Threshold Saturation of Spatially-Coupled Codes on Intersymbol-Interference Channels

Size: px
Start display at page:

Download "Threshold Saturation of Spatially-Coupled Codes on Intersymbol-Interference Channels"

Transcription

1 hreshold Saturaton of Spatally-Coupled Codes on Intersymbol-Interference Channels Phong S Nguyen, Arvnd Yedla, Henry D Pfster, and Krshna R Narayanan Department of Electrcal and Computer Engneerng, exas A&M Unversty {psn, yarvnd, hpfster, krn}@tamuedu Abstract Recently, t has been observed that termnated lowdensty-party-check (LDPC) convolutonal codes (or spatallycoupled codes) appear to approach the capacty unversally across the class of bnary memoryless channels hs s facltated by the threshold saturaton effect whereby the belef-propagaton (BP) threshold of the spatally-coupled ensemble s boosted to the maxmum a-posteror (MAP) threshold of the underlyng consttuent ensemble In ths paper, we consder spatally-coupled codes over ntersymbol-nterference (ISI) channels under jont teratve decodng where we emprcally show that threshold saturaton also occurs hs can be observed by frst dentfyng the GEXI curve that naturally obeys the general area theorem From ths curve, the correspondng MAP and the BP threshold estmates are then numercally obtaned Gven the fact that regular LDPC codes can acheve the symmetrc nformaton rate (SIR) under MAP decodng, we conjecture that spatally-coupled codes wth jont teratve decodng can unversally approach the SIR of ISI channels I INRODUCION LDPC convolutonal codes, whch were ntroduced n [] and shown to have excellent BP thresholds n [2], have recently been observed to unversally approach the capacty of varous memoryless channels he fundamental mechansm behnd ths s explaned well n [3], where t s proven analytcally for the BEC that the BP threshold of a partcular spatally-coupled ensemble converges to the maxmum a-posteror (MAP) threshold of the underlyng ensemble A smlar result was also observed ndependently n [4] and stated as a conjecture Such a phenomenon s now called threshold saturaton va spatal couplng and has also been emprcally observed for general bnary-nput symmetrc-output memoryless (BMS) channels [5] In fact, threshold saturaton seems to be qute general and has now been observed n a wde range of problems In the realm of channels wth memory and partcularly ntersymbol nterference (ISI) channels, the capacty may not be achevable va equprobable sgnalng For lnear codes, a popular practce s to compare nstead wth the symmetrc nformaton rate (SIR), whch s also known as C ud [6], because ths rate s achevable by random lnear codes wth maxmum-lkelhood (ML) decodng For LDPC codes over ISI channels, a jont teratve BP decoder that operates on a large graph representng both the channel and the code constrants [6], [7] can perform qute well Progress has also been made on the desgn of SIR-approachng rregular LDPC hs materal s based upon work supported by the Natonal Scence Foundaton under Grant No he work of P Nguyen was also supported n part by a Vetnam Educaton Foundaton fellowshp Any opnons, fndngs, conclusons, or recommendatons expressed n ths materal are those of the authors and do not necessarly reflect the vews of the Natonal Scence Foundaton codes for some specfc ISI channels [8], [9], [0], [], [2] However, channel parameters must be known at the transmtter for such desgns and therefore unversalty appears dffcult to acheve Snce spatally-coupled codes and the threshold saturaton effect have now shown benefts n many communcaton problems, t s qute natural to consder them as a potental canddate to unversally approach the SIR of ISI channels wth low decodng complexty In fact, the combnaton of spatallycoupled codes and ISI channels was recently consdered by Kudekar and Kasa [3] for the smple dcode erasure channel (DEC) from [2], [4] hey provded a numercal evdence that the jont BP threshold of the spatally coupled codes can approach the SIR over the DEC However, the EXIlke curves they consdered were not equpped wth an area theorem and therefore could not be drectly connected wth the MAP threshold of the underlyng ensemble hus, the threshold saturaton effect was only ndrectly observed In ths paper, we frst focus on the case of general ISI channels where, by dervng the approprate GEXI curve and assocated area theorem, the MAP threshold upper bound can be computed and threshold saturaton can be seen As a consequence, t s possble for spatally-coupled codes to closely approach the SIR of ISI channels under jont teratve BP decodng because regular LDPC codes can acheve the SIR under MAP decodng [5] Also, we revst the transmsson of the spatally-coupled codes over the DEC as a specal case For ths channel, wth the MAP threshold estmated from the (G)EXI curve, threshold saturaton can also be observed to occur exactly Furthermore, the smplcty of the DEC allows us to provde a rgorous analyss of the upper bound on the MAP threshold A ISI Channels and the SIR II BACKGROUND Let the nput alphabet X be fnte, {X } Z be the dscretetme nput sequence (e, X X ) and {Y } Z be the dscretetme output sequence wth Y R Many ISI channels of nterest can be modeled by Y = ν t=0 a t X t + N, where the channel memory s ν, {a t } ν t=0 s the set of tap coeffcents and {N } Z s a sequence of ndependent nose random varables One can also wrte the above as Y = Z + N where Z = ν t=0 a t X t s the ISI channel output wthout nose In ths paper, we restrct ourselves to the class of bnary-nput ISI channels Often, the tap coeffcents are represented through a transform doman polynomal a(d) = ν t=0 a t D t he subject of Secton IV s the dcode erasure channel (DEC), whch s bascally a st-order dfferentator whose

