorithm in Wireless Sensor Networks Using Fuzzy Logic Sesors 205, 5, 954-9559; do:0.3390/s508954 Artcle OPEN ACCESS sesors ISSN 424-8220 www.mdp.com/joural/sesors The Balaced Cross-Layer Desg Routg Algorthm Wreless Sesor Networks Usg Fuzzy Logc Ng L *, José-Ferá

More information

Multi-target and Fuzzy Cloud computing Resource Scheduling

Multi-target and Fuzzy Cloud computing Resource Scheduling Advaced Scece ad Techology Letters Vol.111 (NGCIT 15), pp.94-98 http://dx.do.org/1.147/astl.15.111.19 Mult-target ad Fuzzy Cloud coputg Resource Schedulg Yu Zhag School of forato scece ad techology, Zheg

More information

On the Robustness of Next Generation GNSS Phase-only Real-Time Kinematic Positioning

On the Robustness of Next Generation GNSS Phase-only Real-Time Kinematic Positioning O the Robustess of Next Geerato GNSS Phase-oly Real-Tme Kematc Postog Leard HUISMAN 1, Peter J.G. TEUNISSEN 1,2, Des ODIJK 1 1 Curt Uversty of Techology GNSS Research Lab Ket Street, Betley WA 6845, Perth,

More information

Fault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks

Fault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks 1 Fault-Tolerat Small Cells Locatos Plag G/G Heterogeeous Wreless Networks Tamer Omar 1, Zakha Abchar, Ahmed E. Kamal 3, J. Morrs Chag ad Mohammad Aluem 1 Departmet of Techology Systems, East Carola Uversty,

More information

Assignment#4 Due: 5pm on the date stated in the course outline. Hand in to the assignment box on the 3 rd floor of CAB.

Assignment#4 Due: 5pm on the date stated in the course outline. Hand in to the assignment box on the 3 rd floor of CAB. MATH Assgmet#4 Due: 5pm o the date stated the course outle. Had to the assgmet box o the 3 rd floor of CAB.. Let deote the umber of teror regos of a covex polygo wth sdes, dvded by all ts dagoals, f o

More information

Joint Centralized Power Control and Cell Sectoring for Interference Management in CDMA Cellular Systems in a 2D Urban Environment

Joint Centralized Power Control and Cell Sectoring for Interference Management in CDMA Cellular Systems in a 2D Urban Environment Wreless Sesor Network, 010,, 599-605 do:10.436/ws.010.8071 Publshed Ole August 010 (http://www.scrp.org/joural/ws) Jot Cetralzed Power Cotrol ad Cell Sectorg for Iterferece Maagemet CDMA Cellular Systems

More information

Study of algorithms to optimize frequency assignment for autonomous IEEE access points

Study of algorithms to optimize frequency assignment for autonomous IEEE access points IJCSI Iteratoal Joural of Computer Scece Issues, Vol. 7, Issue 5, September 1 ISSN (Ole): 1694-814 64 Study of algorthms to optmze frequecy assgmet for autoomous IEEE 8.11 access pots Mchael Fsterbusch

More information

New Solution for Reducing Doppler Shift in the MB-OFDM System Over High Mobility Multipath Channel for V&I Communication

New Solution for Reducing Doppler Shift in the MB-OFDM System Over High Mobility Multipath Channel for V&I Communication Amerca Joural of Appled Sceces Orgal Research Paper ew Soluto for Reducg Doppler Shft the MB-OFDM System Over Hgh Moblty Multpath Chael for V&I Commucato 1 Saaa Kabl, 2 Raja Elassal, 3 Fouza Elbahhar,

More information

Evolutionary Algorithm With Experimental Design Technique

Evolutionary Algorithm With Experimental Design Technique Evolutoary Algorthm Wth Expermetal Desg Techque Qgfu Zhag Departmet of Computer Scece Uversty of Essex Wvehoe Park Colchester, CO4 3SQ Uted Kgdom Abstract: - Major steps evolutoary algorthms volve samplg

