Channel Estimation of Various Communication System with Different Modulation Technique

Transcription

1 Volume 4, Issue, January 4 ISS: 77 8X Internatonal Journal of Advanced Research n Computer Scence and Software Engneerng Research Paper Avalale onlne at: Channel Estmaton of Varous Communcaton System wth Dfferent Modulaton Technque Palak K. Patel *, Darshankumar C. Dalwad Assstant Professor E.T. department B.V.M. Engneerng College, V.V.agar, Gujarat, Inda Astract In ths paper, we have measured channel capacty for varous communcaton system such as SISO, SIMO, MIMO wth dfferent modulaton technque lke BPSK, QPSK, 6-QAM and 64-QAM. In the proposed algorthm the Bt Error Rate (BER) s less compared to conventonal algorthm. The proposed algorthm s sutale for fast speed traffc that s when the speed of the mole s very hgh. We have measured the channel varaton and ased on that t error rate s calculated. Here, n ths paper we have appled the Maxmum Rato Comnng technque. Keywords Bt Error Rate(BER), Maxmum Rato Comnng(MRC), Multple Input Multple Output(MIMO), Multple Input Sngle Output(MISO), Sngle Input Multple Output(SIMO), Sngle Input Sngle Output(SISO), Sgnal to ose Rato(SR), Bnary phase shft key( BPSK), Quadrature phase shft key(qpsk), 6-ary Quadrature Ampltude Modulaton( 6-QAM), 64-ary Quadrature Ampltude Modulaton( 64-QAM), I. ITRODUCTIO Communcatng data from one locaton to another requres some form of pathway or medum. Ths medum s called communcaton channel. Channel use two types of meda: Cale (Twsted par wre, cale, fer optc) and Broadcast(mcrowave, Satellte, rado, Infrared). Channel Capacty s the tghtest upper ound on the rate of nformaton that can e relaly transmtted over a communcaton channel. Let X and Y e the random varales representng the nput and output of the channel respectvely. Let Y/X(y/x) e the condtonal dstruton functon of Y gven X, whch s an nherent fxed property of the communcaton channel. Then the choce of the margnal dstruton P X (x) completely determne the jont dstruton [] P X,Y (x,y)=p Y/X (y/x)p X (x) () Whch n turn, nduces a mutual nformaton I(X;Y). The channel capacty s defned as C Sup I (X ; Y ) P x ( x ) () Where the Supremum s taken over all possle choces of P X (x) [] In wreless communcaton, fadng s devaton of the attenuaton affectng a sgnal over certan propagaton meda. The fadng may vary wth tme, geographcal poston or rado frequency. and s often modeled as a random process. In wreless system, fadng may ether e due to multpath propagaton, referred to as multpath nduced fadng, or due to shadowng from ostacles affectng the wave propagaton, sometmes referred to as shadow fadng. The presence of reflectors surroundng a transmtter and recever create multple paths that a transmtted sgnal can traverse. As a result, the recever sees the superposton of multple copes of the transmtted sgnal, each traversng dfferent paths. Each sgnal copy wll experence dfferences n attenuaton, delay and phase shft whle travellng from the source to the recever. Ths can result n ether constructve or destructve nterference, amplfyng or attenuatng a sgnal power seen at the recever. Strong destructve nterference s frequently referred to as a deep fade and may result n temporary falure of communcaton due to a severe drop n the channel sgnal to nose rato. So, Dversty s a way to protect aganst deep fades. Dversty comats fadng y provdng the recever wth multple uncorrelated replcas of the same nformaton earng sgnal. In ths paper, we have used space dversty. Technques appled to comne the multple receved sgnals of a dversty recepton devce n to a sngle mproved sgnal s Selecton comnng(sc), Feedack or Scannng comnng(fc or SC), Maxmum rato comnng(mrc), Equal gan comnng(egc), Zero forcng(zf), Mnmum mean square error(mmse). Here MRC s used as a comnng method to mprove performance n a nose lmted communcaton system where AWG and fadng are ndependent among the dversty ranches. 4, IJARCSSE All Rghts Reserved Page 6