2 d {f} a {x} DE c {L(y)} b {y} random permutaton channel outputs trells nodes bt nodes check nodes Fgure Gallager-anner-Wberg graph of the jont BP decoder for ISI channels he notatons a, b, c, d denote the average denstes of the messages traversng along the graph used n densty evoluton (DE) he quanttes nsde the brackets are erasure rates used n DE for the DEC case output s erased wth probablty ɛ and transmtted perfectly wth probablty ɛ Meanwhle, Secton III consders more general ISI channels among whch the most common s lnear ISI channels wth addtve whte Gaussan nose (AWGN) Unfortunately, no closed-form solutons for the SIR are known n ths case Instead, the numercal method descrbed n [6], [7] s typcally used to gve tght estmates of the SIR B LDPC Ensembles and the Jont BP Decoder he standard rregular LDPC ensemble s charactered by ts degree dstrbuton (dd), whch represents the fracton of nodes (or edges) of partcular degrees [8] From the edge perspectve, the dd par conssts of two polynomals λ(x) = λ x and ρ(x) = ρ x he LDPC ensemble can also be vewed from the node perspectve where ts dd par L(x) = L x and R(x) = R x he desgn rate of an LDPC ensemble s gven by r = L ()/R () When LDPC codes are transmtted over the ISI channel, one can construct a large graph by jonng the code graph and the channel graph together as depcted n Fg Workng on ths jont graph, a jont teratve decoder typcally passes the nformaton back and forth between the channel detector and the LDPC decoder hs technque s termed as turbo equalaton and was frst consdered by Doullard et al n the context of turbo codes [9] For analyss, we also requre the addton of a random scramblng vector to symmetre the effectve channel [20] hs s very smlar to usng a random coset of the LDPC code to allow analyss of the decoder usng the all-ero codeword assumpton; ths technque was also used n [6] where they proved a concentraton theorem and derved the densty evoluton (DE) equatons for ISI channels C Spatally-Coupled Ensembles he class of spatally-coupled ensembles can be defned qute broadly In ths paper, we manly consder two basc varants (see detals n [3]) as dscussed below ) he (l, r, L) ensemble: he (l, r, L) spatally-coupled ensemble (wth l odd so that ˆl = l N) can be constructed 2 from the underlyng (l, r)-regular LDPC ensemble At each poston from [, L] one has M bt nodes and l r M check nodes just lke n the (l, r)-regular case However, each bt node at poston s connected to one check node at each poston from ˆl to +ˆl In dong ths, one also needs to add l r M extra check nodes at each of ˆl extra postons on each sde For example, a jont code/channel graph for the (3, 6, L) ensemble and the ISI channels s shown n Fg 2 2) he (l, r, L, w) ensemble: he (l, r, L, w) can be obtaned wth the ntroducton of a smoothng parameter w One stll places M varable nodes at each poston n [, L] but places l M check nodes at each poston n [, L+w ] Each r bt node at poston s connected unformly and ndependently to a total of l check nodes at postons from the range [, + w ] III GENERAL ISI CHANNELS he man part of ths paper focuses on ISI channels wth general nose models he MAP upper bound for general bnary memoryless symmetrc channels was presented by Méasson et al and conjectured to be tght [2] For general ISI channels, we apply a smlar technque to gve an estmate of the MAP threshold of the underlyng ensemble by frst constructng the BP-GEXI curve that follows an area theorem he BP thresholds of the coupled ensembles are then computed va DE and the threshold saturaton effect s observed In addton, smulatons on the performance of the jont BP decoder for coupled codes of fnte length are conducted to valdate these thresholds A GEXI Curves for the ISI channels When the channel nput X n (X, X 2,, X n ) s chosen unformly at random from a sutable bnary lnear code, the ISI output wthout nose Z at some ndex s a dscrete random varable charactered by ts probablty mass functon p Z () for all n the alphabet Z For example, n the case of a dcode channel, Z = {0, +2, 2} and p Z (0) = 2, p Z (+2) = p Z ( 2) = 4 he channel from Z to Y s a Z -ary nput memoryless channel charactered by ts transton probablty densty p Y Z (y ) Wthout specfyng the ndex, we denote h H(Z Y ) and get h=h(z) p(y ) p()p(y ) log 2 { p( )p(y ) } dy Instead of lookng at a partcular channel, we assume that the channel from Z to Y s from a smooth famly {M(h )} h of Z -ary nput memoryless channels charactered by condtonal entropy h A further assumpton s made that all ndvdual channel famles are parametered n a smooth way by a common parameter 2 ɛ, e, h = H(Z Y )(ɛ) Wth the conventon that y y n y, defne φ (y ) {P Z Y ( y ) Z} and the random vector Φ φ (Y ) Each value of φ s a vector of length Z n the ( Z )- dmensonal probablty smplex he ndex of the vector assocated wth Z s denoted by [] One can see that Φ s a suffcent statstc for estmatng Z gven Y, e, Z Φ (Y ) Y forms a Markov chan 3 he code s proper [8, p 4] and ts dual code contans no codewords nvolvng only 0 s and a run of (ν + ) s 2 For the AWGN case, a convenent choce for ɛ s ɛ = 2σ 2 3 One way to see ths s to wrte P Y Z (y ) = P Z Y ( y ) Φ e [ P Y (y ) = ] P Z ( ) P Z ( ) P Y (y ), where e s the standard bass column vector wth a n the ndex [], and [] apply the result from [8, p 29]

3 L on on L Po s Po s t t trells node bt node check node M Fgure 2 he jont graph for the (l, r, L) ensemble over the ISI channels Illustrated n ths fgure s the case where l = 3 and r = 6 In the setup we consder, the bt transmsson s row by row where the order of transmsson wthn each row s mpled by the (green) arrows he (red) stars are to connect consecutve rows Defnton : Suppose the ntal state n the trells s S0 Let Xn chosen accordng to pxn (xn ) be the nput sequence, Zn be the ISI output sequence wthout nose and Yn be the fnal channel output sequence, e, Y s the result of transmttng Z over the smooth famly {M(h )}h of memoryless channels hen the th GEXI functon s H(Xn Yn (h,, hn ), S0 ) () G (h,, hn ) = h and the average GEXI functon s defned by G(h,, hn ) = n G (h,, hn ) For the case where all channel famles n = are the same, e, h = h, we have dh(xn Yn (h), S0 ) G(h) = n dh Lemma : Assume that all the channel famles are the same4, e, h = h hen, the th GEXI functon s gven by G (h) = p() v a, (v)κ, (v)dv where a, s the dstrbuton of the vector Φ gven Z =, v s a vector of length Z n the ( Z )-dmensonal probablty smplex and the GEXI kernel (for and ) s5 κ, (v) = v[ ] p(y ) p(y ) log2 { v[] p(y ) } dy p() p(y ) log2 { p( )p(y ) } dy p()p(y ) Proof: Suppose the ntal state s S0, we start by wrtng H(Xn Yn, S0 ) = H(Zn Yn, S0 ) = H(Z Yn, S0 ) + H(Z Yn, Z, S0 ) (2) For smplcty of notaton, we drop S0 n all the expressons although the dependency on S0 s always mpled From () and (2), t s clear that G (h) = H(Z Yn (h,, hn ) h =h h We also have H(Z Yn ) = H(Z Y, Φ (Y )) = p( )p(φ )p(y ) φ y p( φ )p(y ) log2 dy dφ p( φ )p(y ) (3) 4 Note that for the case of dfferent channel famles, one can stll compute the th GEXI functon as a functon of the common parameter 5 p(y ) s dependent on h and hence s dependent on where (3) follows from the Bayes theorem and the fact that p(, φ, y ) = p(, φ )p(y φ, ) = p( )p(φ )p(y ) (4) Note that (4) s true snce Y and Φ (Y ) are ndependent gven Z, e, Y Z Φ (Y ) akng dervatve and usng p( φ ) = p( y ), we get d p(y ) dh p( y )p(y ) log2 dy dφ p( y )p(y ) G (h) = p( ) p(φ ) φ y = p() a, (v)κ, (v)dv where κ, (v) = = y y v v[ ] p(y ) d p(y ) log2 { } dy dh v[] p(y ) v[ ] p(y ) h p(y ) log2 { } dy / v[] p(y ) Fnally, by seeng that h H(Z Y ( )) = p( )p(y ) = p() p(y ) log2 { } dy p()p(y ) y we obtan the result Remark : For σ = 0 n the AWGN case (or = 0 n erasure nose), h = 0 and a, s delta at v = e[] where e[] s the standard bass vector At ths extreme, G(0) = 0 snce κ, (v) = 0 At the other extreme σ (or at = for erasure nose), h = H(Z) (eg, 5 for the dcode channel) and G(h) = snce n ths case a, s delta at v[ ] = p( ) ) BP-GEXI curve (wth AWGN): In ths secton, we are partcularly nterested n computng the BP-GEXI functon for ISI channels wth AWGN In ths case, let ΦBP,` denote the extrnsc estmate of Z at the `th round of jont BP decodng If ΦBP,` s used nstead of Φ n the above formulas then one has the BP-GEXI (at the `th round) GBP,` n a smlar manner to [2] and the overall BP-GEXI GBP (h) = lm` GBP,` (h) Also, notce that the two extremes n Remark stll apply when the BP decoder s used nstead of the MAP decoder