More information

Voltage Contingency Ranking for IEEE 39-Bus System using Newton- Raphson Method

Voltage Contingency Ranking for IEEE 39-Bus System using Newton- Raphson Method WSEAS TRANSACTIONS o OWER SSTEMS Haer m, Asma Meddeb, Souad Chebb oltage Cotgecy Rag for IEEE 39-Bus System usg Newto- Raphso Method HAER MII, ASMA MEDDEB ad SOUAD CHEBBI Natoal Hgh School of Egeers of

More information

SAIDI MINIMIZATION OF A REMOTE DISTRIBUTION FEEDER. Kai Zou, W. W. L. Keerthipala and S. Perera

SAIDI MINIMIZATION OF A REMOTE DISTRIBUTION FEEDER. Kai Zou, W. W. L. Keerthipala and S. Perera SAIDI INIIZATIN F A RETE DISTRIBUTIN FEEDER Ka Zou, W. W.. Keerthpala ad S. Perera Uversty of Wollogog School of Electrcal ad Computer Telecommucato Egeerg Wollogog, NSW 2522, Australa Abstract Dstrbuto

More information

On LDPC Code Based Massive Random-Access Scheme for the Gaussian Multiple Access Channel

On LDPC Code Based Massive Random-Access Scheme for the Gaussian Multiple Access Channel O LDPC Code Based Massve Radom-Access Scheme for the Gaussa Multple Access Chael Luza Medova, Ato Glebov, Pavel Ryb, ad Alexey Frolov Ist. for Iformato Trasmsso Problems, Moscow, Russa, pryb@tp.ru Moscow

More information

SIMPLE RANDOM SAMPLING

SIMPLE RANDOM SAMPLING UIT IMPL RADOM AMPLIG mple Radom amplg tructure. Itroducto Obectves. Methods of electo of a ample Lottery Method Radom umber Method Computer Radom umber Geerato Method.3 Propertes of mple Radom amplg Merts

More information

Efficient Utilization of FlexRay Network Using Parameter Optimization Method

Efficient Utilization of FlexRay Network Using Parameter Optimization Method Iteratoal Joural of Egeerg ad Techology, Vol. 8, No. 6, December 2016 Effcet Utlzato of FlexRay Network Usg Parameter Optmzato Method Y. X. Wag, Y. H. Xu, ad Y. N. Xu Abstract FlexRay s a hgh rate of bus

More information

Enhancing Topology Control Algorithms in Wireless Sensor Network using Non-Isotropic Radio Models

Enhancing Topology Control Algorithms in Wireless Sensor Network using Non-Isotropic Radio Models IJCSNS Iteratoal Joural of Computer Scece ad Network Securty, VOL.6 No.8B, August 6 5 Ehacg Topology Cotrol Algorthms Wreless Sesor Network usg No-Isotropc Rado Models Ma.Vctora Que ad Wo-Joo Hwag Departmet

More information

The Institute of Chartered Accountants of Sri Lanka

The Institute of Chartered Accountants of Sri Lanka The Isttute of Chartered Accoutats of Sr Laka Executve Dploma Accoutg, Busess ad Strategy Quattatve Methods for Busess Studes Hadout 0: Presetato ad Aalyss of data Presetato of Data Arragg Data The arragemet

More information

Robust Location Tag Generation from Noisy Location Data for Security Applications

Robust Location Tag Generation from Noisy Location Data for Security Applications Robust Locato Tag Geerato from Nosy Locato Data for Securty Applcatos D Qu, Da Boeh, Sherma Lo, Per Ege, Staford Uversty BIOGRAPHY D Qu s a Ph.D. caddate Aeroautcs ad Astroautcs workg the Global Postog

More information

Aspects of Multistatic Adaptive Pulse Compression

Aspects of Multistatic Adaptive Pulse Compression Aspects of Multstatc Adaptve Pulse Compresso Shao D. Blut, Radar Dvso, Naval Research Laboratory arl Gerlach, Radar Dvso, Naval Research Laboratory ey Words: multstatc radar, MMSE, Doppler msmatch ABSRAC

