a Linear Brock Copr 3 Previous YEARS (ed Ans. Channel Data Rate : It is the bit rate ar, of encoder. If the bit rate at the input of enegae te will be in then channel data rate will be, ee Channel data rate The channel encoder separates " fing bit stream (the output of the source encoder) (R)) = ERs length blocks of L binary digits and maps each imessage block into an N-bit code word where N> L, he extra NL check bits provide the required error | © Swiss are content errors? = tion. There are M = 2" message and thus 2' code ‘words of Jength N bits. The chanriel decoder maps the —————__—_—_ 7 received N —bit word to the most likely code word and Ans. Content Errors : The content errors are Inversely maps the N-bit code word to the corresponding | but errors in the contents of a message ic. 0 my Lit message received as I or vice-versa. ‘Code Word : The encoded block of N bits is called a code word. It contains message bits and redundant bits. a ee aRT-B | } ———————— Q2 Define block length, (RU, 2016} 06) Consider a ae 26 Consider a (6, 3) linear block code w generator matrix is given by £002.04 01 07} peg OO Gasp Ans. Block Length : The number of bits N after codi is called the block length of the code. a 3 bmann IR.TU, 2016} (4) Find the Parity check matrixim distance, we use the property equal ty check of columns of the pari fH are non-zero, and they are conclude that d= 3 3 columns of Hz 0) fo +1} _]0 o| {1} Lo. nimum distance of code is 3. Syndrome computation circuit 6 © +E -6-EhO- Fig. on- Differentiate between systematic and m jn le of each. »yematic codes. Give example of ir 2017 Examples: CChecksums and hash fonctions, combed wih the input data, can be viewed as setemane error-detecting codes. i + Linear codes are usually implemented’ as systematic error-correcting codes (e.g. Reed- Solomon codes in CDs). + Convolutional codes ae implemented ascither systematic or non-systematic codes. Nom- systematic convolutional codes can provide better performance under maximum-iketinood (Viterbi) decoding. —_————— Q8 Given a (6,3) linear block code wil the following parity check matrix H: 1017080 Hzjo 1.4018 \prr0et ix G. (Find the generator maint 1g. (i) Find the code wort for ta ce ee ‘Ans. (i) To obtain th _ ofrBech WV Sem Oe D) and (011) is marion, j ficated by the vector diagram alao. "+8359 Minioam Distance (dau): It ist wring distance etn the Vcd possible if the rec ode vector. This short v= show im the received code vane’ te cea Table : Error controt ‘The generator matriX a] Nanwe of errors hb hse INo, detected/corrected roy [1 [Deter upto 5 errs gefor ool [2 [Correct upt Som pet word | ae die dae lsh Uy ae | due Se seni the codeword for data bit 110 : St errors pes peers M=[110] This is (6,3) code, The three check bits can be iby equation fiar ree fae 1 | 4154 For the (n,k) block code the minimum dit givenas, doin SM ~K +1 x =[180800018010180] =[110] Code vector, X=(m,,m,, m;,c;, ¢, ¢3)=(110 110) (Code word = (110 110) wennn-- 2UW The Parity check matrix of a (7, 4) Hamming code is as under frteti00 mj noon (1011001 it rors. Calculate the syndrom vector for single pee 2012] —— Syndrom Catewtation S= BHT fry liad ora Halton 10 of o10 ort Syndrom for 1" Bit Error S= EH" s=(01) ie, syrom yestor for UO areal Similary syndrom table can Oe ENEving eee h Explain wil 0 place uke p mets sel the need Ipecodink ph rest (ARQ); they must ean ARQ h atic Repeat reQU e-sent sr eorection (FEC), By the fora rr oor and retransmit ih viscovers an ror an illaive too late to be any good: ns where the transmitter immed : on as it is sent (such as most information as soon as it is sent se eras cannotuse ARQ they must tse FE ely no ‘available. (This is also why FEC is used in data ‘systems such as RAID and distributed data store). ‘Applications that use ARQ must have a return ‘chanel applications having no return channel cannot use ARQ. Applications that require extremely low error rates (suchas cigital money ransfers) must use ARQ. Reliability and inspection engineering also make use ofthe theory of erorcorecting codes. Internet ss spi TPP sack, error control is performed . faethe ame carries a CRC-32checksum. janes received with incorrect checksums are by the receiver hardware Contains a checksum, Protecting eader. Packets with ade Caen dropped Within the inorder minimizer OM the 1P¥6 header ‘routing, HZ Processi is sed because curent wine