4 (y e )2 2σ 2 2πσ 2 Next, AWGN mples that p(y ) = and then ɛ p(y ) = ((y ) 2 σ 2 )p(y ) herefore, the correspondng th BP-GEXI s G BP,l (h) = A B where A = B = p() v a BP,l, p() (v) log 2 { p(y ) { (y ) 2 σ 2 } v [] e ( )(2y ) 2σ v 2 } dy dv, [] p(y ) { (y ) 2 σ 2 } log 2 { p( ) p() e ( )(2y ) 2σ 2 } dy In the lmt of l, one can run the DE for ISI channels [6] to obtan the DE-FP and compute the quanttes A and B at ths FP Wth some abuse of notaton, let a (l), b (l), c (l) and d (l) denote the average densty of the bt-to-check, check-tobt, bt-to-trells and trells-to-bt messages, respectvely (see Fg ), at teraton l wth ntal values (at l = 0) beng 0, the delta functon at 0 Also, let n denote the densty of channel nose he DE update equaton for jont BP decodng of a general bnary-nput ISI channels s a (l) = d (l ) λ(b (l ) ), b (l) = ρ(a (l) ), c (l) = L(b (l) ), d (l) = Γ(c (l), n), where for a densty x, λ(x) = λ x ( ), ρ(x) = ρ x ( ) and L(x) = L x he operators and are the standard densty transformatons used n [8, p 8] he map Γ(, ) s not easy to compute n closed form for general trellses and often one needs to resort to the Monte Carlo methods (e, runnng the wndowed BCJR algorthm wth wndow parameter W on a long enough trells - see detals n [6]) to gve the estmates A smlar method was used to upper bound the MAP threshold for turbo codes over BMS channels [22] he denomnator B can be computed ether by numercal ntegraton or by Monte Carlo methods Meanwhle, the numerator A nvolves n the quantty v [] = p (Z = l ) where l denotes the computaton tree of depth l, rooted at ndex, whch ncludes all channel and code constrants assocated wth l teratons of decodng hs computaton tree l excludes the tree root y and s mpled by the decodng schedule n the DE equaton he quantty v [], due to complcatons from the trells, s not easy to obtan n closed form However, one can readly compute v [] as an extra output of the BCJR algorthm (whch s already used n DE) usng v [] s,s Z = α (s ) γ (s, s ) β (s ) where γ (s, s ) s probablty of the nput x that corresponds to the transton from state s (at tme ndex ) to state s at (tme ndex ) gven the computaton tree l Here, α ( ) and β ( ) are the standard forward and backward state probabltes n the BCJR algorthm Note that the scalng constant can be chosen so that v [] = G BP (h) h BP (3, 6) h MAP (3, 6) Area = r h Fgure 3 BP-GEXI curve for a (3, 6)-regular LDPC code over an AWGN dcode channel he upper bound h MAP s obtaned by settng the area under the BP-GEXI curve (the shaded regon) equal to the code rate B Upper Bound for the MAP hreshold he above-mentoned GEXI curve naturally obeys the area theorem H(Z) h G(h)dh = MAP H(Z) 0 G(h)dh = r herefore, one can apply the dscussed boundng technque, e, by fndng the largest value h MAP such that the area under the BP-GEXI curve equals the code rate, H(Z) G BP (h)dh = r, h MAP to obtan the MAP upper bound h MAP h MAP For example, the BP-GEXI curve for the (3, 6)-regular LDPC code over an AWGN dcode channel wth a(d) = ( D)/ 2 followng the analyss n Secton III-A s shown n Fg 3 In ths case, h BP (3, 6) 085 (the correspondng threshold measured n db s σ BP (3, 6) 703 ± 000 db) whle h MAP (3, 6) 0920 (or σ MAP (3, 6) 0959 ± 000 db) C Spatally-Coupled Codes on the ISI Channels Consder the (l, r, L) spatally-coupled ensemble For the ISI channels, the DE equaton for ths ensemble can be obtaned from the protograph chan n a smlar manner to the case of memoryless channels dscussed n [2] For each, j [ ˆl, L + ˆl], let a (l) j (and b(l) j ) denote the average densty of the messages from bt nodes at poston to check nodes at poston j (and the other way around) 6 Wth all the ntal message denstes (at l = 0) beng 0, the DE update equaton (for all [, L]) s a (l) j = d(l ) { b (l ) j }, j [ ˆl, + ˆl], j [ ˆl,+ˆl] j b (l) j = [j ˆl,j+ˆl] c (l) = b (l) j, j [ ˆl,+ˆl] d (l) = Γ(c (l), n) a (l) j, j [ ˆl, + ˆl], where j {j,,j t} x j and {,, t} x denote the operatons x j x j2 x jt and x x 2 x t, respectvely D Smulaton Results We start wth the (l, r, L) crcular ensemble obtaned by consderng all the postons > L of the protograph chan 6 For [, L], set a (l) j = +, the delta functon at +

5 Bt Error Rate σ SIR σ BP (5, 0, 44) σ BP (3, 6, 22) (3, 6, 22), M = 5000 (target) (3, 6, 22), M = 5000 (overall) (3, 6, 22), M = 502 (target) (5, 0, 44), M = 5000 (target) σ BP (3, 6) E b /N 0 (db) Fgure 4 BER and BP thresholds for the (3, 6)-regular, (3, 6, 22) and (5, 0, 44) spatally-coupled ensembles over the AWGN dcode channel to be the same as postons L (smlar to [5]) he order of bt transmssons s left to rght n each length-l row and then start wth the next row (n a total of M rows, see Fg 2) he I max(ν, l ) frst bts n each row are known hese bts wll break the crcular ensemble nto the (l, r, L I) ensemble and also serve as the plot bts to fx the trells state Due to ths fxng, one only needs to run the BCJR ndependently n each row and n a parallel manner [0], [] In our experments, we conduct smulatons over the AWGN dcode channel wth a(d) = ( D)/ 2 and memory ν = Frst, we use the DE n Sec III-C to compute the BP thresholds of the spatally-coupled codng scheme he results n Fg 4 reveal that σ BP (3, 6, 22) s roughly 0959 ± 000 db and approxmately the same as σ BP (3, 6, 44) whose rate loss s smaller Notce that ths s also roughly σ MAP (3, 6) - the MAP threshold estmate of the underlyng (3, 6)-regular ensemble, obtaned by the boundng technque, and s a sgnfcant mprovement over σ BP (3, 6) 703±000 db hs suggests that threshold saturaton occurs for regular ensembles Snce MAP decodng of regular ensembles can acheve the SIR [5], f threshold saturaton occurs, one can unversally approach the SIR of general ISI channels usng coupled codes wth jont teratve decodng o support ths, one can also see that for the (5, 0, 44) ensemble of the same rate as the (3, 6, 22) one, the threshold σ BP (5, 0, 44) 0834±000 db gets very close to the sgnal-to-nose rato (SNR) correspondng to the SIR (σ SIR 0823 ± 000 db usng the numercal method n [6], [7]) Although only smulatons wth the dcode channel are shown, the overall method s readly applcable to channels wth hgher memory Also shown n Fg 4 s the bt error rate (BER) versus SNR plot for the ensembles derved from the (l, r, L) crcular ensembles of fnte M = 502 and M = 5000 For each smulaton, we use l outer = 20 channel updates and between two such channel updates, we run l nner = 5 BP teratons on the code part alone he curves labeled target s the BER for the bts at poston I + (rght after the known bts) n the coupled chan whle the curve labeled overall s the overall BER for all the postons [I +, L] together One mght expect that the overall BER wll get closer to the target BER for large enough M and large enough number of teratons usng able I HRESHOLD ESIMAES OF (l, r)-regular ENSEMBLES OVER HE DEC AND DICODE AWGN CHANNEL FOR AWGN NOISE, HE HRESHOLDS ARE MEASURED IN db (l, r)- DEC Dcode AWGN regular ɛ BP ɛ MAP ɛ SIR σ BP σ MAP σ SIR (3, 6) (5, 0) an nducton argument From Fg 4, one can also observe that the overall BER for (3, 6, 22) and M = 5000 keeps gettng closer to the target BER as SNR slghtly ncreases hose BER curves are sgnfcantly mproved wth respect to ɛ BP (3, 6) - the BP threshold for the underlyng (3, 6)-regular ensemble IV ISI CHANNELS WIH ERASURE NOISE: HE DEC In ths secton, we brefly dscuss threshold saturaton on the DEC For ths channel, the GEXI curves mentoned above becomes dentcal (after scalng) to the EXI curves derved n [23] From these curves, one can also obtan a numercally tght upper bound on the MAP threshold of the underlyng ensemble and observe the threshold saturaton effect A Upper Bound on the MAP hreshold For the DEC and regular LDPC ensembles, the MAP upper bound was frst consdered n [23] In a recent report [24], the authors further provde a closed-form soluton for the BP- EXI and extended BP (EBP) EXI curves that can quckly gves an upper bound ɛ MAP on the MAP threshold by settng the area under the EBP curve (the shaded area n Fg 5) equal to the code rate he tghtness of the boundng technque s strongly suggested by a countng argument and for the (l, r)-regular ensemble, ths upper bound can be shown to quckly approach the erasure rate assocated wth the SIR when ncreasng l, r such that the code rate r s fxed (see arguments n [24] and facts n able I) B Spatally-Coupled Codes for the DEC Consder the (l, r, L, w) spatally-coupled ensemble We also follow the DE equaton dscussed n [3] to compute the BP thresholds of the coupled ensembles he man dfference s that we use the correct EBP-EXI curves wth ther operatonal meanng nstead of the EXI-lke curves used n [3] Let x (l) denote the expected erasure rate at teraton l from bt nodes at poston to check nodes where for [, L], one sets x (l) = 0 o compute both the stable and unstable FPs of DE, one can use the fxed entropy DE procedure outlned n [2, Sec VIII] where the normaled entropy of a constellaton x (l) = (x (l),, x(l) L ), whch s defned as χ(x (l) ) = L L = x (l), s kept constant at every teraton by varyng the channel parameter Wth each FP x obtaned, one obtans the EBP-EXI value of the spatally-coupled ensemble as L L = h EBP (x ) where h EBP ( ) s the EBP-EXI functon defned n [24] he threshold saturaton effect of couplng can be ncely seen by plottng the EBP-EXI curves for the uncoupled and coupled codes For example, Fg 5 shows the EBP curves for the (3, 6, L, 5) ensembles wth varous L along wth the

