On The Iterative Decoding Analysis And Its Application In Coded Cooperative Communication

Since the iterative decoding can make Turbo code and LDPC code have approaching Shan-non limit performance with small decoding complexity,it has become a hot topic in the field of coding research after the advent of Turbo code and the rediscovery of LDPC code.Now,the iterative decoding not only is used in the decoding of error correction codes to obtain the near capacity performance,but also becomes an important method to solve other communication problems and improve the system transmission reliability.But so far,no analytical method can accurately estimate the performance of iterative decoding,and explain why the iterative de-coding mechanism has different decoding trajectories in different signal to noise ratio(SNR)regions.Especially,the performance analysis of the waterfall area has a lot of theoretical prob-lems to be further analyzed.In addition,the complete performance analysis and modeling of the iterative decoding not only can be used as the basis or criterion for optimizing the designs of Turbo code and LDPC code,but also can evaluate the performance of related communication systems with Turbo code or LDPC code.Therefore,(1)this thesis deeply studies the iterative decoding to analyze the key factors affecting the performance of the iterative decoding,accu-rately estimates the performance of the iterative decoding,and improves the performance of the existing codes;(2)this thesis conducts in-depth research on the application of iterative decoding mechanism in network coded cooperation communication,whose aim is to design a reasonable network coded cooperation mechanism to improve both the throughput of the network and the reliability of the transmission.Firstly,the iterative decoding of the Turbo code is considered.By regarding the whole iterative decoding of the Turbo code as a nonlinear dynamic system with feedback,the recur-sive iterative calculation formula of the extrinsic information can be derived,and the equation that the convergence solution should satisfy is directly obtained.However,since the recursive iterative calculation formula of the the extrinsic information is a high-dimensional nonlinear equation,so the stochastic analysis method is used to directly analyze the probability distri-bution of the LLR(Log-likelihood ratio)of the iterative decoder to unveil the Turbo iterative decoding mechanism.The mixture Gaussian model is proposed to accurately approximate the probability density function(PDF)of the LLR of the iterative decoder,and then a BER esti-mator is provided to accurately estimate the entire performance curve of Turbo code.When all 0s information input is assumed,it is found that the stochastic distribution of LLR will con-centrate around two centers,and the one whose center is close to zero(the left sidelobe)will dominate the behavior and the achieved performance of the iterative decoding,which is a fact that is always missed in previous LLR analysis for Turbo codes.The iterative decoding can be viewed as a process that gradually moves the LLR value from the unreliable area to the reliable area.But,as the SNR increases,the left sidelobe becomes smaller and seems to be disap-peared.Therefore,it is proposed to use the noise-free decoding analysis method to determine the sensitive bits that affect the error floor's performance.By comparison,these sensitive bits and the key bits obtained by using the search algorithm of the weight spectrum are almost the same.However,the complexity of the noise-free decoding analysis method is far less than that of the weight spectrum search algorithm.Secondly,the noise-free decoding method is used to analyze the irregular LDPC code.For the irregular LDPC,the LLR reliability ordering is consistent to the degree distribution,which also states the fact that the degree distribution dominates the performance of a irregular LDPC code.And in addition to the degree distribution,there are other factors that affect the performance of the irregular LDPC code,such as the girth and stopping set,so the bits with the same degree have different LLR reliabilities.On this basis,on the LLR reliability criterion based shortening design for irregular LDPC codes is proposed,where the shortened informa-tion indices is determined according to the LLR reliability ordering.Compared with the other shortening methods,the proposed shortening scheme has lower computational complexity,and the numerical simulations verify its validity.Thirdly,how to improve the adaptive network coded cooperation(ANCC)scheme is con-sidered.The ANCC scheme provides a general coded cooperation framework to map the com-munication connections of a wireless network into the coding constraint.However,random selection policy is assumed at each node to select its retrieval subset,which can not avoid ill-ness in code graph at the sink,such as the length-4 cycles and all-zero columns.In this paper,the concept of indicator matrix is proposed to regulate the retrieval subset selection to derive the enhanced ANCC(EANCC)scheme.It is unveiled that the indicator matrix based retrieval subset selection policy can effectively tailor the girth and degree characteristics of the resul-tant code graph.Since the most harmful short cycles in the network coded cooperation can be completely eliminated,the proposed EANCC approach promises better error performance than the existing ANCC scheme and the GANCC scheme.Numerical results show that it performs 13dB better than the ANCC scheme at the bit error rate performance of 10-5 in 7-user network over the Rayleigh block fading channel.Fourthly,the iterative decoding is applied to the coded caching scheme to improve the reliability of the transmission.From channel coding perspective,the coded caching design is reinterpreted.It is shown that,when all the subfiles at the server are corresponded to the system packets,in which the cached and non-cached contents in the placement phase are thought of as the shortened system packets and the punctured system packets,respectively,while those coded contents transmitted in the delivery phase specify the parity packets,the coded caching design can be reformulated as an error control coding problem.The challenges for arbitrary user requests and noncooperative decoding nature at every user will be highlighted to address the design criteria in order to exploit the coding gain residing in the coded caching.The analy-sis unveils that,with some supplementary parity packets included in either the placement phase or the delivery phase,noticeable transmission reliability improvement can be realized.It is shown that the proposed design is able to flexibly fulfill the asymmetric reliable transmission requirement by placing or transmitting some supplementary parity packets only for those users in adverse conditions.It is also shown that the proposed reliable coded caching and the tradi-tional channel coding can be further integrated into the joint network-channel coding(JNCC)framework to fully exploit the benefits from channel coding and the reliable coded caching.In addition,the proposed reliable coded caching can be readily extended to non-uniform popular-ity case.Finally,the iterative decoding is applied to the multi-way relay network(MWRN)to im-prove the reliability of network coded cooperation.In the m-user MWRN,every user is sup-posed to exchange data with all the other m-1 users(full data exchange)by the assistance of the relay,and there is no direct link between any two users.By reformulating the multi-way relaying network in the framework of error control coding,the LDPC based network coded co-operation design issue for MWRNs is presented.On the basis,the requirements,design criteria and practical design were proposed to demonstrate how to devise reliable network coded coop-eration scheme,such that every user is capable of reliably exchanging data with other m-1 users.The analysis shows that,the proposed LDPC based network coded cooperation design exhibits powerful error correction capability,whose resultant parity-check matrix has favorable girth,degree and connectivity properties,especially when compared with the existing schemes.And the overall required number of time slots to accomplish the proposed LDPC based net-work coded cooperation scheme is 2m.And the extension of the proposed design to the JNCC design seems to be a promising solution to the reliable multi-way relaying scheme for adverse channels.