Costs in network cient tly (such as throug! cin’ hast tible-ack) or impete Deep-space telecommanications 7 deep-s i i eep-space missions (My the extreme dlition of signal power oven aso doe and the limited power availability in age Development of er upled with the history distance probes. Where 108 sent dy di starting from 1968, digital error correction segs in the form of (sub-optimally decoded) cong oh and Reed-Muller codes. The Reed-Mutles somal suited to th between 1969 and 1977, The Voyager 1 and Voyager 2 mig, started in 1977, were designed to deliver cole: atid ainongst scientific information of Jupiter and seq a resulted in increased coding requirements, ann it spacecraft were supported by (optimally Viterbedcae convolutional codes that could be Concatenated wi outer Golay (24,12,8) code. = The Voyager 2 craft additionally , supported ion of a Reed-Solomon coq for vanntted Reed-Solomon-Viterbi (RSV) code atin| * ——— party nena emt oe 100) ory Alice oon jo Wehente ool mane receives 100 cmnputes overal eoverrer geen “This mechanism enabley _Themestai beste emt cron, tant four age sie an incorrect master of a ea the two examples above, Bob's cake party : Bob's called Sat atches the parity bitin ts received ale inating S single bit errors, Consider the flowing there are sample with a transmission error nthe send its OR: dat Failed transmission scenario [Type of bit Se | ee odes & act | [peat la isa w ASE TO ae data from sectors that have “B00e errocinthe | Alice some parity bit vale Gore in the spare sectors. RAID second bit, | 1*0" 0"! ‘Aloe ads parity bit and sens 1010 MISSION arety of error correction techoiases to shes a hard drive completely fails. ch as ZFS or Burfs, 25 well as some RAID ves overall pati | | Bob comp jeanne eviexentations, support dats ‘scrubbing and resilvering, = pai blocks tobe detected an Ghopefully) | | | Bob reports incorrect transmission Iecovered before they are used. The recovered datamay | | ane eee pga Ibe re-writtea to exactly the same physical Jocation, or to PF ace wants to tres yoo t ks 2 of hardware, s ity ME =0 ae: decoding: Parity bit checking is used occasionally for hich have 7 bits, leaving01 iranemit: 1 TRANSMISSION ERI Bob receives: 11011 Bob reports correct transmission | though actually incorrect. | Bob observes even parity, as expected, thereby falling to catch the two bt errors. ‘errors and classification of Explain the type of | = /RTU. 2018, 2017, 2012, 2010 transmitted data during teporized as under ig but errors in the ontents Of & message jig f metal, s ion colli ng also with the walls of _ 1d with flow of current acto because they incre 55 Semicg, junctions. The charge carriers, electrons or i a 4 Phenom, the junction region from one side, drift or are ancl atthe junction and are collected on other side. The junction current determines the average interval that, between time when two successive carriers Junction. The exact interval that elapses ig ea fluctuations. This randomness give rise to hos we know that power spectral density of Gaussians Ea - _ receiver input is white Gaussian noise. The ‘Transm errors introduced during a particular intemat Gaussian noise does not affect the performance during subsequent signalling interval (©) Burst Errors : Which is due to impulse noi by long quite intervals followed by high amplitude nog burst. This type of noise occurs from man-made causes such as lightning and qansients. When such noise occurs, it affects Many natural, one symbol or bit and there is usually a dependence errors in successive transmitted symbols. Error contro} schemes for dealing with: random enn | are random error correcting codes and codi designed to correct burst errors are burst over codes, Shot Noise : Shot ni due to the random behaviour of, and holes), In electron tut to the random e eas ‘Oise appears in active deviss ‘ Carri Hon ant recombination ie aaron geoeratin i ‘ole pairs, i rare «Current i Fig, inthe form opae MON devices a.ancory and Coin) Wi p(mi) rete *) = ign ile gees He(m+1) pli) rar (rer Moa) rg.