6 h EBP (ɛ) ɛ BP (3, 6) EBPcurve (3, 6) ɛ MAP (3, 6) L = 33 L = 7 L = 9 L = ɛ Fgure 5 EBP-EXI curves for (3, 6, L, 5) wth L = 2 ˆL + where ˆL = 2, 4, 8, 6, 32, 64, 28, 246 over the DEC As L grows larger, the rate loss becomes neglgble and the curves keep movng left, but they saturate at the MAP threshold of the underlyng regular ensemble EBP curve of the underlyng (3, 6)-regular ensemble From the EBP curves, one can determne ɛ BP (3, 6) and ɛ MAP (3, 6) he BP thresholds of spatally-coupled ensembles for small L due to rate-loss can have larger values, eg, ɛ BP (3, 6, 7, 6) > ɛ MAP (3, 6) However, for a wde range of L, e, L = 33, 65, 29, 257, 53, we observe that ɛ BP (3, 6, L, 5) whch s essentally ɛ MAP (3, 6) whle the rate loss gradually becomes nsgnfcant In [3], Kudekar and Kasa provded a smlar plot based on the EXIlke functon borrowed from the EXI functon of the BEC; though the pcture s smlar, t s not assocated wth an area theorem for the DEC Instead, we use a proper EXI functon h EBP and obtan the MAP threshold estmate ɛ MAP V CONCLUDING REMARKS In ths paper, we consder bnary communcaton over the ISI channels and numercally show that the threshold saturaton effect occurs on both the DEC and dcode channel wth AWGN o do ths, we construct the (G)EXI curves that satsfy the area theorem and obtan an upper bound on the threshold of the MAP decoder he upper bound s conjectured to be tght and, for the DEC, a numercal evdence can be shown to strongly support ths conjecture he observed threshold saturaton effect s valuable because by changng the underlyng regular LDPC ensemble combned wth the results of [5], t s shown that the jont BP decodng of spatally-coupled codes can unversally approach the SIR of the ISI channels Also, the convolutonal structure of the codes allows one to consder a wndowed decoder smlar to the one dscussed n [25], [26] All of these propertes suggest that spatally-coupled codes may be compettve n practce for systems wth ISI REFERENCES [] J Felstrom and K S Zgangrov, me-varyng perodc convolutonal codes wth low-densty party-check matrx, IEEE rans Inform heory, vol 45, no 6, pp 28 29, 999 [2] M Lentmaer, A Srdharan, D J Costello, and K S Zgangrov, Iteratve decodng threshold analyss for LDPC convolutonal codes, IEEE rans Inform heory, vol 56, no 0, pp , Oct 200 [3] S Kudekar, Rchardson, and R Urbanke, hreshold saturaton va spatal couplng: Why convolutonal LDPC ensembles perform so well over the BEC, IEEE rans Inform heory, vol 57, no 2, pp , 20 [4] M Lentmaer and G Fettwes, On the thresholds of generaled LDPC convolutonal codes based on protographs, n Proc IEEE Int Symp Inform heory, Austn, X, 200, pp [5] S Kudekar, C Méasson, Rchardson, and R Urbanke, hreshold saturaton on BMS channels va spatal couplng, n Proc Int Symp on urbo Codes & Iteratve Inform Proc, Sept 200, pp [6] A Kavčć, X Ma, and M Mtenmacher, Bnary ntersymbol nterference channels: Gallager codes, densty evoluton and code performance bounds, IEEE rans Inform heory, vol 49, no 7, pp , July 2003 [7] B M Kurkosk, P H Segel, and J K Wolf, Jont message-passng decodng of LDPC codes and partal-response channels, IEEE rans Inform heory, vol 48, no 6, pp , June 2002 [8] H D Pfster and P H Segel, Jont teratve decodng of LDPC codes and channels wth memory, n Proc 3rd Int Symp on urbo Codes & Related opcs, Brest, France, Sept 2003, pp 5 8 [9] N Varnca and A Kavčć, Optmed low-densty party-check codes for partal response channels, IEEE Commun Letters, vol 7, no 4, pp 68 70, 2003 [0] K R Narayanan and N Nangare, A BCJR-DFE based recever for achevng near capacty performance on nter symbol nterference channels, n Proc 43rd Annual Allerton Conf on Commun, Control, and Comp, Montcello, IL, Oct 2004, pp [] J B Soraga, H D Pfster, and P H Segel, Determnng and approachng achevable rates of bnary ntersymbol nterference channels usng multstage decodng, IEEE rans Inform heory, vol 53, no 4, pp , Aprl 2007 [2] H D Pfster and P H Segel, Jont teratve decodng of LDPC codes for channels wth memory and erasure nose, IEEE J Select Areas Commun, vol 26, no 2, pp , Feb 2008 [3] S Kudekar and K Kasa, hreshold saturaton on channels wth memory va spatal couplng, n Proc IEEE Int Symp Inform heory, St Petersburg, Russa, July 20, pp [4] H D Pfster, On the capacty of fnte state channels and the analyss of convolutonal accumulate-m codes, PhD dssertaton, Unversty of Calforna, San Dego, La Jolla, CA, USA, March 2003 [5] J H Bae and A Anastasopoulos, Capacty-achevng codes for fntestate channels wth maxmum-lkelhood decodng, IEEE J Select Areas Commun, vol 27, no 6, pp , Aug 2009 [6] D Arnold and H Loelger, On the nformaton rate of bnary-nput channels wth memory, n Proc IEEE Int Conf Commun, Helsnk, Fnland, June 200, pp [7] H D Pfster, J B Soraga, and P H Segel, On the achevable nformaton rates of fnte state ISI channels, n Proc IEEE Global elecom Conf, San Antono, exas, USA, Nov 200, pp [8] J Rchardson and R L Urbanke, Modern Codng heory Cambrdge Unversty Press, 2008 [9] C Doullard, M Jééquel, C Berrou, A Pcart, P Dder, and A Glaveux, Iteratve correcton of ntersymbol nterference: urbo equalaton, Eur rans elecom, vol 6, no 5, pp 507 5, Sept Oct 995 [20] J Hou, P H Segel, L B Mlsten, and H D Pfster, Capactyapproachng bandwdth-effcent coded modulaton schemes based on low-densty party-check codes, IEEE rans Inform heory, vol 49, no 9, pp , Sept 2003 [2] C Méasson, A Montanar, Rchardson, and R Urbanke, he generaled area theorem and some of ts consequences, IEEE rans Inform heory, vol 55, no, pp , Nov 2009 [22], Maxmum a posteror decodng and turbo codes for general memoryless channels, n Proc IEEE Int Symp Inform heory, Adelade, Australa, 2005, pp [23] C Wang and H D Pfster, Upper bounds on the MAP threshold of teratve decodng systems wth erasure nose, n Proc Int Symp on urbo Codes & Related opcs, Lausanne, Swterland, Sept 2008, pp 7 2 [24] P S Nguyen, A Yedla, H D Pfster, and K R Narayanan, Spatallycoupled codes and threshold saturaton on ntersymbol-nterference channels, 202, to be submtted to IEEE rans on Inform heory, [Onlne] Avalable: [25] A R Iyengar, M Papaleo, P H Segel, J K Wolf, A Vanell- Corall, and G E Coraa, Wndowed decodng of protographbased LDPC convolutonal codes over erasure channels, Oct 200, submtted to IEEE rans on Inform heory [Onlne] Avalable: [26] A R Iyengar, P H Segel, R L Urbanke, and J K Wolf, Wndowed decodng of spatally coupled codes, n Proc IEEE Int Symp Inform heory, St Petersburg, Russa, July 20, pp