2 In MRC, all paths are cophased and summed wth optmal weghtng to maxmze comner output SR. A means of comnng the sgnals from all recever ranches so that sgnals wth a hgher receved power have a larger nfluence on the fnal output. In MRC, comnng technque needs summng crcuts, weghtng and co-phasng. The sgnal from dfferent dversty ranches are co-phased and weghtng efore summng or comnng. The weghts have to e chosen as proportonal to the respectve sgnals level for maxmzng the comned carrer to nose rato(cr). The appled weghtng to the dversty ranches has to e adjusted accordng to the SR. For maxmzng thesr and mnmzng the proalty of error at the output comner, sgnal of d th dversty ranch s weghted efore makng sum wth others y a factor C * d /σ nd. Where σ nd s nose varaton of dversty ranch d th * and C d complex conjugate of channel gan. As a result the phase shfts are compensated n the dversty channels and the sgnals comng from strong dversty ranches whch have low level nose are weghted more comparng to the sgnals from the weak ranches wth hgh level of nose. The term σ nd n weghtng can e neglected condtonng that σ nd has equal value for all d. then realzaton of the comner needs the estmaton of gan n complex channel and t does not need any estmaton of the power of nose. II. TYPES OF COMMUICATIO TECHIQUE The smplest form of rado lnk can e defned n MIMO terms as SISO - Sngle Input Sngle Output. Ths s effectvely a standard rado channel - ths transmtter operates wth one antenna as does the recever. There s no dversty and no addtonal processng requred. The advantage of a SISO system s ts smplcty. SISO requres no processng n terms of the varous forms of dversty that may e used. However the SISO channel s lmted n ts performance. Interference and fadng wll mpact the system more than a MIMO system usng some form of dversty, and the channel andwdth s lmted y Shannon's law - the throughput eng dependent upon the channel andwdth and the sgnal to nose rato [6]. The SIMO or Sngle Input Multple Output verson of MIMO occurs where the transmtter has a sngle antenna and the recever has multple antennas. Ths s also known as receve dversty. It s often used to enale a recever system that receves sgnals from a numer of ndependent sources to comat the effects of fadng. It has een used for many years wth short wave lstenng / recevng statons to comat the effects of onospherc fadng and nterference. SIMO has the advantage that t s relatvely easy to mplement although t does have some dsadvantages n that the processng s requred n the recever. The use of SIMO may e qute acceptale n many applcatons, ut where the recever s located n a mole devce such as a cell phone handset, the levels of processng may e lmted y sze, cost and attery dran [6]. MISO s also termed transmt dversty. In ths case, the same data s transmtted redundantly from the two transmtter antennas. The recever s then ale to receve the optmum sgnal whch t can then use to receve extract the requred data. The advantage of usng MISO s that the multple antennas and the redundancy codng / processng s moved from the recever to the transmtter. In nstances such as cell phone UEs, ths can e a sgnfcant advantage n terms of space for the antennas and reducng the level of processng requred n the recever for the redundancy codng. Ths has a postve mpact on sze, cost and attery lfe as the lower level of processng requres less attery consumpton [6]. MIMO (pronounced my-moh y some and me-moh y others), s the use of multple antennas at oth the transmtter and recever to mprove communcaton performance. It s one of the several forms of smart antenna technology. ote the terms nput and output refer to the rado channel carryng the sgnal, not the devces havng antennas [6]. MIMO technology has attracted attenton n wreless communcaton ecause t offers sgnfcant ncreases n data throughput and lnk range wthout addtonal andwdth or ncreased transmt power. It acheves ths goal y spreadng the same total transmt power over the antennas to acheve an array gan that mproves the spectral effcency [6]. III. DIVERSITY COMBIIG TECHIQUE The capacty of wreless system n a multpath envronment can e ncreased y dversty technques, such as selecton comnng (SC), Maxmum rato comnng (MRC), Equal gan comnng (EGC). A. Maxmum Rato Comnng MRC s the most optmal lnear comnng technque; t s seldom mplemented n a multpath fadng channel ecause the recever complexty for MRC s drectly proportonal to the numer of resolvale paths (.e. ranch sgnals) avalale at the recever.[6] B. Selecton comnng The selecton comnng technque s smlar to the swtched comnng technque except that recevers are requred to montor nstantaneous SR at all ranches. The ranch wth the hghest SR s selected as the output sgnal.[6] C. Equal Gan Comnng EGC s often used n practce as t offers performance comparale to the optmal MRC wth much greater smplcty than MRC, thus makng t hardware feasle and cost relale. EGC apples equal weght to the recever channels; hence no knowledge (measurement) of average channel sgnal to nose rato s needed. Ths dversty technque s commonly used for non-coherent modulaton, and s consdered a good canddate for wreless applcatons as well.[6] 4, IJARCSSE All Rghts Reserved Page 7

