An algorithm to improve the error rate performance of Accumulate-Repeat-Accumulate codes Tae-Ui Kim

Size: px
Start display at page:

Download "An algorithm to improve the error rate performance of Accumulate-Repeat-Accumulate codes Tae-Ui Kim"

Transcription

1 An algorithm to improve the error rate performance of Accumulate-Repeat-Accumulate codes Tae-Ui Kim The Graduate School Yonsei University Department of Electrical and Electronic Engineering

2 An algorithm to improve the error rate performance of Accumulate-Repeat-Accumulate codes Tae-Ui Kim A Thesis Submitted to the Graduate School of Yonsei University in Partial Fulfillment of the Requirements for the Degree of Master of Science Supervised by Professor Hong-Yeop Song, Ph.D. Department of Electrical and Electronic Engineering The Graduate School YONSEI University December 2006

3 This certifies that the thesis of Tae-Ui Kim is approved. Thesis Supervisor: Hong-Yeop Song Jong-Moon Chung Soo-Yong Choi The Graduate School Yonsei University December 2006

4 . כ 2... כ..... כ.. כ.,,,,,,,,,,,,,,,

5 ,,,...,

6 Contents List of Figures iv List of Tables v Abstract vi Introduction. Motivation....2 Overview ARA Codes and concepts of EMD 4 2. Protograph and ARA codes Construction of parity check matrix of ARA codes Concepts of EMD Improvement of waterfall and error rate performance in high SNR 5 3. Analysis of cycles of ARA codes Criterions for new connection Proposed algorithm i

7 3.4 Simulation results Concluding Remarks 30 Bibliography 3 Abstract (in Korean) 34 ii

8 List of Figures 2. Rate /3 RA code Rate /2 ARA code Parity check matrix involved in entire encoding process Reducing procedure of parity check matrix Parity check matrix of ARA codes Parity check matrix and bipartite graph description of a (9,3) code Venn diagram for relationship of C d,s d and L d A Cycle with EMD An example of consecutive parity bits cycle set L 4 with EMD An example for union of two L d s with size 7 and EMD An example for self return distance Selection of a parity bit and a check node All parity cycle produced by new edge connection Implementation of encoder for proposed algorithm Summary of proposed algorithm Performance comparison of RA and ARA codes iii

9 3.9 Performance of ARA code () Performance of ARA code (2) iv

10 List of Tables 3. Local girth distribution of two ARA codes Consecutive parity bits cycle set distribution of two ARA codes v

11 ABSTRACT An Algorithm to Improve the Error Rate Performance of Accumulate-Repeat-Accumulate Codes Tae-Ui Kim Department of Electrical and Electronic Eng. The Graduate School Yonsei University Contrary to the low threshold of ARA codes their high error floor does not make the waterfall last for a long range with respect to the error rate. In the thesis we propose an algorithm to keep up the waterfall steeper and improve the error rate performance of ARA codes in high SNR using EMD concepts. The algorithm analyzes the cycles containing some part of the dual diagonal and defines the consecutive parities cycles which consist of one interleaved bit and parity bits. After identifying the consecutive parities cycles of the ARA codes the algorithm supplies additional EMD to the consecutive parities cycles. Applying the proposed algorithm to ARA codes with two different interleavers achieves 0.35dB and 0.5dB gain respectively over AWGN channel with maximum iteration 200 and which only adds 9 and 5 edges for each code. Simulation results show that the proposed algorithm derives an improvement of waterfall and error rate performance in high SNR with little complexity burden to both encoder and decoder vi

12 without loss of performance in low SNR. Key words : Low-Density Parity-Check codes, Accumulate-Repeat-Accumulate Codes, EMD, stopping set, Error floor vii

13 Chapter Introduction. Motivation The codes based on iterative decoding are paid much attention after Berrou et al. introduced the turbo codes in 993. Another codes with iterative decoding, LDPC (Low- Density Parity-Check) codes are devised by Gallager in 963 and Mackay rediscovered the genuine value of LDPC codes [] [2]. Since the rediscovery many types of LDPC codes have been appeared. And the codes with turbo structure which can be decoded by decoding algorithm for LDPC codes have been prosed. RA (Repeat-Accumulate) codes, ARA (Accumulate-Repeat-Accumulate) codes and ARAA (Accumulate-Repeat- Accumulate-Accumulate) codes are the representatives of the codes [3] [4]. They are subclasses of LDPC codes. Since this type of codes consists of accumlators, repetition codes and interleavers they can be constructed based on the protograph which is tiny size of Tanner graph and the threshold can be derived from the protograph. The threshold of these codes is low. Divsalar et al. showed that these codes can provide near Shannon limit performance by the maximum likelihood decoding [5]. ARA codes are concatenated RA codes with rate outer accumulator to RA codes in order to achieve

14 lower threshold than RA codes. ARA codes are known that they have threshold 0.08 db for rate /2 and maximum bit node degree 5 as the block length goes to infinity [3]. However ARA codes have narrow waterfall region with respect to the error rate due to their high error floor even if the codes provide low threshold. Error floor can be improved by increasing the minimum distance. There are some researches to increase the minimum distance by achieving large girth of the graph [6]. In addition a scheme to improve the error floor by avoiding small size of stopping set was also proposed [7]. In this thesis, we analyzed the cycles produced by dual diagonal in the parity check matrix of ARA codes. Then, we propose a scheme to improve waterfall and performance in high SNR of ARA codes by increasing the EMD (Extrinsic Message Degree) of the particular cycles containing some part of the dual diagonal. We show that the proposed scheme is not only too simple to burden the much complexity to both encoder and decoder but also provide improvement of waterfall and performance in high SNR without loss in low SNR.2 Overview In chapter 2, We introduce the protograph codes and ARA codes as its application. After that we explain how to construct a parity check matrix of ARA codes in order to reduce the erasure bits. End of chapter 2 we give some theoretical definitions, theorems and examples which are bases of the proposed scheme. In chapter 3, we analyze the cycles produced by dual diagonal in the parity check matrix of ARA codes. We can find some particular cycles containing part of the dual diagonal which have small EMD. Based on the analysis we propose an algorithm to replenish the cycles with EMD to avoid 2

15 generation of small size of stopping set. It will enlarge the waterfall region and make it steeper and improve the error rate performance of ARA codes in high SNR ultimately. We present the simulation results to verify the proposal at the end of the chapter. Finally chapter 4 summarizes the remarks of the paper and gives some limits of the proposed scheme. 3

16 Chapter 2 ARA Codes and concepts of EMD In this chapter we first introduce ARA codes based on protograph and some basic theory about error floor and performance in hign SNR 2. Protograph and ARA codes Protograph is a Tanner graph with relatively small number of nodes [3]. A protograph consists of a set of variable nodes V, a set of check nodes C and a set of edges E. Each edge e E connects a variable node v e V to a check node c e C. The edges are permuted, so the mappings e (v e,c e ) V C construct an interleaver. One can obtain a larger graph with desired size by copy and permutation of a protograph. The resulting large graph is called derived graph and the corresponding LDPC code is protograph code. As a simple example we consider the protograph shown in Figure 2.a. This protograph code may be recognized as the Tanner graph of LDPC code (N = 3,K =)and give rise to rate /3 RA codes with encoder in Figure 2.b. In Figure 2.a black circles are variables nodes connected to the channel, white circles are erasures not connected to the channel and circles with plus sign are check nodes. 4

17 D (a) Protograph of RA (b) Encoder for RA Figure 2.: Rate /3 RA code D XO OX Rep 3 Rep 3 D OOX (a) Protograph of ARA (b) Encoder for ARA Figure 2.2: Rate /2 ARA code ARA codes are concatenated RA codes with rate outer accumulator having interleaver between them and can be also constructed based on a protograph and corresponding encoder. Figure 2.2 represents rate /2 ARA codes with regular repetition 3 and periodic puncturing pattern OOX, where O indicates the puncturing position. Outer accumulator provides precoding gain to achieve lower threshold than RA codes [3] [5]. But outer accumulator and puncturing for desired rate bring into production of many erasure bits. Protograph 2.2a of ARA code has erasure bits as many as /5 code length but erasure bits more than /5 are produced through the encoding process in real since the encoder adopts puncturing and outer accumulator. Hence the parity check matrix corresponding to protograph does not match to encoder and produced bits. In the next section we will explain how to construct a parity check matrix of ARA codes. 5