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:

More information

arxiv: v1 [cs.it] 3 Oct 2011

arxiv: v1 [cs.it] 3 Oct 2011 Unversal Codes for the Gaussan MAC va Spatal Couplng Arvnd Yedla, Phong S. Nguyen, Henry D. Pfster, and Krshna R. Narayanan Department of Electrcal and Computer Engneerng, Teas A&M Unversty Emal: {yarvnd,psn,hpfster,krn}@tamu.edu

More information

Application of Nonbinary LDPC Codes for Communication over Fading Channels Using Higher Order Modulations

Application of Nonbinary LDPC Codes for Communication over Fading Channels Using Higher Order Modulations Applcaton of Nonbnary LDPC Codes for Communcaton over Fadng Channels Usng Hgher Order Modulatons Rong-Hu Peng and Rong-Rong Chen Department of Electrcal and Computer Engneerng Unversty of Utah Ths work

More information

Lecture 3: Shannon s Theorem

Lecture 3: Shannon s Theorem CSE 533: Error-Correctng Codes (Autumn 006 Lecture 3: Shannon s Theorem October 9, 006 Lecturer: Venkatesan Guruswam Scrbe: Wdad Machmouch 1 Communcaton Model The communcaton model we are usng conssts

More information

Lecture 10 Support Vector Machines II

Lecture 10 Support Vector Machines II Lecture 10 Support Vector Machnes II 22 February 2016 Taylor B. Arnold Yale Statstcs STAT 365/665 1/28 Notes: Problem 3 s posted and due ths upcomng Frday There was an early bug n the fake-test data; fxed

More information

Chapter 7 Channel Capacity and Coding

Chapter 7 Channel Capacity and Coding Chapter 7 Channel Capacty and Codng Contents 7. Channel models and channel capacty 7.. Channel models Bnary symmetrc channel Dscrete memoryless channels Dscrete-nput, contnuous-output channel Waveform

More information

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS

More information

Limited Dependent Variables

Limited Dependent Variables Lmted Dependent Varables. What f the left-hand sde varable s not a contnuous thng spread from mnus nfnty to plus nfnty? That s, gven a model = f (, β, ε, where a. s bounded below at zero, such as wages

More information

Chapter 7 Channel Capacity and Coding

Chapter 7 Channel Capacity and Coding Wreless Informaton Transmsson System Lab. Chapter 7 Channel Capacty and Codng Insttute of Communcatons Engneerng atonal Sun Yat-sen Unversty Contents 7. Channel models and channel capacty 7.. Channel models

More information

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 8 Luca Trevisan February 17, 2016

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 8 Luca Trevisan February 17, 2016 U.C. Berkeley CS94: Spectral Methods and Expanders Handout 8 Luca Trevsan February 7, 06 Lecture 8: Spectral Algorthms Wrap-up In whch we talk about even more generalzatons of Cheeger s nequaltes, and

More information

Error Probability for M Signals

Error Probability for M Signals Chapter 3 rror Probablty for M Sgnals In ths chapter we dscuss the error probablty n decdng whch of M sgnals was transmtted over an arbtrary channel. We assume the sgnals are represented by a set of orthonormal

More information

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U) Econ 413 Exam 13 H ANSWERS Settet er nndelt 9 deloppgaver, A,B,C, som alle anbefales å telle lkt for å gøre det ltt lettere å stå. Svar er gtt . Unfortunately, there s a prntng error n the hnt of

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

Finite Length Weight Enumerator Analysis of Braided Convolutional Codes

Finite Length Weight Enumerator Analysis of Braided Convolutional Codes Fnte Length Weght Enumerator Analyss of Braded Convolutonal Codes Saeedeh Moloud, Mchael Lentmaer, and Alexandre Graell Amat Department of Electrcal and Informaton Technology, Lund Unversty, Lund, Sweden

More information

Hidden Markov Models & The Multivariate Gaussian (10/26/04)

Hidden Markov Models & The Multivariate Gaussian (10/26/04) CS281A/Stat241A: Statstcal Learnng Theory Hdden Markov Models & The Multvarate Gaussan (10/26/04) Lecturer: Mchael I. Jordan Scrbes: Jonathan W. Hu 1 Hdden Markov Models As a bref revew, hdden Markov models

More information

Lossy Compression. Compromise accuracy of reconstruction for increased compression.

Lossy Compression. Compromise accuracy of reconstruction for increased compression. Lossy Compresson Compromse accuracy of reconstructon for ncreased compresson. The reconstructon s usually vsbly ndstngushable from the orgnal mage. Typcally, one can get up to 0:1 compresson wth almost

More information

NUMERICAL DIFFERENTIATION

NUMERICAL DIFFERENTIATION NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the

More information

Notes on Frequency Estimation in Data Streams

Notes on Frequency Estimation in Data Streams Notes on Frequency Estmaton n Data Streams In (one of) the data streamng model(s), the data s a sequence of arrvals a 1, a 2,..., a m of the form a j = (, v) where s the dentty of the tem and belongs to

More information

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results.