3 IV. MRC APPROACHED O PROPOSED CHAEL CAPACITY Fg. Block Dagram of MRC Technque A.AWG channel If the average receved power s pw and the nose power spectral densty s capacty s [] C aw g n p W log W ts / s p s the receved sgnal to nose rato (SR). Ths result s known as the Shannon Hartley theorem. W When the SR s large (SR >> db), the capacty C awgn W log W ts/ s p W Hz, the AWG channel (3) s logarthmc n power &approxmately lnear n andwdth. Ths s called andwdth lmted regme[]. p When the SR s small (SR << db), the capacty Cawgn log e s lnear n power ut nsenstve to andwdth. Ths s called the power lmted regme[]. B. Rcan channel The Rcan fadng channel can e descred y two parameters: k and [3]. k s the rato etween the power n the drect path and the power n the other, scattered, paths[4]. s the total power from oth paths ( v ), and acts as a scalng factor to the dstruton. The receved sgnal ampltude (not the receved sgnal power) R s then rce dstruted wth parameters k v and [5]. The resultng PDF s k ( k) Where, I f x (k ) x (k ) x k (k ) exp( k ) I ( x). s the th order modfed Bessel functon of the frst knd. C. Raylegh channel Raylegh fadng s a reasonale model when there are many ojects n the envronment that scatter the rado sgnal efore t arrves at the recever. The central lmt theorem holds that, f there s suffcently much scatter, the channel mpulse response wll e well modeled as a Gaussan process rrespectve of the dstruton of the ndvdual components. If there s no domnant component to the scatter, then such a process wll have zero mean and phase evenly dstruted etween to Π radans. The envronment of the channel response wll therefore e Raylegh dstruted[]. Callng ths random varale R, t wll have a proalty densty functon[], (4) 4, IJARCSSE All Rghts Reserved Page 8

4 r r PR (r) e, r (5) Where, E(R ) Often the gan and phase elements of a channel s dstruton are convenently represented as a complex numer. In ths case, Raylegh fadng s exhted y the assumpton that the real and magnary parts of the response are modeled y dependent and dentcally dstruted zero mean Gaussan process so that the ampltude of the response s the sum of two such processes[]. The requrement that there e many scatters present means that Raylegh fadng can e a useful model n heavly ult-up cty centers where there s no lne of sght etween transmtter and recever and many uldngs and other ojects attenuate, reflect, refract and dffract the sgnal[]. D. Maxmum rato comnng MRC s a method of dversty comnng n whch: The sgnals from each channel are added together. The gan of each channel s made proportonal to the rms sgnal level and nversely proportonal to the mean square nose level n that channel. Dfferent proportonalty constants are used for each channel. It s also known as rato-squared comnng and pre-detecton comnng. MRC s the optmum comner for ndependent AWG channels. MRC can restore a sgnal to ts orgnal shape. Here, we assume that the channel s a flat fadng Raylegh multpath channel and the modulaton s BPSK. ) We have receve antenna and one transmt antenna. ) The channel s flat fadng, t means that the multpath channel has only one tap. So, the convoluton operaton reduces to a smple multplcaton.[7] 3) The channel experenced y each receve antenna s randomly varyng n tme. For the th receve antenna, each transmtted symol gets multpled y a randomly varyng complex numer h, As the channel under consderaton s Raylegh channel, the real and magnary parts of h are Gaussan dstruted havng mean h and varance.[7] 4) The channel experence y each receve antenna s ndependent from the channel experenced y other receve antennas.[7]. On each receve antenna, the nose n has the Gaussan proalty densty functon wth h p(n) e (n ) wth and. The nose on each receve antenna s ndependent from the nose on the other receve antenna.[7] 5) At each receve antenna, the channel h known at the recever.[7] h 6) In the presence of channel h, the nstantaneous t energy to nose rato at th E receve antenna s convenence, let us defne, h E (6) for rotatonal (7) Here, on the recever antenna, the receved sgnal s, y h x n y = receved symol on the th receve antenna h = channel on the th receve antenna x = transmtted symol n = nose on th receve antenna Expressng t n matrx form, the receved sgnal s, y hx n where, y [y y...y ] T s the receved symol from all the receved antenna. h [h h...h ] T s the channel on all the receve antenna n [nn...n ] T s the nose on all the receve antenna x s transmtted symol The equalzed symol s, (8) 4, IJARCSSE All Rghts Reserved Page 9