More information

PROVIDING high-rate and reliable communications is an

PROVIDING high-rate and reliable communications is an 1764 IEEE TRANSACTIONS ON MOBILE COMPUTING, VOL. 5, NO. 12, DECEMBER 26 OMAR: Utlzg Multuser Dversty Wreless Ad Hoc Networks Jafeg Wag, Studet Member, IEEE, Hogqag Zha, Studet Member, IEEE, Yuguag Fag,

More information

Determination of the Multi-slot Transmission in Bluetooth Systems with the Estimation of the Channel Error Probability

Determination of the Multi-slot Transmission in Bluetooth Systems with the Estimation of the Channel Error Probability Determnaton of the Mult-slot Transmsson n Bluetooth Systems wth the Estmaton of the Channel Error Probablty K Won Sung and Chae Y. Lee Det. of Industral Engneerng, KAIST, 373-1 Kuseong Dong, Taejon, Korea

More information

Block-based Feature-level Multi-focus Image Fusion

Block-based Feature-level Multi-focus Image Fusion Block-based Feature-level Mult-focus Image Fuso Abdul Bast Sddqu, M. Arfa Jaffar Natoal Uversty of Computer ad Emergg Sceces Islamabad, Paksta {bast.sddqu,arfa.affar}@u.edu.pk Ayyaz Hussa, Awar M. Mrza

More information

Development and Studies of Instant Messaging System Based on Android Platform

Development and Studies of Instant Messaging System Based on Android Platform Revsta de la Facultad de Igeería U.C.V., Vol. 3, N 4, pp. 48-54, 07 Abstract Developmet ad Studes of Istat Messagg System Based o Adrod Platform Zhpg Zhou, Hogya Zhag Software Departmet, Hebe Software

More information

ISSN (Print), ISSN (Online) Volume 5, Issue 1, January (2014), IAEME AND TECHNOLOGY (IJARET)

ISSN (Print), ISSN (Online) Volume 5, Issue 1, January (2014), IAEME AND TECHNOLOGY (IJARET) Iteratoal INTERNATIONAL Joural JOURNAL of Advaced OF Research ADANED Egeerg RESEARH ad Techology IN ENGINEERING (IJARET), ISSN 976 648(Prt), ISSN 976 6499(Ole) olume 5, Issue, Jauary (4), IAEME AND TEHNOLOGY

More information

Fault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks

Fault-Tolerant Small Cells Locations Planning in 4G/5G Heterogeneous Wireless Networks Ths artcle has bee accepted for publcato a future ssue of ths joural, but has ot bee fully edted. Cotet may chage pror to fal publcato. Ctato formato: DOI 10.1109/TVT.016.613, IEEE Trasactos o Vehcular

More information

Long Number Bit-Serial Squarers

Long Number Bit-Serial Squarers Log Number Bt-Seral Squarers E. Chaotaks, P. Kalvas ad K. Z. Pekmestz are th the Natoal Techcal Uversty of Athes, 7 73 Zographou, Athes, Greece. E-mal: lchaot, paraskevas, pekmes@mcrolab.tua.gr Abstract

More information

Simulation of rainfall-runoff process by artificial neural networks and HEC-HMS model (case study Zard river basin)

Simulation of rainfall-runoff process by artificial neural networks and HEC-HMS model (case study Zard river basin) Proceedgs of The Fourth Iteratoal Ira & Russa Coferece 43 Smulato of rafall-ruoff process by artfcal eural etworks ad HEC-HMS model (case study Zard rver bas Mehrdad Akbarpour MSc. Graguate, Water Structures

More information

Improved NSGA-II Based on a Novel Ranking Scheme

Improved NSGA-II Based on a Novel Ranking Scheme HTTP://ITE.GOOGLE.COM/ITE/JOURNALOFCOMPUTING/ Imroved NGA-II Based o a Novel Rakg cheme Ro G. L. D ouza, K. Chadra ekara, ad A. Kadasamy 9 Abstract No-domated ortg Geetc Algorthm (NGA) has establshed tself

