Scribd
Upload
9 views22 pages

Module

The document discusses the transition from analog to digital communications, highlighting advancements in digital technology and the reliability of digital communication systems. It presents a conceptual model of digital communication, focusing on the characteristics of Additive White Gaussian Noise (AWGN) channels and the importance of minimizing symbol error rates in receivers. Additionally, it covers signal representation, orthonormal basis functions, and the Gram-Schmidt orthogonalization procedure for constructing orthogonal signal sets.

Uploaded by

Harsha Vardhan NH
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
9 views22 pages

Module

The document discusses the transition from analog to digital communications, highlighting advancements in digital technology and the reliability of digital communication systems. It presents a conceptual model of digital communication, focusing on the characteristics of Additive White Gaussian Noise (AWGN) channels and the importance of minimizing symbol error rates in receivers. Additionally, it covers signal representation, orthonormal basis functions, and the Gram-Schmidt orthogonalization procedure for constructing orthogonal signal sets.

Uploaded by

Harsha Vardhan NH
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
oem Shankara.M,N. 2 EGE Dept, SCE Bangalore Associate Professor ravishankarmn@[Link] Module 2.- Signalling Over AWGN Channels Introduction The conversion of analog waveforms into coded pulses represents the transition from analog communications to digital communications, This transition has empowered by several factors like: 1, Advancement and improvement of digital silicon chips, digital signal processing, and computers, 2, Improved reliability in digital communications 3. Broadened range of multiplexing of users, which is enabled by the use of digital modulation] techniques. j4. Different communic ion networks, uses digital communications as the preferred choice. Conceptual Model of Digital Communication System Figure shows the conceptual model of digital communication system. Théimessage source emits the sec with the symbols = 0 and m; = L| belonging to an alphabet of M symbols denoted as my oe If M=4 > m, = 00, my = 01, m=10 and my =11. ae equally likely, Hence p, = P(emitted) = = foralli rq a xX tm) £8), sof 0), ‘vector | Estimate Message Vector é faveform ctor zee et [OL tine] meninee SOY gree [AY ewer FL ee Fa The output of the message sources numbers as output and is of the a a Sig ier which preduces a vector of real where NSM |The VE real ed aay nite energy signal s(t) for every symbol emitted by |the message sol gai e fle energy of the signal ig given by E, = SF sq? at ‘The Channel is as: yg nee two characteristics: Channel is linear, wit ih bandwidth that is large enough to accommgdate the transmission of the modulator output s(t) wit dition, Transmitted — Recnived signal signal >The transmitted signal 56)’ is perturbed by an additive, zero-mean, 3 + x). I stationary, white, Gaussian noise process denoted by w(t) and such a | channel is referred as AWGN ( additive white Gaussian noise ) channel. The received signal x(t) = s(t) + w(t) Whita Gaussian noise The Receiver observes the received signal x(t) for a duration T and makes a wid best estimate of the transmitted signal lowing to the presence of channel noise, the receiver will inevitably make occasional ‘trors. ‘The requirement, therefore, is to design the receiver so as to minimize the average probability of} symbol error and is defined + Page 1| This supporting material forthe SCE students, Ary use for eonmereal purpose hs to be permite by the author ORevi Sankara MLN. A iw is ae fc This isa Ravi Shan (fi Ol1 sent) and P(A = 1[0 sent Géomet ‘Any set of M energy signals, (s(t)}, ean be expressed as linear com functions, where N Inner product 1). interval 0 so sas u(t). (Analyzer forseconstructing the signal vector (Si) lem, | ae | 3 “ (t) can be completely described by the tor s;is called Signal vector. The set {siicizum Ded, vector s; with its coefficients as ‘Sin Sa reptééents the co-ordinates