For now, let us focus on a specific model of neurons. These are simplified from reality but can achieve remarkable results. Neural Networks : Dervaton compled by Alvn Wan from Professor Jtendra Malk s lecture Ths type of computaton s called deep learnng and s the most popular method for many problems, such as computer vson

More information

Problem Set 9 Solutions

Problem Set 9 Solutions Desgn and Analyss of Algorthms May 4, 2015 Massachusetts Insttute of Technology 6.046J/18.410J Profs. Erk Demane, Srn Devadas, and Nancy Lynch Problem Set 9 Solutons Problem Set 9 Solutons Ths problem

More information

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE

CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE CHAPTER 5 NUMERICAL EVALUATION OF DYNAMIC RESPONSE Analytcal soluton s usually not possble when exctaton vares arbtrarly wth tme or f the system s nonlnear. Such problems can be solved by numercal tmesteppng

More information

Numerical Heat and Mass Transfer

Numerical Heat and Mass Transfer Master degree n Mechancal Engneerng Numercal Heat and Mass Transfer 06-Fnte-Dfference Method (One-dmensonal, steady state heat conducton) Fausto Arpno f.arpno@uncas.t Introducton Why we use models and

More information

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud

Resource Allocation with a Budget Constraint for Computing Independent Tasks in the Cloud Resource Allocaton wth a Budget Constrant for Computng Independent Tasks n the Cloud Wemng Sh and Bo Hong School of Electrcal and Computer Engneerng Georga Insttute of Technology, USA 2nd IEEE Internatonal

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

Global Sensitivity. Tuesday 20 th February, 2018

Global Sensitivity. Tuesday 20 th February, 2018 Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values

More information

Composite Hypotheses testing

Composite Hypotheses testing Composte ypotheses testng In many hypothess testng problems there are many possble dstrbutons that can occur under each of the hypotheses. The output of the source s a set of parameters (ponts n a parameter

More information

Electrical double layer: revisit based on boundary conditions

Electrical double layer: revisit based on boundary conditions Electrcal double layer: revst based on boundary condtons Jong U. Km Department of Electrcal and Computer Engneerng, Texas A&M Unversty College Staton, TX 77843-318, USA Abstract The electrcal double layer

More information

Power law and dimension of the maximum value for belief distribution with the max Deng entropy

Power law and dimension of the maximum value for belief distribution with the max Deng entropy Power law and dmenson of the maxmum value for belef dstrbuton wth the max Deng entropy Bngy Kang a, a College of Informaton Engneerng, Northwest A&F Unversty, Yanglng, Shaanx, 712100, Chna. Abstract Deng

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

Inductance Calculation for Conductors of Arbitrary Shape

Inductance Calculation for Conductors of Arbitrary Shape CRYO/02/028 Aprl 5, 2002 Inductance Calculaton for Conductors of Arbtrary Shape L. Bottura Dstrbuton: Internal Summary In ths note we descrbe a method for the numercal calculaton of nductances among conductors

More information

x = , so that calculated

x = , so that calculated Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to

More information

Vapnik-Chervonenkis theory

Vapnik-Chervonenkis theory Vapnk-Chervonenks theory Rs Kondor June 13, 2008 For the purposes of ths lecture, we restrct ourselves to the bnary supervsed batch learnng settng. We assume that we have an nput space X, and an unknown

More information

Parametric fractional imputation for missing data analysis. Jae Kwang Kim Survey Working Group Seminar March 29, 2010

Parametric fractional imputation for missing data analysis. Jae Kwang Kim Survey Working Group Seminar March 29, 2010 Parametrc fractonal mputaton for mssng data analyss Jae Kwang Km Survey Workng Group Semnar March 29, 2010 1 Outlne Introducton Proposed method Fractonal mputaton Approxmaton Varance estmaton Multple mputaton

More information

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification

2E Pattern Recognition Solutions to Introduction to Pattern Recognition, Chapter 2: Bayesian pattern classification E395 - Pattern Recognton Solutons to Introducton to Pattern Recognton, Chapter : Bayesan pattern classfcaton Preface Ths document s a soluton manual for selected exercses from Introducton to Pattern Recognton

More information

The Geometry of Logit and Probit

The Geometry of Logit and Probit The Geometry of Logt and Probt Ths short note s meant as a supplement to Chapters and 3 of Spatal Models of Parlamentary Votng and the notaton and reference to fgures n the text below s to those two chapters.

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix Lectures - Week 4 Matrx norms, Condtonng, Vector Spaces, Lnear Independence, Spannng sets and Bass, Null space and Range of a Matrx Matrx Norms Now we turn to assocatng a number to each matrx. We could

More information

Stat260: Bayesian Modeling and Inference Lecture Date: February 22, Reference Priors

Stat260: Bayesian Modeling and Inference Lecture Date: February 22, Reference Priors Stat60: Bayesan Modelng and Inference Lecture Date: February, 00 Reference Prors Lecturer: Mchael I. Jordan Scrbe: Steven Troxler and Wayne Lee In ths lecture, we assume that θ R; n hgher-dmensons, reference

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA

4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA 4 Analyss of Varance (ANOVA) 5 ANOVA 51 Introducton ANOVA ANOVA s a way to estmate and test the means of multple populatons We wll start wth one-way ANOVA If the populatons ncluded n the study are selected

More information

Implicit Integration Henyey Method

Implicit Integration Henyey Method Implct Integraton Henyey Method In realstc stellar evoluton codes nstead of a drect ntegraton usng for example the Runge-Kutta method one employs an teratve mplct technque. Ths s because the structure

More information

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

More information

On an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1

On an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1 On an Extenson of Stochastc Approxmaton EM Algorthm for Incomplete Data Problems Vahd Tadayon Abstract: The Stochastc Approxmaton EM (SAEM algorthm, a varant stochastc approxmaton of EM, s a versatle tool

More information

Convergence of random processes

Convergence of random processes DS-GA 12 Lecture notes 6 Fall 216 Convergence of random processes 1 Introducton In these notes we study convergence of dscrete random processes. Ths allows to characterze phenomena such as the law of large

More information

Open Systems: Chemical Potential and Partial Molar Quantities Chemical Potential

Open Systems: Chemical Potential and Partial Molar Quantities Chemical Potential Open Systems: Chemcal Potental and Partal Molar Quanttes Chemcal Potental For closed systems, we have derved the followng relatonshps: du = TdS pdv dh = TdS + Vdp da = SdT pdv dg = VdP SdT For open systems,

More information

Report on Image warping

Report on Image warping Report on Image warpng Xuan Ne, Dec. 20, 2004 Ths document summarzed the algorthms of our mage warpng soluton for further study, and there s a detaled descrpton about the mplementaton of these algorthms.

More information

ECE 534: Elements of Information Theory. Solutions to Midterm Exam (Spring 2006)

ECE 534: Elements of Information Theory. Solutions to Midterm Exam (Spring 2006) ECE 534: Elements of Informaton Theory Solutons to Mdterm Eam (Sprng 6) Problem [ pts.] A dscrete memoryless source has an alphabet of three letters,, =,, 3, wth probabltes.4,.4, and., respectvely. (a)

More information

Supplementary Notes for Chapter 9 Mixture Thermodynamics

Supplementary Notes for Chapter 9 Mixture Thermodynamics Supplementary Notes for Chapter 9 Mxture Thermodynamcs Key ponts Nne major topcs of Chapter 9 are revewed below: 1. Notaton and operatonal equatons for mxtures 2. PVTN EOSs for mxtures 3. General effects

More information

An Upper Bound on SINR Threshold for Call Admission Control in Multiple-Class CDMA Systems with Imperfect Power-Control

An Upper Bound on SINR Threshold for Call Admission Control in Multiple-Class CDMA Systems with Imperfect Power-Control An Upper Bound on SINR Threshold for Call Admsson Control n Multple-Class CDMA Systems wth Imperfect ower-control Mahmoud El-Sayes MacDonald, Dettwler and Assocates td. (MDA) Toronto, Canada melsayes@hotmal.com

More information

More metrics on cartesian products

More metrics on cartesian products More metrcs on cartesan products If (X, d ) are metrc spaces for 1 n, then n Secton II4 of the lecture notes we defned three metrcs on X whose underlyng topologes are the product topology The purpose of

More information

Lecture 7: Boltzmann distribution & Thermodynamics of mixing

Lecture 7: Boltzmann distribution & Thermodynamics of mixing Prof. Tbbtt Lecture 7 etworks & Gels Lecture 7: Boltzmann dstrbuton & Thermodynamcs of mxng 1 Suggested readng Prof. Mark W. Tbbtt ETH Zürch 13 März 018 Molecular Drvng Forces Dll and Bromberg: Chapters

More information

Lecture 12: Discrete Laplacian

Lecture 12: Discrete Laplacian Lecture 12: Dscrete Laplacan Scrbe: Tanye Lu Our goal s to come up wth a dscrete verson of Laplacan operator for trangulated surfaces, so that we can use t n practce to solve related problems We are mostly

More information

Consider the following passband digital communication system model. c t. modulator. t r a n s m i t t e r. signal decoder.