More information

Infinite Series Forms of Double Integrals

Infinite Series Forms of Double Integrals Iteratoal Joural of Data Evelopmet Aalyss ad *Operatos Research*, 4, Vol., No., 6- Avalable ole at http://pubs.scepub.com/jdeaor/// Scece ad Educato Publshg DOI:.69/jdeaor--- Ifte Seres Forms of Double

More information

Optimal Packetization Interval for VoIP Applications Over IEEE Networks

Optimal Packetization Interval for VoIP Applications Over IEEE Networks 145 Afrca Joural of Iformato ad Commucato Techology, Vol. 2, No. 4, December 2006 Optmal Packetzato Iterval for VoIP Applcatos Over IEEE 802.16 Networks Sheha Perera Departmet of Electrcal & Computer Egeerg

More information

A New Method for Detection and Evaluation of Winding Mechanical Faults in Transformer through Transfer Function Measurements

A New Method for Detection and Evaluation of Winding Mechanical Faults in Transformer through Transfer Function Measurements [Dowloaded from www.aece.ro o Wedesday, Jue 0, 20 at 6:04:09 (UTC) by 27.28.226.42. Redstrbuto subject to AECE lcese or copyrght. Ole dstrbuto s expressly prohbted.] Advaces Electrcal ad Computer Egeerg

More information

THE FOURIER SERIES USED IN ANALYSE OF THE CAM MECHANISMS FOR THE SHOEMAKING MACHINES (PART I)

THE FOURIER SERIES USED IN ANALYSE OF THE CAM MECHANISMS FOR THE SHOEMAKING MACHINES (PART I) ANNALS OF HE UNIVERSIY OF ORADEA FASCICLE OF EXILES, LEAHERWORK HE FOURIER SERIES USED IN ANALYSE OF HE CAM MECHANISMS FOR HE SHOEMAKING MACHINES (PAR I) IOVAN-DRAGOMIR Ala, DRIȘCU Maraa, Gheorghe Asach

More information

A QoS-based Scheme for Planning and Dimensioning of Optical Label Switched Networks

A QoS-based Scheme for Planning and Dimensioning of Optical Label Switched Networks Ths fu text aer was eer revewed at the drecto of IEEE Commucatos Socety subject matter exerts for ubcato the ICC 27 roceedgs. A QoS-based Scheme for Pag ad Dmesog of Otca Labe Swtched Networks Yasse Khf,

More information

Increasing Geometric Resolution of Data Supplied by Quickbird Multispectral Sensors

Increasing Geometric Resolution of Data Supplied by Quickbird Multispectral Sensors Sesors & Trasducers 013 by IFS htt://www.sesorsortal.com Icreasg Geometrc Resoluto of Data Suled by Quckbrd Multsectral Sesors Claudo PRENTE ad Raffaele SNTMRI Deartmet of Sceces ad Techologes, Uversty

More information

Time-Frequency Entropy Analysis of Arc Signal in Non-Stationary Submerged Arc Welding

Time-Frequency Entropy Analysis of Arc Signal in Non-Stationary Submerged Arc Welding Egeerg, 211, 3, 15-19 do:1.4236/eg.211.3213 Publshed Ole February 211 (http://www.scrp.org/joural/eg) Tme-Frequecy Etropy Aalyss of Arc Sgal o-statoary Submerged Arc Weldg Abstract Kuafag He 1, Swe Xao

More information

The 3-dB Transcoding Penalty in Digital Cellular Communications

The 3-dB Transcoding Penalty in Digital Cellular Communications The -db Trascodg Pealty Dgtal Cellular Commucatos Jerry D. Gbso Electrcal & Computer Egeerg, UC Sata Barbara gbso@ece.ucsb.edu Dedcated to Toby Berger o the occaso of hs 70 th brthday Abstract- I spte

More information