5 It s ntutve to note that the term, h H H H H H h y h hx h n h n x x H H H H h h h h h h h h h.e. sum of the channel powers across all the receve antennas.[7] E. Effectve E / wth MRC In the presence of channel h, the nstantaneous t energy to nose rato at th receve antenna s (9). h E () Gven that we are equalzng the channel wth h H, wth the receve antenna case, the effectve t energy to p E (9) 4, IJARCSSE All Rghts Reserved Page h E () Effectve t energy to nose rato n a receve antenna case s tmes the t energy to nose rato for sngle antenna case[8]. F. Bt Error Rate wth MRC We know that, f h s a Raylegh dstruted random varale then h s a ch-squared random varale wth two degrees of freedom. The pdf of s p( ) E Snce the effectve t energy to nose rato s the sum of such random varales, the PDF of s a ch-squared random varale wth degrees of freedom. The pdf of s p E e E E e BER computaton n AWG, wth t energy to nose rato of E E P erfc,, the t error rate for BPSK n AWG derved as, () (3) (4) (5) Gven that the effectve t energy to nose rato wth MRC s, the total t error rate s the ntegral of the condtonal BER ntegrated over all possle values of [8]. Ths equaton reduces to Pe erfc p d E Pe erfc e d E e k o p p k p k (6) (7) (8)

6 V. SIMULATIO RESULTS A. BER V/S SR for SISO and SIMO system wth BPSK Modulaton Technque BER perfoemancde of MRC Scheme - SISO MRC (Tx:,Rx:) MRC (Tx:,Rx:4) - BER SR[dB] Fg. BER V/S SR for SISO and SIMO wth BPSK Modulaton Technque Fgure shows the t error rate performance of SISO and SIMO system wth BPSK modulaton technque. From fgure, t s seen that at SR db, the value of BER for SISO system s around - and for SIMO (one transmtter antenna and two recever antenna) system BER s around -3 and -5 for SIMO wth one transmtter antenna and four recever antenna. So, from fgure, we say that as the numer of antenna s ncreased the BER s reduced and performance of the overall system s etter. B. BER V/S SR for SISO and SIMO system wth QPSK Modulaton Technque BER perfoemancde of MRC Scheme - SISO MRC (Tx:,Rx:) MRC (Tx:,Rx:4) - BER SR[dB] Fg. 3 BER V/S SR for SISO and SIMO wth QPSK Modulaton Technque Fgure 3 shows the t error rate performance of SISO and SIMO system wth QPSK modulaton technque. From fgure 3, t s seen that at SR db, the value of BER for SISO system s around - and for SIMO (one transmtter antenna and two recever antenna) system BER s around -3 and -5 for SIMO wth one transmtter antenna and four recever antenna. So, from fgure 3, we say that as the numer of antenna s ncreased the BER s reduced and performance of the overall system s etter. C. BERV/S SR for SISO and SIMO system wth 6-QAM Modulaton Technque 4, IJARCSSE All Rghts Reserved Page

7 BER perfoemancde of MRC Scheme - SISO MRC (Tx:,Rx:) MRC (Tx:,Rx:4) - BER SR[dB] Fg. 4 BER V/S SR for SISO and SIMO wth 6-QAM Modulaton Technque Fgure 4 shows the t error rate performance of SISO and SIMO system wth 6-QAM modulaton technque. From fgure 4, t s seen that at SR db, the value of BER for SISO system s around - and for SIMO (one transmtter antenna and two recever antenna) system BER s around - and -3 for SIMO wth one transmtter antenna and four recever antenna. So, from fgure 4, we say that as the numer of antenna s ncreased the BER s reduced and performance of the overall system s etter. Compared to fgure (BPSK) and fgure 3 (QPSK) t error rate n fgure 4 (for 6-QAM) s hgh ecause as M-ary numer ncreased t error rate s also ncreased. D. BER V/S SR for SISO and SIMO system wth 64-QAM Modulaton Technque BER perfoemancde of MRC Scheme - SISO MRC (Tx:,Rx:) MRC (Tx:,Rx:4) - BER SR[dB] Fg. 5 BER V/S SR for SISO and SIMO wth 64-QAM Modulaton Technque 4, IJARCSSE All Rghts Reserved Page