18 2.2 Construction of parity check matrix of ARA codes In the thesis we use ARA code in Figure 2.2. Entire encoded bits produced through the encoding process are composed of K information bit, K erased outer accumulator output bits and 3K parity bits of which 2/3 are erased parity bits. Hence the ratio of erasure bits for originally encoded codeword is 3/5. The ratio is very high and that degrades the peformance of the ARA codes. Erasure bits slow down the convergence speed since they pass no message by sum-product algorithm. Therefore we need to reduce the portion of erasure bits. The Figure 2.3 shows that the parity check matrix of ARA codes involved in all the bits produced through the encoding process, where the length of information bit k is 8 and the portion of erasure bits are 3/5. Two dual diagonal represents the inner and outer accumulator repectively. Outer accumulator produces 8 intermediate erasure bits e 0,e,,e 7 and 24 parity bits p 0,p,,p 23 with 8 information bits. Even numbered information bits and odd numbered intermediate erasure bits are repeated 3 times and put into the interleaver. Inner accumulator produces the 24 parity bits with interleaved input bits and transmit the bits indexed by 3k +2,k =0,,, 7. Only 6 bits of 40 bits are transmitted to the channel in total. But we can reconstruct the parity check matrix such that the codewords have small number of erasure bits. Let us see the parity check equations which produce intermediate erasure bits. 6

19 e 0 = k 0 e = e 0 + k e 2 = e + k 2 e 3 = e 2 + k 3. We can rewrite the parity check equations without even numbered punctured intermediate erasure bits as following e = k 0 + k e 3 = e 2 + k 3 = e + k 2 + k 3. These parity check equations give the parity check matrix with small number of intermediate erasure bits. Then we reorder the information bits by column permutations. By the procedures we obtain the reduced parity check matrix of outer accumulate code such as Figure2.4 We can apply these reducing procedure to the inner accumulate code in the same way. Finalizing whole procedure we get the parity check matrix in Figure 2.5 of which size are N = 5 2 K and M = 3 2K where K is the length of information. Now we call this reduced matrix as a parity check matrix of ARA codes and corresponding codeword of length N = 5 2K as a codeword of ARA codes. Note that the length of transmitted codeword is 2K since 2K erased bits are not transmitted. 7

20 k 0 k k 2 k 3 k 4 k 5 k 6 k 7 e 0 e e 2 e 3 e 4 e 5 e 6 e 7 p 0 p 23 p 22 p 2 p 20 p 9 p 8 p 7 p 6 p 5 p 4 p 3 p 2 p p 0 p 9 p 8 p 7 p 6 p 5 p 4 p 3 p 2 p Figure 2.3: Parity check matrix involved in entire encoding process 8

21 k 0 k k 2 k 3 k 4 k 5 k 6 k 7 e 0 e e 2 e 3 e 4 e 5 e 6 e 7 k k 3 k 5 k 7 k 0 k 2 k 4 k 6 e e 3 e 5 e 7 Figure 2.4: Reducing procedure of parity check matrix Parity check matrix with reduced erasure bits has erasure bits as much as /5 code length while the original parity check matrix has 3/5. The procedure needs a caution. When we carry out the method for the inner accumulate code consecutive three rows are merged in to one row. If a column has two s in these consecutive rows two s are deleted by modulo two sum since the bits corresponding to the column participate in merged check equation twice. To avoid this phenomenon we construct interleaver after reducing the parity check matrix. It doesn t makes the codes lose bit node degree, which may also degrade the performance of the code. ARA codes have interleaver. There are many ways to construct an interleaver such as interleavers for turbo codes. Under belief propagation PEG algorithm provides a graph with comparatively good performance [6]. It is reasonable to construct an interleaver by PEG algorithm since ARA codes are subclasses of LDPC codes. In Figure 2.5 PEG interleaver means that 3 repetition and interleaver are constructed using PEG algorithm from already given graph by protograph. Although parity check matrix and codeword have reduced portion of erasure bits the portion is still high. These erasure bits slow down the speed of convergence and degrade 9

22 d, d 3,3 v c Figure 2.5: Parity check matrix of ARA codes the performance of the iterative decoder. In addition ARA codes have small minimum distance bring into high error floor which let the codes loses their advantages. In the next section we will study the researches concerned in error floor and minimum distance of LDPC codes. 2.3 Concepts of EMD Minimum distance determines the performance of codes at high SNR. Large minimum distance reduces the probability that the decoded codeword satisfies the all check equation but it is not correct codeword. In order to derive the technique which increases minimum distance we introduce some definitions and theorems [7]. Figure 2.6 is an example for the definitions and theorems. The parity check matrix H can be divided by 2 matrices H and H 2. H represents message bits and H 2 represents parity bits. 0

23 v0 v v2 v3 v4 v5 v6 v7 v H H 2 (a) Parity check matrix c 0 c c 2 c 3 c 4 c 5 (b) Bipartite graph Figure 2.6: Parity check matrix and bipartite graph description of a (9,3) code Definition 2. Cycle A cycle of length 2d is a set of d variable nodes and d check nodes connected by edges such that a path exist and the path travels through every node in the set and connects each node to itself without traversing an edge twice. Definition 2.2 Cycle set C d A set of variable nodes in a bipartite graph is a set C d if it has d elements, and one or more cycles are formed between this set and its neighboring check node set. A set of d variable nodes does not form a set C d set only if no cycles exist between these variables and their check neighbors. The maximum cycle length that is possible to exist in a C d is 2d. Figure 2.6 shows a length-6 cycle of a bold line and a length-4 cycle of a dotted line. Note that the variable node set {v 4,v 5,v 6 } is also a set C 3 although v 5 is not contained in the length-4 cycle.

24 Definition 2.3 Stopping set S d A stopping set S d is called a set S d if it has d elements and all its neighbors are connected to it at least twice. Variable nodes set {v 0,v 4,v 6 } in Figure 2.6 is a set S 3 because all its neighbors c 0,c,c 3 and c 5 are connected to this set at least twice. In a bipartite graph without singly connected variable nodes, every stopping set contains cycles. Edges of variable node play a role of leaving from the node and arriving at node. This combination of leaving and arriving makes the cycles. Furthermore in a bipartite graph without singly connected variable nodes, stopping sets are comprised of multiple cycles in general. The only stopping set formed by a single cycle is one that consists of all degree-2 nodes. Definition 2.4 Linearly dependent set L d A variable node set is called an L d set if it is comprised of exactly d elements whose columns are linearly dependent but any subset of these columns is linearly independent, where the columns are column vectors of parity check matrix H corresponding to variables in L d. Variable node set {v 0,v 4,v 6 } in Figure 2.6 is an L 3 set. The N tuple vector with d s at positions corresponding to the variable nodes in L d and 0 at other positions satisfies all the parity check equations by the definition of L d. Hence a code with minimum distance d min has at least one L dmin because the variable nodes in the set are identical to the positions at which the minimum distance codewords have s. But L d with d<d min cannot exist since there is no codeword with hamming weight less than d min. Next theorem gives the relationships between L d s and S d s. 2

25 C d d S L d Figure 2.7: Venn diagram for relationship of C d,s d and L d Theorem 2. A set of variable nodes which form a set L d must form S d. Proof: The binary sum of all columns corresponding to the variable nodes in L d is the all-zero vector. Thus any neighbor of an L d is connected by the variable nodes in the set even number of times, that means at least twice. Theorem 2. implies that preventing small size of stopping set also prevents small d min. Since L dmin always forms S dmin a code with minimum distance d min always have L dmin and S dmin. Therefore, if we forbid a code to have all stopping sets S d for d t we can assure ourselves that the minimum distance d min of the code will be larger than t. However the converse is not true. Even if the small stopping set S d with d<texists minimum distance d min can be larger than t. Figure 2.7 shows the relationship between C d, S d and L d. Di, et al [8] showed that after we decode the received word in an iterative fashion until either the codeword has been recovered or until the decoder fails to progress further, a set of variable nodes left as erased is exactly equal to the maximum size stopping set that is a subset of originally erased symbol by the channel. The role of the stopping set under belief propagation is applicable to ARA codes since the codes contains erasure bits 3

26 in general. Even if the codes does not contain erasure bits it makes the theory applicable that the received bits with very small reliability can be considered as erasure bits. As we have known that preventing small size of stopping set also prevent small d min our purpose is to increase size of stopping set. Now we are going to present another definition to obtain a large size of stopping set. Definition 2.5 Extrinsic message degree (EMD) An extrinsic constraint node of a variable node set is a constraint node that is singly connected to this set. The EMD of a variable node set is the number of extrinsic constraint nodes of this variable node set. The EMD of stopping set is zero since every constraint nodes of variable nodes in the stopping set are connected to the set. A set of variable nodes with large EMD it will require more additional closure constraint nodes to be a stopping set. Therefore it will result in increased minimum distance that we make a set of variable nodes have large EMD. Figure 2.8 shows a cycle with EMD 2. The set requires additional variable nodes so that it may be a stopping set. Figure 2.8: A Cycle with EMD 2 4