Consider the following passband digital communication system model. c t. modulator. t r a n s m i t t e r. signal decoder. PASSBAND DIGITAL MODULATION TECHNIQUES Consder the followng passband dgtal communcaton system model. cos( ω + φ ) c t message source m sgnal encoder s modulator s () t communcaton xt () channel t r a n

More information

One-sided finite-difference approximations suitable for use with Richardson extrapolation

One-sided finite-difference approximations suitable for use with Richardson extrapolation Journal of Computatonal Physcs 219 (2006) 13 20 Short note One-sded fnte-dfference approxmatons sutable for use wth Rchardson extrapolaton Kumar Rahul, S.N. Bhattacharyya * Department of Mechancal Engneerng,

More information

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009

College of Computer & Information Science Fall 2009 Northeastern University 20 October 2009 College of Computer & Informaton Scence Fall 2009 Northeastern Unversty 20 October 2009 CS7880: Algorthmc Power Tools Scrbe: Jan Wen and Laura Poplawsk Lecture Outlne: Prmal-dual schema Network Desgn:

More information

Low Complexity Soft-Input Soft-Output Hamming Decoder

Low Complexity Soft-Input Soft-Output Hamming Decoder Low Complexty Soft-Input Soft-Output Hammng Der Benjamn Müller, Martn Holters, Udo Zölzer Helmut Schmdt Unversty Unversty of the Federal Armed Forces Department of Sgnal Processng and Communcatons Holstenhofweg

More information

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence

More information

Solution Thermodynamics

Solution Thermodynamics Soluton hermodynamcs usng Wagner Notaton by Stanley. Howard Department of aterals and etallurgcal Engneerng South Dakota School of nes and echnology Rapd Cty, SD 57701 January 7, 001 Soluton hermodynamcs

More information

DC-Free Turbo Coding Scheme Using MAP/SOVA Algorithms

DC-Free Turbo Coding Scheme Using MAP/SOVA Algorithms Proceedngs of the 5th WSEAS Internatonal Conference on Telecommuncatons and Informatcs, Istanbul, Turkey, May 27-29, 26 (pp192-197 DC-Free Turbo Codng Scheme Usng MAP/SOVA Algorthms Prof. Dr. M. Amr Mokhtar

More information

Errors for Linear Systems

Errors for Linear Systems Errors for Lnear Systems When we solve a lnear system Ax b we often do not know A and b exactly, but have only approxmatons  and ˆb avalable. Then the best thng we can do s to solve ˆx ˆb exactly whch

More information

Bayesian predictive Configural Frequency Analysis

Bayesian predictive Configural Frequency Analysis Psychologcal Test and Assessment Modelng, Volume 54, 2012 (3), 285-292 Bayesan predctve Confgural Frequency Analyss Eduardo Gutérrez-Peña 1 Abstract Confgural Frequency Analyss s a method for cell-wse

More information

Markov Chain Monte Carlo (MCMC), Gibbs Sampling, Metropolis Algorithms, and Simulated Annealing Bioinformatics Course Supplement

Markov Chain Monte Carlo (MCMC), Gibbs Sampling, Metropolis Algorithms, and Simulated Annealing Bioinformatics Course Supplement Markov Chan Monte Carlo MCMC, Gbbs Samplng, Metropols Algorthms, and Smulated Annealng 2001 Bonformatcs Course Supplement SNU Bontellgence Lab http://bsnuackr/ Outlne! Markov Chan Monte Carlo MCMC! Metropols-Hastngs

More information

High resolution entropy stable scheme for shallow water equations

High resolution entropy stable scheme for shallow water equations Internatonal Symposum on Computers & Informatcs (ISCI 05) Hgh resoluton entropy stable scheme for shallow water equatons Xaohan Cheng,a, Yufeng Ne,b, Department of Appled Mathematcs, Northwestern Polytechncal

More information

Joint Statistical Meetings - Biopharmaceutical Section

Joint Statistical Meetings - Biopharmaceutical Section Iteratve Ch-Square Test for Equvalence of Multple Treatment Groups Te-Hua Ng*, U.S. Food and Drug Admnstraton 1401 Rockvlle Pke, #200S, HFM-217, Rockvlle, MD 20852-1448 Key Words: Equvalence Testng; Actve

More information

Markov Chain Monte Carlo Lecture 6

Markov Chain Monte Carlo Lecture 6 where (x 1,..., x N ) X N, N s called the populaton sze, f(x) f (x) for at least one {1, 2,..., N}, and those dfferent from f(x) are called the tral dstrbutons n terms of mportance samplng. Dfferent ways

More information

2 Finite difference basics

2 Finite difference basics Numersche Methoden 1, WS 11/12 B.J.P. Kaus 2 Fnte dfference bascs Consder the one- The bascs of the fnte dfference method are best understood wth an example. dmensonal transent heat conducton equaton T

More information

Linear Approximation with Regularization and Moving Least Squares

Linear Approximation with Regularization and Moving Least Squares Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...

More information

= z 20 z n. (k 20) + 4 z k = 4

= z 20 z n. (k 20) + 4 z k = 4 Problem Set #7 solutons 7.2.. (a Fnd the coeffcent of z k n (z + z 5 + z 6 + z 7 + 5, k 20. We use the known seres expanson ( n+l ( z l l z n below: (z + z 5 + z 6 + z 7 + 5 (z 5 ( + z + z 2 + z + 5 5

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

More information

Assortment Optimization under MNL

Assortment Optimization under MNL Assortment Optmzaton under MNL Haotan Song Aprl 30, 2017 1 Introducton The assortment optmzaton problem ams to fnd the revenue-maxmzng assortment of products to offer when the prces of products are fxed.

More information

THE ARIMOTO-BLAHUT ALGORITHM FOR COMPUTATION OF CHANNEL CAPACITY. William A. Pearlman. References: S. Arimoto - IEEE Trans. Inform. Thy., Jan.

THE ARIMOTO-BLAHUT ALGORITHM FOR COMPUTATION OF CHANNEL CAPACITY. William A. Pearlman. References: S. Arimoto - IEEE Trans. Inform. Thy., Jan. THE ARIMOTO-BLAHUT ALGORITHM FOR COMPUTATION OF CHANNEL CAPACITY Wllam A. Pearlman 2002 References: S. Armoto - IEEE Trans. Inform. Thy., Jan. 1972 R. Blahut - IEEE Trans. Inform. Thy., July 1972 Recall

More information

Lecture 4: Universal Hash Functions/Streaming Cont d

Lecture 4: Universal Hash Functions/Streaming Cont d CSE 5: Desgn and Analyss of Algorthms I Sprng 06 Lecture 4: Unversal Hash Functons/Streamng Cont d Lecturer: Shayan Oves Gharan Aprl 6th Scrbe: Jacob Schreber Dsclamer: These notes have not been subjected

More information

The Minimum Universal Cost Flow in an Infeasible Flow Network

The Minimum Universal Cost Flow in an Infeasible Flow Network Journal of Scences, Islamc Republc of Iran 17(2): 175-180 (2006) Unversty of Tehran, ISSN 1016-1104 http://jscencesutacr The Mnmum Unversal Cost Flow n an Infeasble Flow Network H Saleh Fathabad * M Bagheran