= message bit error probability sz iplies that majority of decoding errors are a(t 1 biterors in an m-bit codeword alt “¢ (c+ 1) error, the fraction K/n represent the ofnessage bits. Hence average message bit eror Kk pe pert +E Poe (pe a (i) For Error Detection (We compare the performance of data transmission system using block codes for error detection with systems using direct transmission without error control coding. . Wedo the comparison under the assumed same assumption that Syy» ls Tp remain same for both ‘coded and uncoded systems. Weuse probability of incorrectly decoding a block ofkmessage bits as our measure ‘of comparative performance. Here we assume that (n, k) block code is capable of detecting upto 2t errors per block. . Decoder checks the received codewords for cors and when error is deduced, the decoder may either discard or retransmit the message. We know that data rate r, over channel is a2) whenan(a, 8) eror correcting blosko8e is used. The dat rater, will have to be higher At time ty + A © Sock that woe Seam a acknowledges ~ teknowtedgement (oak, °r © teva serecarlessement WACK ty maatne ‘in received — on IAC oo ‘= ee tanemited ince eae it if not —. aa Case transmit 1 ~- can is wait seodtverhs ae ~ oy ‘To average channel all error rate blocks, Consider the transmission ofW block of data oer channel at rater, bits per second 6. Total time needed to complete transmission: know r, - bits per second rate. i Te, % Total n bits & Delay =24 Te Raza -. for N block time = sna)24 7 Burst correcting efficiency y = —* ~ f Correcting Codes ete it i Ifa code is needed for detecting a burst of en (B) Borst g techn which | then, number of check bits needed must satiggy. a n-k2d better to design ut. His nee ors and/ or single fandom erro Decoding algorithm for correcting burst similar to algorithm for cyclic codes designed iy random errors: (C) Interleaving (i) Block Interleaving 4 codes capable of cor multiple Burst of length nents are © zero comp ition, the fist and ast fined as-a vector whose NO? ned to q consecutive digits his non-zero cof whic! — ~ (00010001000) 1. A primary technique which is efecn, ret overcoming burst errors is interleaving ao hat i capable of correcting all burst errors A i es ealled as q-burst error correcting, or less is cal of heat stone burs coecting abil ‘Theorem : The number of parity check bits of a q burst err correcting code must have at least 2q is n-k22 Proof : We can proof this theorem by showing that following wo statements are true. @ Anecessary condition for a(n, k) linear code to beable to correct al burst of length q or less can ‘be a code vector. (@ Thenumber of parity check digits ofa (n, k) linear canal ieee , 2 am pr" | comes [Deana coting ee 3. Before the data stream is applied to chang data goes through a process of interleaving, error correction coding. 4. At the receiving side, data is decoded ss bits are evaluated in a manner to take: code has no burst of length b or less as a code . ‘vector is at least b is n_K>b. ! ‘Toprove part (i) consider a code vector v it a ; with burst Sinicicngmsteneects | % Gopetrhatat ) sry of cae and om tS: Ths in "Group of k/bts is loaded into shiftreg . then cote and v, must be in same coset. fone of Fo reni2=d into rows with [6 decors isa coset leader, then other will ke on ‘ coal a aa This ode wl not be able to To prove part (ii) consider vector ‘are confined 4 52) veto no ty ase ei therefore, there Process of codingiis held stored in first regi ee ora OR Udo burst| oe) ©2¢\ parity bits/ column oe ee) k bits/ column lea nterleaved code = (An, Xk) S15, 52) =U5, 35) ss Each row is @ 15 bit code word ————___. 36 | 41 6] 7 37 o7| 2 38. we | | 68] 73 ee 69 | 74 ah 70 | 75 Sai=5 with a burst error correcting ability = 10. searangement of codewords by interleaving as shown ‘rows and 5 columns. = Tetus assume that errors occur at position 5, 37, 43 469, The decoder operates on column and in each row x12 errors are possible. By this interleaving, these errors are corrected. §, 69 are random errors and 37 and 43 are burst 10, As we see that information bits in of particular codeword is /bits apart that original bits stream orwe can say two adjacent bits are /-bits distance apart like. 1. Whencaing ie complete, the entire content ofk x7 eee come well as Fx 1 parity bits are transmitted over channel. 