27 Chapter 3 Improvement of waterfall and error rate performance in high SNR In this chapter we analyze the cycles in ARA codes and find particular cycles with small value of EMD containing some part of the dual diagonal. Then we propose a scheme to keep up the waterfall steeper and improve the error rate performance of ARA codes in high SNR by selective accumulation. 3. Analysis of cycles of ARA codes Now we are going to analyze the cycles in the bipartite graph of ARA codes. The bipartite graph structure is semi-deterministic due to protograh of the codes. Dual diagonal made by accumulator is such an example. Through the analysis of the dual diagonal which produces parity bits we can find some particular cycles with small value of EMD containing some part of the dual diagonal. The cycles passing through parity bits are comprised of interleaved bits and parity bits but we can find a special type of cycles passing through parity bits. 5

28 p i c j pi c j pi 2 c j 2 b k c j 3 i L4 (a) Consecutive parity bits cycle L i 4 (b) Appearance in parity check matrix Figure 3.: An example of consecutive parity bits cycle set L 4 with EMD 2 Definition 3. Consecutive parity bits cycle set L d A set of variable nodes and check nodes which consist of d consecutive parity bits, iterleaved bit and d consecutive check nodes, which contains some part of dual diagonal of the parity check matrix. Figure3. is an example for L 4 = {p i,p i+,p i+2,b k,c j,c j+,c j+2,c j+3 } where p s are parity node, b is interleaved bit and c s are check nodes. Superscript i is just an index of set according to size order. The parity bits provide no EMD since the degree of d parity bits in L d is 2 and all of them participate in the cycles. Thus EMD of consecutive parity bits cycle set is determined by degree of a interleaved bit node. Let us E(L d ) denote EMD of L d. We can calculate the E(L d ) as following. E(L d )=Degree(b k ) 2 The set size of L d is not only small but its EMD is also relatively small compared to the cycles of same size in the conventional LDPC codes. Since L d with small d has 6

29 small EMD and small size it is likely to be a small size of stopping set. In addition the parity bits receive low reliable message due to their low degree of two. Furthermore the union of some L d s also has small EMD compared to the cycles of same size in the conventional LDPC codes as well. But the union of L d s more than three are complicated to consider and the size and EMD of union are not too small. Thus in the thesis we are going to only consider the unions of two consecutive parity bits cycle sets. Let us consider unions of two L d s. Depending on the graph structure the event that EMD of the union is less than sum of individual EMD of the component set may occur. This event can happen when the two component set of uinon L i d and Lj d share the extrinsic check nodes or one set contains extrinsic check nodes of the other. Figure 3.2 shows an example that the union of an L 0 3 and an L 4 share the extrinsic check nodes. The size of the union is 7 and the EMD is 3. Their individual EMD s are 3 and 2 but the EMD of union is not 5 since they share a check node c s. These unions are especially weak in term of EMD compared to any other unions of L d s. Let us define deficient EMD for union of L i d and Lj d as measure of weakness in terms of EMD. Deficient EMD of union : E Def (i, j) E(L i d )+E(Lj d ) E(L i d Lj d ) The larger E Def, The weaker the union is in terms of EMD. We have investigated the specific cycles in ARA codes so far. Our proposed scheme to improve error floor of ARA codes is based on the consecutive parity bits cycles. We want to increase EMD of the consecutive parity bits cycle so that they may be a large size of stopping set. In order to do we will make a new edge connection to consecutive 7

30 c m 0 L 3 p n pi L 4 c j cm pn pi c j cm 2 bt pi 2 c j b k c j 2 c s Figure 3.2: An example for union of two L d s with size 7 and EMD 3 parity bits cycles and it can be identified by appearing new s down the dual diagonal. We have problems which L d, parity and check node do we choose. Next section will give the criterions. 3.2 Criterions for new connection It is preferred that new path induced by new edges involves more nodes in order that the new edges may make the L d s be a large stopping set when we add new edges to L d.we are going to define self return distance l s. Definition 3.2 Self return distance l s Self return distance is the shortest number of edges from L d to L d by spreading a tree from edge outside L d. Once the consecutive parity bits cycles are found we can calculate initial self return distance ls ini. We make a new connection so that adding a new edge does not get l s shorter 8

31 ini ls 4 ls 7 Figure 3.3: An example for self return distance than ls ini. Large l s lets the more variable nodes participate in stopping set including L d and increase EMD of union set. But l s does not have to be maximized since proper number of variable nodes involved in new path is enough. Figure 3.3 shows an example with l ini s =4. A new path involves in 3 variable nodes. The union of L d and these variable nodes will require more additional nodes to be a stopping set than the union Adding a new edge means we have to choose a parity bits and a check node. We keep the following rule when we choose a parity and a check nodes. Parity bits selection ARA codes have erasure bits in general as explained in chapter 2. We take a ARA code in Figure 2.2 as an example. A parity bit is connected to two check nodes and the check nodes are connected to another 4 variable nodes. These four variable nodes are composed of one parity bit, information bits and erasure bits. The more erasure bits a check node are connected to, the lower reliability the check nodes gives to a parity bit since 9

32 erasure bits passes no message under belief propagation. We call this scenario erasure and information bit connectivity. Therefore we choose a parity bits which is likely to receive low reliability from their check nodes considering erasure and information bit connectivity. If there are multiple choices of parity bits we select randomly. Check node selection Once selecting a parity bit we are faced to choose a check node to be connected to a chosen parity bit. Selecting a check node we keep the erasure and information bit connectivity in mind. In the sense of the connectivity it is also reasonable to choose a check node which has more information bits than erasure bits. Figure 3.4 shows selection of a parity bit and a check node in terms of erasure and information connectivity. But selection of a check node is more complicated. We have to consider things related to l s, EMD and cycles since there are so many candidate check nodes satisfying erasure and information connectivity. We prepare some criterions to choose a good check node.. Candidates have to satisfy erasure and check nodes connectivity. 2. New path induced by selection a pair of parity and check node does not make l s shorter than l ini s. 3. New connection makes E Def (i, j) decreased and E(L i d ) increased for a set Li d to be added with new edge compared to previous graph. 4. If there are multiple candidates choose a check nodes that gives maximum cycle length 20

33 Condition 3 means that new connection not only increase EMD of L d but does not also make the cases that union individual EMD of component set such as Figure 3.2. For every candidate check node to be connected to a consecutive cycle L i d the E(Li d ) and E(i, j) for j i is calculated from resulting graph. Assuming that new connection by the candidate is done new E(L i d ) have to be increased and new E(i, j) have to be decreased compared to previous E(L i d ) and E(i, j) respectively. Once we connect a new edge the new cycle comprised of only parity bits is produced as shown in Figure 3.5. Since this cycle consists of only parity bits the message passing through the cycles has low reliability. In order to overcome the resulting bad effect it is recommended to increase the length of the resulting all parity bits cycle. It is realized by permitting a check node apart from the previous check nodes of a chosen parity by some value D s. Then the length of resulting all parity bits cycle is 2(D s +). p p k k e p p k e e p k e e p k e e Figure 3.4: Selection of a parity bit and a check node 2

34 L 4 D s Figure 3.5: All parity cycle produced by new edge connection 3.3 Proposed algorithm Based on the previous discussion we will summarize the algorithm in Figure 3.7. Let us T, A max, A i and S denote total number of consecutive parity bits cycles, maximum number of added edges, number of added edges to L i and length of shortest all parity cycle respectively. In setp. all L d s are identified by analyzing the shortest cycles passing through the parity bits. For encoder implementation of proposed algorithm we take selective accumulations with chosen parity by a value D s corresponding to its added check nodes. Figure 3.6 shows the encoder implementation for proposed algorithm. The encoder for proposed algorithm contains only additional block for D s compared to original encoder Figure 2.2b The proposed algorithm raises little encoding and decoding complexity. The number of added s in the parity check matrix is not too many parity bits are selected 22

35 D s Figure 3.6: Implementation of encoder for proposed algorithm by the algorithm for ARA codes with K = 50 and code rate R =/2 in next section. And for a chosen parity the maximum number of added edges A max cannot be large since too many edges for a parity result in many all parity bits cycles. 23

36 step. Analysis Identify all L d s and index them from 0 to T according to increasing size. Determine D s and A max step 2. Initialization Calculate E(L d ), E Def (i, j) for 0 i, j T and l ini s A i =0 for all i stpe 3. Adding new edges begin for union if E Def (i, j) > 0 for 0 i<j T if A i <A max L i new edge such that l s ls ini, S>2Dand E Def(p, q) not increased A i = A i + end if A j <A max L j new edge such that l s ls ini, S>2Dand E Def(p, q) not increased A j = A j + begin for single cycle end if A i <A max for 0 i, j T L i new edge such that l s ls ini, S>2Dand E Def(p, q) not increased A i = A i + Figure 3.7: Summary of proposed algorithm 24