of M message points in sgh dimensional Euclidean space, with N mutually . perpendicular axis labeled as ©), 2 ... Ox cg withM message points is called Signal space diagram or Signal 2; results in three signals s\(t), 52(0) and ,(t) and two orthonormal basis functions (1) and ®2(0) pmstaial for teSCE students, ny use for commerca purpose has io be permite SCE Dept, SCE Ban Bala) ravisha nkarmn @sapthagin. ed Ravi Shankara. Associate Professor _* + product 3 ; Length of Signal Vector and Inner product “ on enh of any signal ECTS is define es ‘The length of the signal wetor is defined by ee Wait = sit]= (68) to be the inner product or dot product ofS} with its i Sia Sit sun 7] 2 Isyy 2+ Sig 245137 SUN Us. Sie oS Sin Sin ilsIF =)" : Gis) _ sits [Cosine of angle between s, ands; Cos(0,) = Teas] = Tete 5 *** If the inner product is zero, the signals are said to, Relation between Energy content of a The energy of the signal s(t) is given by B, g, =| He [dsorco]a s))® where s, are elements of s, Hence the rey Of be Siena (OSGiaLto'te squared length of the signal vector representing it yee ae ispceihSvgnok i ayonges Oo) Gram-Schmidt Orthogonalization (GSO) To show that it is possible to construct a set of ‘N’ orthonormal basis functions 4(t), (1), .. from the linearly independent signals (s(0).i= 1,2, uN IPs) (8),$o(1), «1. Su(l) ate real valued energy signals, each of duration Tiss supporting material for the SCE students. us or comm avi Shankara. ECE Dept, SCE Bangalore ‘ssoclate Professor ; ravishankarmn@sapthagirieduin Ssociate Professor sane Ganga a0 320 - 521010 & onc = & 2101 aL | S22 % | Let the intermediate function gp(t) = sp(t) — S210; (#) such that g2(') is orthogonal to ,(t) | Q. Explain the Gram-Schmidt Onhiogonaliation procedure. 2 i 7 oe 82) _ TG) 8B .0 “29.0 on = ! 2 ora re® VEz = S21? i f i 1 i BAO Bea 23. WN | © energy oF gi | | aSrormal set from which one ean generate orthogonal signals as & i is functions. | 3, oy and (S31 Sx)* gyaye 401 (S20 5 etsT so-raoe [Stal at 1) 22 o,() * Jot) = SH esas 4010 ae f 2: 3 Oi() é and 83(0 = $3010) +Sy202(0 = 40014) finear combinalio” of the basis pra the constellation dig and express s1( $20) and 530.95 TEs) = 033) (oy,s1) = G95 92 ~ (onus) = C295 & st” functions (0 and @,(t); given st = ‘Solution: 11,812) = 3,0) 82 = Given si o 540 [oetoat s(t) = S10 * sno= 30 and 3 a(t) = S11 + 83202(0) = 3 &i(t)-3 O28) 3. Cheek whether the signals Ob. and os are orthdze nal? “Ol (0), of iC Functions. Express the signals 040 1 40) 1 Solution: Consider the integral | > Hence (1) and ©3(0) are ontbogonal Wenchotten. [o,(O'at= a= Jat =2) and; [ws covers [ o,(oJtae= (1)? + (-1)2 = 2) heat ing orthonormal functions are Page 6 ff ankaral ECE Dept, S Associate Professor 10 = 0 +0,(0; S29 = —0,() + 6,0) and « 830 = 0,(0) - 0 (0) '[Link] the orthogonal signal pair shown, construct the signal constellat s2ft) “ye tion for s(t) and s(t) Solution: [Energy of signal s,(t)is givenas B= J [s,(tl*d ‘The first basis function is given by 9, (0 = Let go(t) be the intermediate function such that so t rl g0(t) = s2(0) - Sarr) . six = i S200, (a=, Hence g(t) =s2(t) ‘The coefficientsare _ fal E Poe and Sy = Ey = VT The seat vecorstH ag =| Tee Sy fl- [val 5. Using Gram-Schmidt ais procedure, express si(t) and s2(t) interms of orthogonal basis Tunctions. Solution: Given number of ifdlependent signals M = 2 and the basis functions required to represent these signals N=2 Energy ofsignal s,(Wis givenas E, = ff's,(? dt= Gada er Ko ‘The first basis function is given by 0,(t) = si(0 2 AL i ‘ ve VRPT (0 = ostsT Ths leasupperting material for the SCE students, Ary us for camera pose hs abe permite by the author CR Shankara CE Bangalore ravishankarmn@[Link] Page 7 Ec el naa es on such that s re orthogonal to each other Leis te fmnetion such that so that ga(t) and y(t) a S21) = = Sud) aa Foafd = seoo,coar f aa(z)ace lo 1 2D = 820) = 5,0,(t) = [2a=0Cr - (o-at Ge + . Energy of g, (tis given as 8, [* ca aus ff (2A)? dt= Aer fi The s ee he second basis function is given by (1) = = Ps, 6).Consider the signal s(t), sO /Foaud sa(t) as AS in gure Ue Gram-Schmit Orthogonalization | procedure to obtain the bi elon Citrate responds } ata oor add a) sa) 1 1 af 1 | oT i Omg Or TS 0 Tee 3 3 3 ony” sigilal space diagram. Solution: By inspection; s3(t) > se(t) & not linearly independent, Hence the number of linearly indepertient signals N= Since N 3 and Number of distinct symbo's M= 4 |. Only thee basis fmetions are required to represent the g given signal, Energy of the signal sy(t) is : [tscova= [ova Intermediate function Balt) = 5,(0 - 529,10 The first basis function (2) is 0400 = sO 1 3 er ECE Dept, SCE Bangalore jassociote Professor oats fes501 visha nkarmn@sapthagiriedy 7 aan = Sz a noes fol a Fenos lb For the period F B20 = 5,0) -5,,0,() = 1 - fF fi For the period galt) = s(t) ~ 5,,0,(0 = - ft. ()=1 art) = 5,(t) ~5,0,(0 =0 ~[E Energy of ga(t) is computed for'the period For the period Fst only as ext) [Link] other intervals 3 a 3 a zx Energy of g,(t) is given as a= [e2(oytat = [> xypear= b 0.0) = 4 fo-- ‘Second Intermediate function a1 = £2 s5(00, (Dat = fS(A)(ODde= 0 ae For the rg: ex(t) fe ae * Terra Ta for 2 sest (0) and (9 forms an orthonormal set, The signals s1(1), s(t), ss(t) and s(t) ‘do not form a linearly indeperfent Set since sa(t) = s1(t) + s3(t). ale) ale) a ol : ! oO g Tr " ore BCE Dept SCE Banat. ravishankar Ravi Shankara. Associate Professor. ‘ombination of the three basis functions as given below. Any ofthese four signals ate expressed as a linear ¢ sa(t) = 531 Pr(t) + SsxP a(t) + 833 3(t) Oi), sal) sar A) +822 PaO) 10) + S3(0) = 811 D(H) + 831 DC) + S202) + S53 s(t) Noe:s=VE sa = JB andsy = VE 7. Using Gram: Schmidt orthogonalization procedure, find the set of orthogonal basis functions to represent three signals as shown in figure. Also express each of the signals in terms of basis functions. ae af?) ssf T Rete bbbbounus T Soh Tee ade si(t) is BS if [s,(OPat lo First Intermediate function = sribal sO fn oO -(ists2} Second Intermediate function g,(t) = s3(t) ~s510,(t) — 5320 (0) : 1 fa -[ (09, (0ar= [ @)@at=3 and sn [ ss(00,@oae = f @eaae lo F Forthe period OSt<1; — g,(t) 3 83(t) ~5410,(t) 5320, (t) Forthe period 1552; galt) = 55(0) ~54,04(0) ~5y909(0 = 3300) = (-3).(-1) =0 8300 = s5(t) = $3:01(0 ~ 532020 30). (0) =3 ¥ commercial purposehas to be permitted by heauthor (Ravi ShanksraMN. Page 10| 3-3(1)-(-3).(0) = For the period 2/2 of additive noise process W(t). Covariance is a measure of ae in one variable are associated with changes in the second variable. ‘Covariance measures the degr fing propery ofS funtion fre which two variables are linearly associated, Since X; are mutually uncorrelated, the Covariance of X, are given by Covl x.) = [(%)— 1, Mu ~ Ha, )] = ECA) $y OM — Su] Covle Xu] = E[(S, + W, ~ Sy) Su + My -54)] = B[W, Wy] ft 1 eel we, (oat f Ww, oa] f Show that the elements of Vestor N: (Cocreltor outa N) are satay Independent [Associate Professor i pa cova Xe] = le [coe swcow@d)aedu cov[x, Xe] = 28 fF O40, (OSCE wat du to each other) forj#k (ince) Okare orthogonal different and it can 7 cove, Se] = 92 [9,20 OC ae = 0 «the Comelator oumputs are highly uncorrelated a1 nd each Correlator o/p are ‘This shows that be concluded that they are statistically independent. west siays questlon pa “ion Sess am pans a ranand varance of he comelator output Ako show that Se corielato' ‘outputs are statistically |} independent Explain how to conver continuous Show that for a nosy input, te mean value of aj {-1.2.... Nhave variance equsl to PSD (No/2) ofthe aditve noise process wi) AWGNchannel into veetorchannel Ate v mcltor sutput % depends only on sy and alle COmeItOr OMIPUIS 1 = PDF of X is defined as 4 i> Tana Since the elements are statistical expif independént)*the, conditional PBF of X given that the signal si(®) or f cofditional PDF of individual elements onesponding symbol was seas x and.3;ar€ saitple [Link] random variable X and-X}. The vector x is called as observation vector and x; is called an element: Using equation (C) in equation yt At Nw ecard = (Ee) ol een] on o9-# 5-0 ofan Likelihood Function The conditional PDF fx(x|m) ofeach transmitted message my is known as likelihood function denoted by lm). Hom.) fetclons) 