More information

A Robust Method for Calculating the Correlation Coefficient

A Robust Method for Calculating the Correlation Coefficient A Robust Method for Calculatng the Correlaton Coeffcent E.B. Nven and C. V. Deutsch Relatonshps between prmary and secondary data are frequently quantfed usng the correlaton coeffcent; however, the tradtonal

More information

Time-Varying Systems and Computations Lecture 6

Time-Varying Systems and Computations Lecture 6 Tme-Varyng Systems and Computatons Lecture 6 Klaus Depold 14. Januar 2014 The Kalman Flter The Kalman estmaton flter attempts to estmate the actual state of an unknown dscrete dynamcal system, gven nosy

More information

Pulse Coded Modulation

Pulse Coded Modulation Pulse Coded Modulaton PCM (Pulse Coded Modulaton) s a voce codng technque defned by the ITU-T G.711 standard and t s used n dgtal telephony to encode the voce sgnal. The frst step n the analog to dgtal

More information

Chapter 6. Supplemental Text Material

Chapter 6. Supplemental Text Material Chapter 6. Supplemental Text Materal S6-. actor Effect Estmates are Least Squares Estmates We have gven heurstc or ntutve explanatons of how the estmates of the factor effects are obtaned n the textboo.

More information

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity

Week3, Chapter 4. Position and Displacement. Motion in Two Dimensions. Instantaneous Velocity. Average Velocity Week3, Chapter 4 Moton n Two Dmensons Lecture Quz A partcle confned to moton along the x axs moves wth constant acceleraton from x =.0 m to x = 8.0 m durng a 1-s tme nterval. The velocty of the partcle

More information

An Application of Fuzzy Hypotheses Testing in Radar Detection

An Application of Fuzzy Hypotheses Testing in Radar Detection Proceedngs of the th WSES Internatonal Conference on FUZZY SYSEMS n pplcaton of Fuy Hypotheses estng n Radar Detecton.K.ELSHERIF, F.M.BBDY, G.M.BDELHMID Department of Mathematcs Mltary echncal Collage

More information

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers

Psychology 282 Lecture #24 Outline Regression Diagnostics: Outliers Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.

More information

VQ widely used in coding speech, image, and video

VQ widely used in coding speech, image, and video at Scalar quantzers are specal cases of vector quantzers (VQ): they are constraned to look at one sample at a tme (memoryless) VQ does not have such constrant better RD perfomance expected Source codng

More information

Quantum and Classical Information Theory with Disentropy

Quantum and Classical Information Theory with Disentropy Quantum and Classcal Informaton Theory wth Dsentropy R V Ramos rubensramos@ufcbr Lab of Quantum Informaton Technology, Department of Telenformatc Engneerng Federal Unversty of Ceara - DETI/UFC, CP 6007

More information

Feature Selection: Part 1

Feature Selection: Part 1 CSE 546: Machne Learnng Lecture 5 Feature Selecton: Part 1 Instructor: Sham Kakade 1 Regresson n the hgh dmensonal settng How do we learn when the number of features d s greater than the sample sze n?

More information

Lecture 14 (03/27/18). Channels. Decoding. Preview of the Capacity Theorem.

Lecture 14 (03/27/18). Channels. Decoding. Preview of the Capacity Theorem. Lecture 14 (03/27/18). Channels. Decodng. Prevew of the Capacty Theorem. A. Barg The concept of a communcaton channel n nformaton theory s an abstracton for transmttng dgtal (and analog) nformaton from

More information

Lecture 5 Decoding Binary BCH Codes

Lecture 5 Decoding Binary BCH Codes Lecture 5 Decodng Bnary BCH Codes In ths class, we wll ntroduce dfferent methods for decodng BCH codes 51 Decodng the [15, 7, 5] 2 -BCH Code Consder the [15, 7, 5] 2 -code C we ntroduced n the last lecture

More information

Entropy Coding. A complete entropy codec, which is an encoder/decoder. pair, consists of the process of encoding or

Entropy Coding. A complete entropy codec, which is an encoder/decoder. pair, consists of the process of encoding or Sgnal Compresson Sgnal Compresson Entropy Codng Entropy codng s also known as zero-error codng, data compresson or lossless compresson. Entropy codng s wdely used n vrtually all popular nternatonal multmeda

More information

A Hybrid Variational Iteration Method for Blasius Equation

A Hybrid Variational Iteration Method for Blasius Equation Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 1 (June 2015), pp. 223-229 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) A Hybrd Varatonal Iteraton Method

More information

Hidden Markov Models

Hidden Markov Models Hdden Markov Models Namrata Vaswan, Iowa State Unversty Aprl 24, 204 Hdden Markov Model Defntons and Examples Defntons:. A hdden Markov model (HMM) refers to a set of hdden states X 0, X,..., X t,...,

More information

Lecture 21: Numerical methods for pricing American type derivatives

Lecture 21: Numerical methods for pricing American type derivatives Lecture 21: Numercal methods for prcng Amercan type dervatves Xaoguang Wang STAT 598W Aprl 10th, 2014 (STAT 598W) Lecture 21 1 / 26 Outlne 1 Fnte Dfference Method Explct Method Penalty Method (STAT 598W)

More information

CS : Algorithms and Uncertainty Lecture 17 Date: October 26, 2016

CS : Algorithms and Uncertainty Lecture 17 Date: October 26, 2016 CS 29-128: Algorthms and Uncertanty Lecture 17 Date: October 26, 2016 Instructor: Nkhl Bansal Scrbe: Mchael Denns 1 Introducton In ths lecture we wll be lookng nto the secretary problem, and an nterestng

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

Assignment 2. Tyler Shendruk February 19, 2010

Assignment 2. Tyler Shendruk February 19, 2010 Assgnment yler Shendruk February 9, 00 Kadar Ch. Problem 8 We have an N N symmetrc matrx, M. he symmetry means M M and we ll say the elements of the matrx are m j. he elements are pulled from a probablty

More information

Digital Signal Processing

Digital Signal Processing Dgtal Sgnal Processng Dscrete-tme System Analyss Manar Mohasen Offce: F8 Emal: manar.subh@ut.ac.r School of IT Engneerng Revew of Precedent Class Contnuous Sgnal The value of the sgnal s avalable over

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011 Stanford Unversty CS359G: Graph Parttonng and Expanders Handout 4 Luca Trevsan January 3, 0 Lecture 4 In whch we prove the dffcult drecton of Cheeger s nequalty. As n the past lectures, consder an undrected

More information

ECE559VV Project Report

ECE559VV Project Report ECE559VV Project Report (Supplementary Notes Loc Xuan Bu I. MAX SUM-RATE SCHEDULING: THE UPLINK CASE We have seen (n the presentaton that, for downlnk (broadcast channels, the strategy maxmzng the sum-rate

More information

Adiabatic Sorption of Ammonia-Water System and Depicting in p-t-x Diagram

Adiabatic Sorption of Ammonia-Water System and Depicting in p-t-x Diagram Adabatc Sorpton of Ammona-Water System and Depctng n p-t-x Dagram J. POSPISIL, Z. SKALA Faculty of Mechancal Engneerng Brno Unversty of Technology Techncka 2, Brno 61669 CZECH REPUBLIC Abstract: - Absorpton

More information

Uncertainty in measurements of power and energy on power networks

Uncertainty in measurements of power and energy on power networks Uncertanty n measurements of power and energy on power networks E. Manov, N. Kolev Department of Measurement and Instrumentaton, Techncal Unversty Sofa, bul. Klment Ohrdsk No8, bl., 000 Sofa, Bulgara Tel./fax:

More information

Lecture Notes on Linear Regression

Lecture Notes on Linear Regression Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume

More information