37 3.4 Simulation results In this section we present simulation results for ARA codes applying proposed algorithm over AWGN channel. We generate two ARA codes in Figure 2.2 with information bit K = 50 and code rate R =/2using PEG interleaver. The two ARA codes have local girth distributions shown in table 3.. Analyzing the shortest cylces passing through parity bits of ARA codes we can obtain distribution of consecutive parity bits cycles shown in the table 3.2. Based on the Table 3. and 3.2 we can expect that the code (2) tends to show better performance than the code (). local girth ARA code () ARA code (2) Table 3.: Local girth distribution of two ARA codes ARA code () ARA code (2) L L L L Table 3.2: Consecutive parity bits cycle set distribution of two ARA codes Following simulation results give performance evaluations. We transmitted all-zero codewords and used sum-product algorithm for decoding with maximum iteration 00 for Figure 3.8 and 200 for Figure 3.9 and 3.0. ARA codes are concatenated RA code with rate outer accumulator to achieve lower threshold than RA codes in the 25

38 0 0 Comparison to RA Code and Irregular LDPC code BER and FER RA code FER Irregular LDPC FER ARA code () FER ARA code (2) FER RA code BER Irregular LDPC BER ARA code () BER ARA code (2) + BER Eb/No Figure 3.8: Performance comparison of RA and ARA codes waterfall. Figure 3.8 shows the gain of ARA codes in the waterfall region, where RA codes with same repetition as one of ARA codes is used. Figure 3.8 also shows the performance comparison to irregular LDPC codes with degree distribution λ(x) = x x x x x x x 4 [9]. In the waterfall region ARA codes present better performance than irregular LDPC code but they give high error floor and narrow waterfall region with respect to error rate. We applied proposed algorithm to the ARA codes () and (2). The maximum number of added edges for each consecutive cycle is set to and D s is set to 0 for ARA code 26

39 () and 30 for ARA code (2) repectively. Thus the resulting the length of all parity bits cycle is guaranteed to be more than 20 and 60 for each code. We obtain results in Figure 3.9 for ARA code () with 9 edges added and Figure 3.0 with 5 edges added. The ARA code (2) shows better performance than ARA code () as expected. It is caused by distribution of local girth and consecutive parity bits cycles. The ARA code () contains more small size of consecutive parity bits cycles than ARA code (2) relatively. The proposed scheme provides 0.35 db gain for ARA code () and 0.5 db gain for ARA code (2) at FER 0 4. In terms of waterfall proposed algorithm makes the waterfall steeper and preserved longer than the original one. For comparison we added same number of edges to randomly selected parity bits using PEG algorithm which is known that gives reasonably good performance under iterative decoding. Adding new edges using PEG algorithm does not deteriorate the girth condition surely but gives no gain. The results are nearly same as original codes. On the observation of the simulation results we discuss the relationship between the measure of performance improvement and the distribution of consecutive parity bits cycles. It is significant that applying scheme to ARA code () derives more improvement than to ARA code (2). Proposed scheme utilizes the short consecutive parity bits cycles to improve waterfall and error rate performance in high SNR. That is the reason why application to ARA code () is more effective than ARA code (2). The improvement is caused by effective utilization of consecutive parity bits cycles. If there is no consecutive parity bits cycles we cannot use the propose scheme. But it is hardly possible to remove all the consecutive cycles. 27

40 0 0 Performance of proposed scheme,200 Iterations BER and FER ARA code () FER ARA code () + PEG 9 edges FER ARA code () + proposed FER ARA code () BER ARA code () + PEG 9 edges BER ARA code () + proposed BER Eb/No Figure 3.9: Performance of ARA code () 28

41 0 0 Performance of proposed scheme, 200 Iterations BER and FER ARA code (2) FER ARA code (2) + PEG 5edges FER ARA code (2) + proposed FER ARA code (2) BER ARA code (2) + PEG 5edges BER ARA code (2) + proposed BER Eb/No Figure 3.0: Performance of ARA code (2) 29

42 Chapter 4 Concluding Remarks ARA codes have particular cycles with small EMD and small size containing some part of dual diagonal. This cycles tends to be small stopping sets and result in small minimum distance of codewords. In the thesis we define consecutive parity bits cycle L d and propose a scheme to improve waterfall and error rate performance of ARA codes in high SNR by increasing their EMD. We verified that the proposed scheme is available to improve the performance by the simulation results. The proposed scheme is attractive since the scheme achieves the improvement of waterfall and error rate performance in high SNR with no loss in low SNR. In addition hardly the scheme burdens complexity to both encoder and decoder. However the proposed scheme have some limits.. If ARA codes does not have enough short consecutive cycles the proposed scheme is not effective. Because the proposed scheme utilize the short consecutive cycles. 2. Applying to RA codes is not effective. Through the analysis of cycles in RA codes we have known that RA codes does not have enough short consecutive cycles. It 30

43 is one reason why the proposed scheme is not useful for RA codes as. But we conjecture that other reasons make it unuseful for RA codes. 3. Analyzing the cycles and finding consecutive parity bits cycles are practical when the code length is moderate. Tree spreading of a graph has exponentially increasing complexity. If the code length is long proposed algorithm is no longer useful. 3

44 Bibliography [] R. G. Gallager, Low-density parity-check codes, IRE Trans. Inform. Theory, vol. IT-B, pp. 2 28, Jan [2] D. J. C. MacKay and R. M. Neal, Near Shannon limit performance of low density parity check codes, Electronics Letters, vol. 33, pp , Mar [3] A. Abbasfar, D. Divsalar and K. Yao Yang, Accumulate-repeat-accumulate codes, Golobal Telecommunications Conference, GLOBECOM, 04, IEEE, vol. 52, pp [4] H. Jin, A. Khandekar and R. McEliece, Irregular repeat-accumulate codes, in Proc. 2nd International Symposium on Turbo Codes, 2000, pp. 8. [5] A. Abbasfar, D. Divsalar and K. Yao Yang, Maximum likelihood decoding analysis of accumulate-repeat-accumulate codes, Global Telecommunicaitons Conference 2004, vol., pp [6] X. Hu, E. Eleftheriou and D.-M. Arnold, Progressive Edge-Growth Tanner Graphs, IEEE GlobeCom, vol. 2, pp , Nov

45 [7] T. Tian, C. Jones, J. D. Villasenor and R. D. Wesel, Construction of irregular LDPC codes with low error floors, ICC 03, vol. 5, pp [8] C. Di, D. Proietti, E. Telatar, T. Richardson and R, Urbanke, Finite length analysis of low-density parity-check codes on the bianry channel, IEEE Trans. Inform. Theory, vol 48, pp , June [9] T. Richardson, A. Shokrollahi and R. Urbanke, Design of capacity-approaching irregular low-density parity-check codesl, IEEE Trans. Inform. Theory, vol 47, pp , Feb

46 כ SNR Accumulate-Repeat-Accumulate ARA כ waterfall. EMD SNR ARA waterfall. ARA EMD EMD. ARA FER dB 0.5dB. SNR. : LDPC, Accumulate-Repeat-Accumulate, EMD, stopping, 34

Practical Polar Code Construction Using Generalised Generator Matrices

Practical Polar Code Construction Using Generalised Generator Matrices Practical Polar Code Construction Using Generalised Generator Matrices Berksan Serbetci and Ali E. Pusane Department of Electrical and Electronics Engineering Bogazici University Istanbul, Turkey E-mail:

More information

Efficient design of LDPC code using circulant matrix and eira code Seul-Ki Bae

Efficient design of LDPC code using circulant matrix and eira code Seul-Ki Bae Efficient design of LDPC code using circulant matrix and eira code Seul-Ki Bae The Graduate School Yonsei University Department of Electrical and Electronic Engineering Efficient design of LDPC code using

More information

Introduction to Low-Density Parity Check Codes. Brian Kurkoski

Introduction to Low-Density Parity Check Codes. Brian Kurkoski Introduction to Low-Density Parity Check Codes Brian Kurkoski kurkoski@ice.uec.ac.jp Outline: Low Density Parity Check Codes Review block codes History Low Density Parity Check Codes Gallager s LDPC code

More information

An Introduction to Low Density Parity Check (LDPC) Codes

An Introduction to Low Density Parity Check (LDPC) Codes An Introduction to Low Density Parity Check (LDPC) Codes Jian Sun jian@csee.wvu.edu Wireless Communication Research Laboratory Lane Dept. of Comp. Sci. and Elec. Engr. West Virginia University June 3,

More information

Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor

Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor Construction of low complexity Array based Quasi Cyclic Low density parity check (QC-LDPC) codes with low error floor Pravin Salunkhe, Prof D.P Rathod Department of Electrical Engineering, Veermata Jijabai

More information

Constructions of Nonbinary Quasi-Cyclic LDPC Codes: A Finite Field Approach

Constructions of Nonbinary Quasi-Cyclic LDPC Codes: A Finite Field Approach Constructions of Nonbinary Quasi-Cyclic LDPC Codes: A Finite Field Approach Shu Lin, Shumei Song, Lan Lan, Lingqi Zeng and Ying Y Tai Department of Electrical & Computer Engineering University of California,