1,2, 0M For convenience, the Likelihood function is expressed in natural log and given by sia supporting material for the SCE students. ry wsefor omer | ris is asupporuing! ‘ I purposehas w bepermitedby the author CRa ShakeraMN, Page 14 Dept, SCE Bangalore karmn@[Link] Lemde mtn} o|-2 54 la iessnge, points, for the ‘insmitted signal s,(t) constitutes a message constellation, ra The representation of received signal x(t) is cong due to te peers ofad x(t) is applied to N bank of correlators , the correlitoxbutput defines thvobservation vector x and is given as ‘Mlustrating the effect of 2 ‘noise perturbation on signal point. the location ofthe received signal point. « » B v ‘esi | is a ened in the safe Euclidean space as received signal point. Due to the See Eeeanth ‘cived Signal point, falters about the message point ina random fishion and may lic y ‘ietan ‘The signal detection probléf a as“ Given an observation vector x, perform a mapping from x to an estimate 7 ofthe transmitted yymbo! m, in a way that would the minimize probability of error in the decision] j making process “. # Maximum Likelihood detector (decoder) | (a) Given the observation vector x, for the chosen decision rf = m,, the probability of error is given by P.(mjlx) = 1 - P((m, sentlx) ‘On the basis of the above, the optimum decision rule Choose 7A = m, if stated as rms ina srrarignar ero SEs An wear cnet Preis obeprniy ator Ek SHAAN " Page 15| ECE Dept, SCE Banga Tore : ravisha nkarmn@sapthagir clin —Rule(1) referred to as Maximum A posteriori Probability (MAP) rule and the system used to implement this rule is called a MAP decoder. ‘ [(b) The MAP rule can be represented considering Likelihood probabilities using Baye's Theorem as Hy fy (xIm, set Ram, if Eafe (olny) | fe@) Where m4 is the prior probability of transmitting symbol my, ismaximum fork={ --—~~—Rule(2) ‘A(xima) is the conditional PDF of the random observation vector X-When the symbol m, is transmitted | {X(X) is the unconditional PDF of X te jc) Iecan be noted from rule-2 that (i) the denominator term f(x) is independent of transmitted symbol (43 the prior probability 74 =m; when all the source symbols greased | Lil) the conditional PDF fx(xim,) bears one-to-one relia ee | The Gecision rule can be restated in terms of Lim | Set =m, if L(m)is masdnetifr ksi Rule 3 is referred to as Maximum. Likelihood rule ae pe System used to implement this rule is called a | Maximum Likelihood Decoder. * % wt }(¢) Graphical representation of Maximum Likelihood rile | Let Z denotes the Ne -dimensiorial [Link] this space is known as | Observation space and this space is Corre: ‘are — into M “yesh regions denoted by Z), -2ss, The decision rule can be [Link] a Set M= mj; ObsBrValtion Vector x I Pin the region wae maximum for k= (e) Using t eFélationship othe Likelihood function for the symbol my Pinimum, ‘L(m) attains its maximum value. 7 Observation Vector x est régionzZ, if Yorn is minimum for k =i — —Rule(S) fi w (Also Dens) = Ie sll? Where jc sl the Euclidean distance between the observation vector x at the receiver input and the fransmitted vector s,. Hence the decision rule can be restated as Observation Vector x lies in the region Z, if the Euclidean distance||x ~ s,l| is minimum ECE Dept, SCE Bangalore ravishankarmn@[Link] The Rule (6) implies that the Maxi Maximum Likelihood rule is simply to choose the message point closest to the received signal point, N y > 65 — 52) = G5) ~ 25 2S) 1y 2 2" ~28 54+ Gu) Ye)? =2( D4 5-7 ou)" J fat fA ft ‘The first term in equation (B) gives the energy of the signal and is independ 2 ‘he second term is the inner product of observation vector x and trareriyg fl eure hence it is ignored. ynal vector Sy and the third term tion of RHS. sa is the energy of the signal s(t). Minimization of LHS is equivalent toma imi Se <3)? it 2 ‘The Maximum Likelihood (ML) decision ¢ x " ne if | Sox; see | ts maximum ant ket The figure illustrates the above statement for M= 4 sig eisigna's are [Link] Bn af 5 the wansmitted signbls,s*(), sa(0 receiver which st@fewn as a Correlation receiver. The output of the correlation vector, @ knowledge of the transmitied signal and onan (i probability of occurrence of input symbols Ves oC The correlation receiver works on Maximum Likelihood(ML) decision and consists of two subsystems namely the detector : and vector receives, fare ‘The detector part of the receiver consists of N bank correlntors, . supplied with orthonorimal basis functions generated locally +n period 0S t $ T to produce the observation] and these correlatays operate on the received signal x(t) for the u(t) are equally likely, the optimum | ‘mizes the average probability of error ‘S| receiver is an estimate of the input based orf() the observation Page 17| The vector roeiver & implemented inthe Om 1 maximum likelihood detector that, operates On bservation vector x to produge an estimate *t oF the transmitted symbol m, The N elements of the vector x are multiplied by comesporsting N elements of each of the M signal observation. [vectors Si, Sooo Sye-aind the resultant products \ are summed in the accumulators to form the inner products {(Xsq)} and these are comected by subtracting half the signal energy (E/2) and likely hood decision is finally maximum employed to get the estimate 1, The vector x lies inside the decision region ne Pastas Filter Recei is time reversed and delayed version oftth in nature and the noise in the channel The detector in correlation receiver ¥(T) = xOb (T= Hae -- Consider the detector based A bank of correlators. output of j*correlator is given by, [ xmayoa -- lo =) By comparing equations (B) & (C), for y(T) to be equal to x), it must satisfies the following, jh(T~0 = (1) for OSt5T andj=1,2,...M Equivalently b() = OT =t) andj = 1,2,..M This isa supportin x(t) , A flier whizh is matched to known signal of duation T sec, characterizedby an impulse response which input oi) sky meets wn Ne filter. The matched filter is causal Ah()== Time reversal-—+ h(-t ACOs crime iy too Meee) Detector usa Mateo et | ren (Observation vector x ost) Matched gankara.M. Professor ECE Dept, SCE Bangalore le ss ravishankarmn@[Link] Fa [poo the above described condition it can be concludes as follows -paiven a pulse O(1) occupying the interval 0 <1 < T, a LTT fier & snid to be matched to M(t) if its tmpulse response h(t) satisfies the condition nt) = (Tt) for OSt Obiain the tmaninain likelyhood decision rule forthe signa detection prot Obtain the decision rae for maxitmam likelyhood decoding and explain the cortelatinn receiver > Explain the conelation rediver using product modulator or mtegitor aud nated filer, | PWath neat diagrams and necessary equations eplin matched fier PExplain the matched filter receiver with relevant mathematical theory 4 a and the cofresponding tris AWGN. The signal vectors 0,1,0): Determine the ov signal vascd oh dimen ti ‘avi Shankara. Associate Professot receiver selects any one arbitrarily. r matched 9) For the signal shown below, determine the impulse response of 2 filte its to s(t), Plot the matched filter output as a function of ti determine its pea Solution: 4 RS d delay 28 version ‘The impulse response of the matched filter is time®reversed ani : of the amin signal, b(t) = s(T-1) and the output of the'filter, y(t) = x(t) * be. 0 The peak value of the output is (a°T/4) (refer class notes for detailed solution to the problem) | 10) Consider the signal s(Q shown in figure. @ Determine the impulse responsé ofa filter matched to this so Solution: The impuise response of te matched filter is time-reversed and delayed version ofthe input signal, h(t) = s(Te0)'and_y es the ouput of the filter, y(t v [Givens +1 ford © 5 a4 7,2--m ootane, . wis where gauncarr node OD Mtar=O anol PP=Me Te St & ap & corrnetato cOnsrbtici A Vella x xKesttm —»@ : wt on xy 4 “ aa [] so] ae Law ie ey t vee 1% & 0 are Saupe valu Gf rewcton, veetow KE Po = plarseat |anecetved) > © bm ag Prteakelety 6 Cng, cere) v pny Peto ststeure D@ Y Me etttua/e & meceke fyntod viens he plrat set | recetueh IE plany sem Lx xeretec) F bepa--m be OO : The hues rub gia oy BOQ 4 ated Uax poircrio¥ > The Leelibord 6x Grtetwt) J tes Coretetrmnal Por gf mecetved Vern x Ge Tuas ar* ato4 trans

You might also like