8 Fgure 5 shows the t error rate performance of SISO and SIMO system wth 64-QAM modulaton technque. From fgure 5, t s seen that at SR db, the value of BER for SISO system s around - and for SIMO (one transmtter antenna and two recever antenna) system BER s around - and -3 for SIMO wth one transmtter antenna and four recever antenna. So, from fgure 4, we say that as the numer of antenna s ncreased the BER s reduced and performance of the overall system s etter. Compared to fgure (BPSK) and fgure 3 (QPSK) t error rate n fgure 5 (for 64-QAM) s hgh ecause as M-ary numer ncreased t error rate s also ncreased. E. BER V/S SR for MIMO system wth 8-transmtter and -recever antenna MIMO: TX=8, RX=, 6-QAM - SER E/ Fg. 6 SER (Symol Error Rate) V/S SR for MIMO wth 8-transmtter and -recever antenna Fg. 6 SER (Symol Error Rate) V/S SR for MIMO wth 8-transmtter and -recever antenna From fgure 6, t s seen that as the SR ncreased SER decreased. In fgure 6, we have take 8-transmtter and -recever antenna. From fgure 6, we say that at SR db the value of SER s around - and BER s very less. F. Channel Capacty V/S SR for MRC Technque Capacty of suoptmally selected antennas 9 8 ps/hz SR[dB] Fg. 7 Channel Capacty V/S SR for MRC Technque From fgure 7, t s seen that as SR ncreased the channel capacty s also ncreased. As we ncreased the numer of antenna the channel capacty s also ncreased. 4, IJARCSSE All Rghts Reserved Page 3

9 VI. COCLUSIO In ths paper, we have measured channel capacty for varous communcaton system such as SISO, SIMO, MIMO wth dfferent modulaton technque lke BPSK, QPSK, 6-QAM and 64-QAM. In the proposed algorthm the Bt Error Rate (BER) s less compared to conventonal algorthm. As we ncreased the numer of antenna as oth transmtter and recever sde the BER s decreased. It s also conclude that as we ncreased the numer of antenna the channel capacty s also ncreased. The proposed algorthm s sutale for fast movng vehcle as well as for more fadng envronment. ACKOWLEDGMET The work of Palak K. Patel has een supported y Darshankumar C. Dalwad, M.E. Gold Medalst from S. P. Unversty. Rght from the tme I started workng on my paper, he has een very helpful n organzng my deas and puttng together a very fne pece of work. He s also provdng the help for how to wrte the MATLAB code. Also, I am grateful for hs patence n readng, and correctng my errors n wrtng. REFERECES [] M. R. McKay, A. J. Grant, I. B. Collngs, Performance Analyss of MIMO-MRC n Doule-Correlated Raylegh Envronments, ovemer 5 [] S. Catreux, L. J. Greensten, V. Erceg, Some results and nsghts on the performance gans of MIMO System, 3 [3] S. Wu Km, Z. Wang, Maxmum Rato Dversty Comnng Recever Usng Sngle Rado Frequency Chan and Sngle Matched Flter, 7 [4] P.K.Patel, D.C.Dalwad, Comparatve analyss of maxmum rato comnng technque on SISO and SIMO, n natonal conference PEPSI-, SVIT, Vasad, Gujarat, 9 th January, pp Chandrasekaran, Rutgers unversty, Dversty Technque n wreless Communcaton, Sprng 5 [6] ME Thess, Palak Patel, Gujarat technologcal unversty, Performance analyss of dfferent dversty comnng technques wth MIMO channel n wreless, June. R. Adve, otes of recever dversty. [8 ]R. John, [DIG-COMM-BARRY-LEE-MESSERSCHMITT], Dgtal communcaton, thrd edton. [9] SKLAR, Dgtal communcaton [ ]The handook of electrcal engneerng research & educaton Assocaton,996, P.D-49.ISB [ ]Davd Tse, PramodVshwanath(5), Fundamental of wreless communcaton, Camrdge unversty press, U.K. [ ]John G. proaks(995). Dgtal communcaton(3 rd ed). Sngapore:MCGraw-Hll Book Co.pp [3] Ad, A. and tepedelenloglu, C. and kaveh, M. and Gannaks, G. on the estmaton of the k parameter for the rce fadng dstruton, IEEE communcatons letters, March, p [4] Statstcal propertes of a sne wave plus random nose So Rce-Bell syst. Tech J, 948. [5] Rchards, M.A, Rce dstruton for RCS, Georga nsttute of Technology (sep 6). 4, IJARCSSE All Rghts Reserved Page 4