More information

Structured Low-Density Parity-Check Codes: Algebraic Constructions

Structured Low-Density Parity-Check Codes: Algebraic Constructions Structured Low-Density Parity-Check Codes: Algebraic Constructions Shu Lin Department of Electrical and Computer Engineering University of California, Davis Davis, California 95616 Email:shulin@ece.ucdavis.edu

More information

Lecture 4 : Introduction to Low-density Parity-check Codes

Lecture 4 : Introduction to Low-density Parity-check Codes Lecture 4 : Introduction to Low-density Parity-check Codes LDPC codes are a class of linear block codes with implementable decoders, which provide near-capacity performance. History: 1. LDPC codes were

More information

LDPC Codes. Slides originally from I. Land p.1

LDPC Codes. Slides originally from I. Land p.1 Slides originally from I. Land p.1 LDPC Codes Definition of LDPC Codes Factor Graphs to use in decoding Decoding for binary erasure channels EXIT charts Soft-Output Decoding Turbo principle applied to

More information

LDPC Codes. Intracom Telecom, Peania

LDPC Codes. Intracom Telecom, Peania LDPC Codes Alexios Balatsoukas-Stimming and Athanasios P. Liavas Technical University of Crete Dept. of Electronic and Computer Engineering Telecommunications Laboratory December 16, 2011 Intracom Telecom,

More information

Enhancing Binary Images of Non-Binary LDPC Codes

Enhancing Binary Images of Non-Binary LDPC Codes Enhancing Binary Images of Non-Binary LDPC Codes Aman Bhatia, Aravind R Iyengar, and Paul H Siegel University of California, San Diego, La Jolla, CA 92093 0401, USA Email: {a1bhatia, aravind, psiegel}@ucsdedu

More information

Time-invariant LDPC convolutional codes

Time-invariant LDPC convolutional codes Time-invariant LDPC convolutional codes Dimitris Achlioptas, Hamed Hassani, Wei Liu, and Rüdiger Urbanke Department of Computer Science, UC Santa Cruz, USA Email: achlioptas@csucscedu Department of Computer

More information

ECEN 655: Advanced Channel Coding

ECEN 655: Advanced Channel Coding ECEN 655: Advanced Channel Coding Course Introduction Henry D. Pfister Department of Electrical and Computer Engineering Texas A&M University ECEN 655: Advanced Channel Coding 1 / 19 Outline 1 History

More information

Low Density Parity Check (LDPC) Codes and the Need for Stronger ECC. August 2011 Ravi Motwani, Zion Kwok, Scott Nelson

Low Density Parity Check (LDPC) Codes and the Need for Stronger ECC. August 2011 Ravi Motwani, Zion Kwok, Scott Nelson Low Density Parity Check (LDPC) Codes and the Need for Stronger ECC August 2011 Ravi Motwani, Zion Kwok, Scott Nelson Agenda NAND ECC History Soft Information What is soft information How do we obtain

More information

Modern Coding Theory. Daniel J. Costello, Jr School of Information Theory Northwestern University August 10, 2009

Modern Coding Theory. Daniel J. Costello, Jr School of Information Theory Northwestern University August 10, 2009 Modern Coding Theory Daniel J. Costello, Jr. Coding Research Group Department of Electrical Engineering University of Notre Dame Notre Dame, IN 46556 2009 School of Information Theory Northwestern University

More information

Iterative Encoding of Low-Density Parity-Check Codes

Iterative Encoding of Low-Density Parity-Check Codes Iterative Encoding of Low-Density Parity-Check Codes David Haley, Alex Grant and John Buetefuer Institute for Telecommunications Research University of South Australia Mawson Lakes Blvd Mawson Lakes SA

More information

Recent Results on Capacity-Achieving Codes for the Erasure Channel with Bounded Complexity

Recent Results on Capacity-Achieving Codes for the Erasure Channel with Bounded Complexity 26 IEEE 24th Convention of Electrical and Electronics Engineers in Israel Recent Results on Capacity-Achieving Codes for the Erasure Channel with Bounded Complexity Igal Sason Technion Israel Institute

More information

Performance Analysis and Code Optimization of Low Density Parity-Check Codes on Rayleigh Fading Channels

Performance Analysis and Code Optimization of Low Density Parity-Check Codes on Rayleigh Fading Channels Performance Analysis and Code Optimization of Low Density Parity-Check Codes on Rayleigh Fading Channels Jilei Hou, Paul H. Siegel and Laurence B. Milstein Department of Electrical and Computer Engineering

More information

On the minimum distance of LDPC codes based on repetition codes and permutation matrices 1

On the minimum distance of LDPC codes based on repetition codes and permutation matrices 1 Fifteenth International Workshop on Algebraic and Combinatorial Coding Theory June 18-24, 216, Albena, Bulgaria pp. 168 173 On the minimum distance of LDPC codes based on repetition codes and permutation

More information

Maximum Likelihood Decoding of Codes on the Asymmetric Z-channel

Maximum Likelihood Decoding of Codes on the Asymmetric Z-channel Maximum Likelihood Decoding of Codes on the Asymmetric Z-channel Pål Ellingsen paale@ii.uib.no Susanna Spinsante s.spinsante@univpm.it Angela Barbero angbar@wmatem.eis.uva.es May 31, 2005 Øyvind Ytrehus

More information

Low-density parity-check codes

Low-density parity-check codes Low-density parity-check codes From principles to practice Dr. Steve Weller steven.weller@newcastle.edu.au School of Electrical Engineering and Computer Science The University of Newcastle, Callaghan,

More information

Design of regular (2,dc)-LDPC codes over GF(q) using their binary images

Design of regular (2,dc)-LDPC codes over GF(q) using their binary images Design of regular (2,dc)-LDPC codes over GF(q) using their binary images Charly Poulliat, Marc Fossorier, David Declercq To cite this version: Charly Poulliat, Marc Fossorier, David Declercq. Design of

More information

Construction of Protographs for QC LDPC Codes With Girth Larger Than 12 1

Construction of Protographs for QC LDPC Codes With Girth Larger Than 12 1 Construction of Protographs for QC LDPC Codes With Girth Larger Than 12 1 Sunghwan Kim, Jong-Seon No School of Electrical Eng. & Com. Sci. Seoul National University, Seoul, Korea Email: {nodoubt, jsno}@snu.ac.kr

More information

Fountain Uncorrectable Sets and Finite-Length Analysis

Fountain Uncorrectable Sets and Finite-Length Analysis Fountain Uncorrectable Sets and Finite-Length Analysis Wen Ji 1, Bo-Wei Chen 2, and Yiqiang Chen 1 1 Beijing Key Laboratory of Mobile Computing and Pervasive Device Institute of Computing Technology, Chinese

More information

Quasi-Cyclic Asymptotically Regular LDPC Codes

Quasi-Cyclic Asymptotically Regular LDPC Codes 2010 IEEE Information Theory Workshop - ITW 2010 Dublin Quasi-Cyclic Asymptotically Regular LDPC Codes David G. M. Mitchell, Roxana Smarandache, Michael Lentmaier, and Daniel J. Costello, Jr. Dept. of

More information

Graph-based codes for flash memory

Graph-based codes for flash memory 1/28 Graph-based codes for flash memory Discrete Mathematics Seminar September 3, 2013 Katie Haymaker Joint work with Professor Christine Kelley University of Nebraska-Lincoln 2/28 Outline 1 Background

More information

On Bit Error Rate Performance of Polar Codes in Finite Regime

On Bit Error Rate Performance of Polar Codes in Finite Regime On Bit Error Rate Performance of Polar Codes in Finite Regime A. Eslami and H. Pishro-Nik Abstract Polar codes have been recently proposed as the first low complexity class of codes that can provably achieve

More information

Design of Non-Binary Quasi-Cyclic LDPC Codes by Absorbing Set Removal

Design of Non-Binary Quasi-Cyclic LDPC Codes by Absorbing Set Removal Design of Non-Binary Quasi-Cyclic LDPC Codes by Absorbing Set Removal Behzad Amiri Electrical Eng. Department University of California, Los Angeles Los Angeles, USA Email: amiri@ucla.edu Jorge Arturo Flores

More information

An Efficient Algorithm for Finding Dominant Trapping Sets of LDPC Codes

An Efficient Algorithm for Finding Dominant Trapping Sets of LDPC Codes An Efficient Algorithm for Finding Dominant Trapping Sets of LDPC Codes Mehdi Karimi, Student Member, IEEE and Amir H. Banihashemi, Senior Member, IEEE Abstract arxiv:1108.4478v2 [cs.it] 13 Apr 2012 This