4 Generally transmission carried out row by row in order, Cy Ciy coer Ce ane. 12. Here we see that parity bits received data is again received in sa intransmission and error correction is ‘The parity bits are discarded. tor enue , wala omer Seve bars (15, 7) be Gena = edna ee cole Rel 4" error correcting abitity of Be 4 7 as Sin 2 gins double error cx Hero we have ='5, eso Ai) Convotationa is isan another interleaving 2. In fig. we have transmitter sane four switches, There are teal ies ‘0 total | ines onthe beth oneoae and receiver side On the transmitter side, we have 0 storage ‘elements on line element ro Land storage jamneses As we move on transmitter sideline to line or receiver side, first line consists of 1)S storage element, second line (/—2)S and soon. The sum of storage elements on a specific line of transmitter and receiver is constant and equalito d-)s. |. _Here in fig. we have four switches that operates in step and more from line to line atrate of ‘Thus each switch move from linel.,thentoline2 and so on. . Let us considera single line fon transmiter side. Suppose that during a particular bit interval of d(K) there is a switch contact at both input and output sides of lineparted and sy" ey ‘ositches were comes tit hit of shit rene 4 until the next time consider tat initially all io one (atboth transmitter and as de) In this case received sequence is Same iter: 12, With shif register in transmit line has ((- 1) delay and therefore output segment ‘Suppose that rwo successive input bt stream are ‘transmitted d{(k) and d (k + 1) then d(k + 1) will ‘S=15 meaning 15 bits are interleaved between Source Code : There are two types of source code. Sourse codes are often divided into two broad categories 2 Convolutional codes information bits is followed derived rom he block check bits, are the information block B Tech Sem) CE oder | SHEL i icon message check Block of k ee <—| message = Maid codewords OPTS Bit rate=r, a (ae Error:control coding is the calculated gy redundancy. The functional blocks that accomplish errors coding are the channel encoder and the channel dex {Error control coding. Channel bit error probas ae=P {d # k d,} and message bit error probability Pib ak, b} ‘ b, - binary output r, bits/sec. ‘The channel encoder and decoder are func blocks in system that by acting together, reduee the probability of error. The encoder divides the inp bits into block of k message bits and replace exh message block D, with an n bit codeword Cw by act check bits to each message block. ‘The mapping coding and decoding are to be chosen such that control coding lowers the overall probability of Important aspects. of error controlling code: 1 Itis possible to detect and correct errosty’ ces called check bits to the message stream.les are as follows: where a s¢ pinary codes symbol 0 and 1 lable code is called an instantaneous 1y codeward is recognizable without ent code symbols, Prefix free codes instantaneous codes, caid to be optinal if t is instantaneous and n average L fora given source with a given ment for the source symbol Gaat i coding efficiency ? Show that the fading efficiency is maximum when P(0) = Pi. [R.TU, 2018, Dec. 2013] Design (m k) hamming code with a © inimum distance Of dj, = 3 and message length of 4 bits. [R.TU. Dec. 2013] iency : {ficiency is Jing efficiency : The code ef Pott aio of message bits to the number of bet. There message (mm ms ae ities [P(m,), PM) «+ POM) Letn, Ne umber of symbols in the i h message. average length of the message or the average length code word is than given by. : t= )1,P(q,) letters/message H shouldbe minimum to have an efficient transmission eficiency, then can be defined as ‘i hen Prove of coding efficiency is maximum W 1 =P) : LELH(e) be the entropy of the source in bine log m be the maximum average inform with each letter in bits/letter, bitvinensaage Dee/etion the minimum average Hence the coding efficieney Me fue fil LT loge We know that Hx) will be mane are equiproble: imum when symboe And the coding efficiency will be maim HQ) will be maximum, So we can conclue thar meen ve thet efficiency will be maximum when re P(0) = PCI) Let us see an example to prove it {M]=[m, m,] rou Tos y) _ Efficiency n=") j \y 10 -| \to4 \o + 4 (a) Generator Matrix (@) (b) Code Word ae c=MP tes v [6yC,C).= MyM MyM) cs C= MOM, Ms \ : 1 y= MOM, OM, C= MOM OMe Code table 5,3.2/5-20. 5 (©) Minimum: distance: Risequaltonininan sett rode vector. S° |. [000] Hence proved: jo 0 of a particular (7 4) ir The parity check matrix linear Block code is given BY [1110 100 H=|1101010 {on oot (a) Find the generator matrix G- (0) List all the code vectors: (o) What is the minimum distance beewee” the code vectors ? (@) How many errors cam De detected ee vi many can be corrected x Tt