More information

Belief-Propagation Decoding of LDPC Codes

Belief-Propagation Decoding of LDPC Codes LDPC Codes: Motivation Belief-Propagation Decoding of LDPC Codes Amir Bennatan, Princeton University Revolution in coding theory Reliable transmission, rates approaching capacity. BIAWGN, Rate =.5, Threshold.45

More information

BOUNDS ON THE MAP THRESHOLD OF ITERATIVE DECODING SYSTEMS WITH ERASURE NOISE. A Thesis CHIA-WEN WANG

BOUNDS ON THE MAP THRESHOLD OF ITERATIVE DECODING SYSTEMS WITH ERASURE NOISE. A Thesis CHIA-WEN WANG BOUNDS ON THE MAP THRESHOLD OF ITERATIVE DECODING SYSTEMS WITH ERASURE NOISE A Thesis by CHIA-WEN WANG Submitted to the Office of Graduate Studies of Texas A&M University in partial fulfillment of the

More information

Codes designed via algebraic lifts of graphs

Codes designed via algebraic lifts of graphs p./40 Codes designed via algebraic lifts of graphs Clemson Mini-Conference on Discrete Mathematics Oct. 3, 2008. Christine A. Kelley Department of Mathematics University of Nebraska-Lincoln email: ckelley2@math.unl.edu

More information

Low-Density Parity-Check codes An introduction

Low-Density Parity-Check codes An introduction Low-Density Parity-Check codes An introduction c Tilo Strutz, 2010-2014,2016 June 9, 2016 Abstract Low-density parity-check codes (LDPC codes) are efficient channel coding codes that allow transmission

More information

Capacity-approaching codes

Capacity-approaching codes Chapter 13 Capacity-approaching codes We have previously discussed codes on graphs and the sum-product decoding algorithm in general terms. In this chapter we will give a brief overview of some particular

More information

Low-density parity-check (LDPC) codes

Low-density parity-check (LDPC) codes Low-density parity-check (LDPC) codes Performance similar to turbo codes Do not require long interleaver to achieve good performance Better block error performance Error floor occurs at lower BER Decoding

More information

STUDY OF PERMUTATION MATRICES BASED LDPC CODE CONSTRUCTION

STUDY OF PERMUTATION MATRICES BASED LDPC CODE CONSTRUCTION EE229B PROJECT REPORT STUDY OF PERMUTATION MATRICES BASED LDPC CODE CONSTRUCTION Zhengya Zhang SID: 16827455 zyzhang@eecs.berkeley.edu 1 MOTIVATION Permutation matrices refer to the square matrices with

More information

Decoding of LDPC codes with binary vector messages and scalable complexity

Decoding of LDPC codes with binary vector messages and scalable complexity Downloaded from vbn.aau.dk on: marts 7, 019 Aalborg Universitet Decoding of LDPC codes with binary vector messages and scalable complexity Lechner, Gottfried; Land, Ingmar; Rasmussen, Lars Published in:

More information

On the Block Error Probability of LP Decoding of LDPC Codes

On the Block Error Probability of LP Decoding of LDPC Codes On the Block Error Probability of LP Decoding of LDPC Codes Ralf Koetter CSL and Dept. of ECE University of Illinois at Urbana-Champaign Urbana, IL 680, USA koetter@uiuc.edu Pascal O. Vontobel Dept. of

More information

Analysis of Sum-Product Decoding of Low-Density Parity-Check Codes Using a Gaussian Approximation

Analysis of Sum-Product Decoding of Low-Density Parity-Check Codes Using a Gaussian Approximation IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 47, NO. 2, FEBRUARY 2001 657 Analysis of Sum-Product Decoding of Low-Density Parity-Check Codes Using a Gaussian Approximation Sae-Young Chung, Member, IEEE,

More information

Extended MinSum Algorithm for Decoding LDPC Codes over GF (q)

Extended MinSum Algorithm for Decoding LDPC Codes over GF (q) Extended MinSum Algorithm for Decoding LDPC Codes over GF (q) David Declercq ETIS ENSEA/UCP/CNRS UMR-8051, 95014 Cergy-Pontoise, (France), declercq@ensea.fr Marc Fossorier Dept. Electrical Engineering,

More information

Message-Passing Decoding for Low-Density Parity-Check Codes Harish Jethanandani and R. Aravind, IIT Madras

Message-Passing Decoding for Low-Density Parity-Check Codes Harish Jethanandani and R. Aravind, IIT Madras Message-Passing Decoding for Low-Density Parity-Check Codes Harish Jethanandani and R. Aravind, IIT Madras e-mail: hari_jethanandani@yahoo.com Abstract Low-density parity-check (LDPC) codes are discussed

More information

Weaknesses of Margulis and Ramanujan Margulis Low-Density Parity-Check Codes

Weaknesses of Margulis and Ramanujan Margulis Low-Density Parity-Check Codes Electronic Notes in Theoretical Computer Science 74 (2003) URL: http://www.elsevier.nl/locate/entcs/volume74.html 8 pages Weaknesses of Margulis and Ramanujan Margulis Low-Density Parity-Check Codes David

More information

RCA Analysis of the Polar Codes and the use of Feedback to aid Polarization at Short Blocklengths

RCA Analysis of the Polar Codes and the use of Feedback to aid Polarization at Short Blocklengths RCA Analysis of the Polar Codes and the use of Feedback to aid Polarization at Short Blocklengths Kasra Vakilinia, Dariush Divsalar*, and Richard D. Wesel Department of Electrical Engineering, University

More information

THIS paper provides a general technique for constructing

THIS paper provides a general technique for constructing Protograph-Based Raptor-Like LDPC Codes for the Binary Erasure Channel Kasra Vakilinia Department of Electrical Engineering University of California, Los Angeles Los Angeles, California 90024 Email: vakiliniak@ucla.edu

More information

Optimal Rate and Maximum Erasure Probability LDPC Codes in Binary Erasure Channel

Optimal Rate and Maximum Erasure Probability LDPC Codes in Binary Erasure Channel Optimal Rate and Maximum Erasure Probability LDPC Codes in Binary Erasure Channel H. Tavakoli Electrical Engineering Department K.N. Toosi University of Technology, Tehran, Iran tavakoli@ee.kntu.ac.ir

More information

Polar Codes: Graph Representation and Duality

Polar Codes: Graph Representation and Duality Polar Codes: Graph Representation and Duality arxiv:1312.0372v1 [cs.it] 2 Dec 2013 M. Fossorier ETIS ENSEA/UCP/CNRS UMR-8051 6, avenue du Ponceau, 95014, Cergy Pontoise, France Email: mfossorier@ieee.org

More information

Codes on graphs and iterative decoding

Codes on graphs and iterative decoding Codes on graphs and iterative decoding Bane Vasić Error Correction Coding Laboratory University of Arizona Prelude Information transmission 0 0 0 0 0 0 Channel Information transmission signal 0 0 threshold

More information

Lower Bounds on the Graphical Complexity of Finite-Length LDPC Codes

Lower Bounds on the Graphical Complexity of Finite-Length LDPC Codes Lower Bounds on the Graphical Complexity of Finite-Length LDPC Codes Igal Sason Department of Electrical Engineering Technion - Israel Institute of Technology Haifa 32000, Israel 2009 IEEE International

More information

Integrated Code Design for a Joint Source and Channel LDPC Coding Scheme

Integrated Code Design for a Joint Source and Channel LDPC Coding Scheme Integrated Code Design for a Joint Source and Channel LDPC Coding Scheme Hsien-Ping Lin Shu Lin and Khaled Abdel-Ghaffar Department of Electrical and Computer Engineering University of California Davis

More information

CHAPTER 3 LOW DENSITY PARITY CHECK CODES

CHAPTER 3 LOW DENSITY PARITY CHECK CODES 62 CHAPTER 3 LOW DENSITY PARITY CHECK CODES 3. INTRODUCTION LDPC codes were first presented by Gallager in 962 [] and in 996, MacKay and Neal re-discovered LDPC codes.they proved that these codes approach

More information

Error Floors of LDPC Coded BICM

Error Floors of LDPC Coded BICM Electrical and Computer Engineering Conference Papers, Posters and Presentations Electrical and Computer Engineering 2007 Error Floors of LDPC Coded BICM Aditya Ramamoorthy Iowa State University, adityar@iastate.edu

More information

Research Letter Design of Short, High-Rate DVB-S2-Like Semi-Regular LDPC Codes

Research Letter Design of Short, High-Rate DVB-S2-Like Semi-Regular LDPC Codes Research Letters in Communications Volume 2008, Article ID 324503, 4 pages doi:0.55/2008/324503 Research Letter Design of Short, High-Rate DVB-S2-Like Semi-Regular LDPC Codes Luca Barletta and Arnaldo

More information

Turbo Code Design for Short Blocks

Turbo Code Design for Short Blocks Turbo Code Design for Short Blocks Thomas Jerkovits, Balázs Matuz Abstract This work considers the design of short parallel turbo codes (PTCs) with block lengths in the order of (a few) hundred code bits.

More information

Spatially Coupled LDPC Codes

Spatially Coupled LDPC Codes Spatially Coupled LDPC Codes Kenta Kasai Tokyo Institute of Technology 30 Aug, 2013 We already have very good codes. Efficiently-decodable asymptotically capacity-approaching codes Irregular LDPC Codes

More information

Bifurcations in iterative decoding and root locus plots

Bifurcations in iterative decoding and root locus plots Published in IET Control Theory and Applications Received on 12th March 2008 Revised on 26th August 2008 ISSN 1751-8644 Bifurcations in iterative decoding and root locus plots C.M. Kellett S.R. Weller

More information

Making Error Correcting Codes Work for Flash Memory

Making Error Correcting Codes Work for Flash Memory Making Error Correcting Codes Work for Flash Memory Part I: Primer on ECC, basics of BCH and LDPC codes Lara Dolecek Laboratory for Robust Information Systems (LORIS) Center on Development of Emerging

More information

THE seminal paper of Gallager [1, p. 48] suggested to evaluate

THE seminal paper of Gallager [1, p. 48] suggested to evaluate IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 50, NO. 11, NOVEMBER 2004 2657 Extrinsic Information Transfer Functions: Model and Erasure Channel Properties Alexei Ashikhmin, Member, IEEE, Gerhard Kramer,

More information

Chalmers Publication Library

Chalmers Publication Library Chalmers Publication Library Error Floor Analysis of Coded Slotted ALOHA over Packet Erasure Channels This document has been downloaded from Chalmers Publication Library (CPL). It is the author s version

More information

Turbo Codes are Low Density Parity Check Codes

Turbo Codes are Low Density Parity Check Codes Turbo Codes are Low Density Parity Check Codes David J. C. MacKay July 5, 00 Draft 0., not for distribution! (First draft written July 5, 998) Abstract Turbo codes and Gallager codes (also known as low

More information

Adaptive Cut Generation for Improved Linear Programming Decoding of Binary Linear Codes

Adaptive Cut Generation for Improved Linear Programming Decoding of Binary Linear Codes Adaptive Cut Generation for Improved Linear Programming Decoding of Binary Linear Codes Xiaojie Zhang and Paul H. Siegel University of California, San Diego, La Jolla, CA 9093, U Email:{ericzhang, psiegel}@ucsd.edu

More information

From Stopping sets to Trapping sets

From Stopping sets to Trapping sets From Stopping sets to Trapping sets The Exhaustive Search Algorithm & The Suppressing Effect Chih-Chun Wang School of Electrical & Computer Engineering Purdue University Wang p. 1/21 Content Good exhaustive

More information

Construction of Type-II QC LDPC Codes Based on Perfect Cyclic Difference Set

Construction of Type-II QC LDPC Codes Based on Perfect Cyclic Difference Set Chinese Journal of Electronics Vol24, No1, Jan 2015 Construction of Type-II QC LDPC Codes Based on Perfect Cyclic Difference Set ZHANG Lijun 1,LIBing 2 and CHENG Leelung 3 (1 School of Electronic and Information

More information

EE229B - Final Project. Capacity-Approaching Low-Density Parity-Check Codes

EE229B - Final Project. Capacity-Approaching Low-Density Parity-Check Codes EE229B - Final Project Capacity-Approaching Low-Density Parity-Check Codes Pierre Garrigues EECS department, UC Berkeley garrigue@eecs.berkeley.edu May 13, 2005 Abstract The class of low-density parity-check

More information

Coding Techniques for Data Storage Systems

Coding Techniques for Data Storage Systems Coding Techniques for Data Storage Systems Thomas Mittelholzer IBM Zurich Research Laboratory /8 Göttingen Agenda. Channel Coding and Practical Coding Constraints. Linear Codes 3. Weight Enumerators and

More information

An Introduction to Low-Density Parity-Check Codes

An Introduction to Low-Density Parity-Check Codes An Introduction to Low-Density Parity-Check Codes Paul H. Siegel Electrical and Computer Engineering University of California, San Diego 5/ 3/ 7 Copyright 27 by Paul H. Siegel Outline Shannon s Channel

More information

Low-Density Parity-Check Code Design Techniques to Simplify Encoding

Low-Density Parity-Check Code Design Techniques to Simplify Encoding IPN Progress Report 42-171 November 15, 27 Low-Density Parity-Check Code Design Techniques to Simplify Encoding J. M. Perez 1 and K. Andrews 2 This work describes a method for encoding low-density parity-check

More information

Performance Comparison of LDPC Codes Generated With Various Code-Construction Methods

Performance Comparison of LDPC Codes Generated With Various Code-Construction Methods Performance Comparison of LDPC Codes Generated With Various Code-Construction Methods Zsolt Polgar, Florin rdelean, Mihaly Varga, Vasile Bota bstract Finding good LDPC codes for high speed mobile transmissions

More information

Graph-based Codes and Iterative Decoding

Graph-based Codes and Iterative Decoding Graph-based Codes and Iterative Decoding Thesis by Aamod Khandekar In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California

More information

Achieving Flexibility in LDPC Code Design by Absorbing Set Elimination

Achieving Flexibility in LDPC Code Design by Absorbing Set Elimination Achieving Flexibility in LDPC Code Design by Absorbing Set Elimination Jiajun Zhang, Jiadong Wang, Shayan Garani Srinivasa, Lara Dolecek Department of Electrical Engineering, University of California,

More information

LDPC codes based on Steiner quadruple systems and permutation matrices

LDPC codes based on Steiner quadruple systems and permutation matrices Fourteenth International Workshop on Algebraic and Combinatorial Coding Theory September 7 13, 2014, Svetlogorsk (Kaliningrad region), Russia pp. 175 180 LDPC codes based on Steiner quadruple systems and

More information

Codes on Graphs. Telecommunications Laboratory. Alex Balatsoukas-Stimming. Technical University of Crete. November 27th, 2008

Codes on Graphs. Telecommunications Laboratory. Alex Balatsoukas-Stimming. Technical University of Crete. November 27th, 2008 Codes on Graphs Telecommunications Laboratory Alex Balatsoukas-Stimming Technical University of Crete November 27th, 2008 Telecommunications Laboratory (TUC) Codes on Graphs November 27th, 2008 1 / 31

More information

Pseudocodewords of Tanner Graphs

Pseudocodewords of Tanner Graphs SUBMITTED TO IEEE TRANSACTIONS ON INFORMATION THEORY 1 Pseudocodewords of Tanner Graphs arxiv:cs/0504013v4 [cs.it] 18 Aug 2007 Christine A. Kelley Deepak Sridhara Department of Mathematics Seagate Technology

More information

Introducing Low-Density Parity-Check Codes

Introducing Low-Density Parity-Check Codes Introducing Low-Density Parity-Check Codes Sarah J. Johnson School of Electrical Engineering and Computer Science The University of Newcastle Australia email: sarah.johnson@newcastle.edu.au Topic 1: Low-Density

More information

LDPC Decoder LLR Stopping Criterion

LDPC Decoder LLR Stopping Criterion International Conference on Innovative Trends in Electronics Communication and Applications 1 International Conference on Innovative Trends in Electronics Communication and Applications 2015 [ICIECA 2015]

More information

New Puncturing Pattern for Bad Interleavers in Turbo-Codes

New Puncturing Pattern for Bad Interleavers in Turbo-Codes SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 6, No. 2, November 2009, 351-358 UDK: 621.391.7:004.052.4 New Puncturing Pattern for Bad Interleavers in Turbo-Codes Abdelmounaim Moulay Lakhdar 1, Malika

More information

Trapping Set Enumerators for Specific LDPC Codes

Trapping Set Enumerators for Specific LDPC Codes Trapping Set Enumerators for Specific LDPC Codes Shadi Abu-Surra Samsung Telecommunications America 1301 E. Lookout Dr. Richardson TX 75082 Email: sasurra@sta.samsung.com David DeClercq ETIS ENSEA/UCP/CNRS

More information

On Generalized EXIT Charts of LDPC Code Ensembles over Binary-Input Output-Symmetric Memoryless Channels

On Generalized EXIT Charts of LDPC Code Ensembles over Binary-Input Output-Symmetric Memoryless Channels 2012 IEEE International Symposium on Information Theory Proceedings On Generalied EXIT Charts of LDPC Code Ensembles over Binary-Input Output-Symmetric Memoryless Channels H Mamani 1, H Saeedi 1, A Eslami

More information

Stopping, and Trapping Set Analysis

Stopping, and Trapping Set Analysis LDPC Codes Based on Latin Squares: Cycle Structure, Stopping, and Trapping Set Analysis Stefan Laendner and Olgica Milenkovic Electrical and Computer Engineering Department University of Colorado, Boulder,

More information

Bounds on the Error Probability of ML Decoding for Block and Turbo-Block Codes

Bounds on the Error Probability of ML Decoding for Block and Turbo-Block Codes Bounds on the Error Probability of ML Decoding for Block and Turbo-Block Codes Igal Sason and Shlomo Shamai (Shitz) Department of Electrical Engineering Technion Israel Institute of Technology Haifa 3000,

More information

Minimum Distance Bounds for Multiple-Serially Concatenated Code Ensembles

Minimum Distance Bounds for Multiple-Serially Concatenated Code Ensembles Minimum Distance Bounds for Multiple-Serially Concatenated Code Ensembles Christian Koller,Jörg Kliewer, Kamil S. Zigangirov,DanielJ.Costello,Jr. ISIT 28, Toronto, Canada, July 6 -, 28 Department of Electrical

More information

Low-complexity error correction in LDPC codes with constituent RS codes 1

Low-complexity error correction in LDPC codes with constituent RS codes 1 Eleventh International Workshop on Algebraic and Combinatorial Coding Theory June 16-22, 2008, Pamporovo, Bulgaria pp. 348-353 Low-complexity error correction in LDPC codes with constituent RS codes 1

More information

Random Redundant Soft-In Soft-Out Decoding of Linear Block Codes

Random Redundant Soft-In Soft-Out Decoding of Linear Block Codes Random Redundant Soft-In Soft-Out Decoding of Linear Block Codes Thomas R. Halford and Keith M. Chugg Communication Sciences Institute University of Southern California Los Angeles, CA 90089-2565 Abstract

More information

Spatially Coupled LDPC Codes Constructed from Protographs

Spatially Coupled LDPC Codes Constructed from Protographs IEEE TRANSACTIONS ON INFORMATION THEORY (SUBMITTED PAPER) 1 Spatially Coupled LDPC Codes Constructed from Protographs David G. M. Mitchell, Member, IEEE, Michael Lentmaier, Senior Member, IEEE, and Daniel

More information

Channel Codes for Short Blocks: A Survey

Channel Codes for Short Blocks: A Survey 11th International ITG Conference on Systems, Communications and Coding February 6, 2017 Channel Codes for Short Blocks: A Survey Gianluigi Liva, gianluigi.liva@dlr.de Fabian Steiner, fabian.steiner@tum.de

More information

Capacity-Achieving Ensembles for the Binary Erasure Channel With Bounded Complexity

Capacity-Achieving Ensembles for the Binary Erasure Channel With Bounded Complexity Capacity-Achieving Ensembles for the Binary Erasure Channel With Bounded Complexity Henry D. Pfister, Member, Igal Sason, Member, and Rüdiger Urbanke Abstract We present two sequences of ensembles of non-systematic

More information

Lecture 12. Block Diagram

Lecture 12. Block Diagram Lecture 12 Goals Be able to encode using a linear block code Be able to decode a linear block code received over a binary symmetric channel or an additive white Gaussian channel XII-1 Block Diagram Data

More information

Quasi-Cyclic Low-Density Parity-Check Codes With Girth Larger Than

Quasi-Cyclic Low-Density Parity-Check Codes With Girth Larger Than IEEE TRANSACTIONS ON INFORMATION THEORY, VOL 53, NO 8, AUGUST 2007 2885 n possible values If the parity check is satisfied, the error probability is closely approximated by the probability of two bit errors,

More information

Convergence analysis for a class of LDPC convolutional codes on the erasure channel

Convergence analysis for a class of LDPC convolutional codes on the erasure channel Convergence analysis for a class of LDPC convolutional codes on the erasure channel Sridharan, Arvind; Lentmaier, Michael; Costello Jr., Daniel J.; Zigangirov, Kamil Published in: [Host publication title

More information

Message Passing Algorithm with MAP Decoding on Zigzag Cycles for Non-binary LDPC Codes

Message Passing Algorithm with MAP Decoding on Zigzag Cycles for Non-binary LDPC Codes Message Passing Algorithm with MAP Decoding on Zigzag Cycles for Non-binary LDPC Codes Takayuki Nozaki 1, Kenta Kasai 2, Kohichi Sakaniwa 2 1 Kanagawa University 2 Tokyo Institute of Technology July 12th,

More information

Turbo Codes for Deep-Space Communications

Turbo Codes for Deep-Space Communications TDA Progress Report 42-120 February 15, 1995 Turbo Codes for Deep-Space Communications D. Divsalar and F. Pollara Communications Systems Research Section Turbo codes were recently proposed by Berrou, Glavieux,

More information

Turbo Compression. Andrej Rikovsky, Advisor: Pavol Hanus

Turbo Compression. Andrej Rikovsky, Advisor: Pavol Hanus Turbo Compression Andrej Rikovsky, Advisor: Pavol Hanus Abstract Turbo codes which performs very close to channel capacity in channel coding can be also used to obtain very efficient source coding schemes.

More information

ON THE MINIMUM DISTANCE OF NON-BINARY LDPC CODES. Advisor: Iryna Andriyanova Professor: R.. udiger Urbanke

ON THE MINIMUM DISTANCE OF NON-BINARY LDPC CODES. Advisor: Iryna Andriyanova Professor: R.. udiger Urbanke ON THE MINIMUM DISTANCE OF NON-BINARY LDPC CODES RETHNAKARAN PULIKKOONATTU ABSTRACT. Minimum distance is an important parameter of a linear error correcting code. For improved performance of binary Low

More information

Construction and Performance Evaluation of QC-LDPC Codes over Finite Fields

Construction and Performance Evaluation of QC-LDPC Codes over Finite Fields MEE10:83 Construction and Performance Evaluation of QC-LDPC Codes over Finite Fields Ihsan Ullah Sohail Noor This thesis is presented as part of the Degree of Master of Sciences in Electrical Engineering

More information

Upper Bounding the Performance of Arbitrary Finite LDPC Codes on Binary Erasure Channels

Upper Bounding the Performance of Arbitrary Finite LDPC Codes on Binary Erasure Channels Upper Bounding the Performance of Arbitrary Finite LDPC Codes on Binary Erasure Channels Chih-Chun Wang School of Electrical & Computer Engineering Purdue University West Lafayette, IN 47907, USA Email:

More information

IEEE C802.16e-04/264. IEEE Broadband Wireless Access Working Group <http://ieee802.org/16>

IEEE C802.16e-04/264. IEEE Broadband Wireless Access Working Group <http://ieee802.org/16> 24-9-2 I C82.6e-4/264 Project Title Date Submitted I 82.6 roadband Wireless Access Working Group Irregular Structured DPC Codes 24-8-7 Source(s) Victor Stolpman, Jianzhong (Charlie)

More information

An Integer Programming-Based Search Technique for Error-Prone Structures of LDPC Codes

An Integer Programming-Based Search Technique for Error-Prone Structures of LDPC Codes An Integer Programming-Based Search Technique for Error-Prone Structures of LDPC Codes Abdullah Sarıduman a, Ali E. Pusane a, Z. Caner Taşkın b a Department of Electrical and Electronics Engineering, Boğaziçi

More information

RECURSIVE CONSTRUCTION OF (J, L) QC LDPC CODES WITH GIRTH 6. Communicated by Dianhua Wu. 1. Introduction

RECURSIVE CONSTRUCTION OF (J, L) QC LDPC CODES WITH GIRTH 6. Communicated by Dianhua Wu. 1. Introduction Transactions on Combinatorics ISSN (print: 2251-8657, ISSN (on-line: 2251-8665 Vol 5 No 2 (2016, pp 11-22 c 2016 University of Isfahan wwwcombinatoricsir wwwuiacir RECURSIVE CONSTRUCTION OF (J, L QC LDPC

More information

Decoding Codes on Graphs

Decoding Codes on Graphs Decoding Codes on Graphs 2. Probabilistic Decoding A S Madhu and Aditya Nori 1.Int roduct ion A S Madhu Aditya Nori A S Madhu and Aditya Nori are graduate students with the Department of Computer Science

More information

AN INTRODUCTION TO LOW-DENSITY PARITY-CHECK CODES

AN INTRODUCTION TO LOW-DENSITY PARITY-CHECK CODES AN INTRODUCTION TO LOW-DENSITY PARITY-CHECK CODES Item Type text; Proceedings Authors Moon, Todd K.; Gunther, Jacob H. Publisher International Foundation for Telemetering Journal International Telemetering

More information

Construction of LDPC codes

Construction of LDPC codes Construction of LDPC codes Telecommunications Laboratory Alex Balatsoukas-Stimming Technical University of Crete July 1, 2009 Telecommunications Laboratory (TUC) Construction of LDPC codes July 1, 2009

More information