RC-LDPC Codes Construction And Error Flat Analysis In Deep Space Communication

Due to the features of long communication distance, low signal to noise ratio (SNR) of the receiving end and long transmission delay, deep space exploration requires channel coding technology of high performance to provide reliable communications. Low Density Parity Check (LDPC) Code has the advantages of high performance, moderate complexity as well as parallel coding and decoding, therefore it is highly valued in the field of deep space communication. The performance of LDPC code approximates the Shannon limit, but Shannon coding only gives the theoretical structure but does not propose the specific construction method, therefore LDPC coding of excellent performance and low complexity remains an unsolved difficult problem. Error floor is an important issue in the theory and application of LDPC code, represented by the sudden drop of bit error rate curve slope in the medium-high section of SNR range. Deep space communication requires achieving an extremely low bit error rate, and therefore, how to effectively eliminate error floor is of great significance to deep space communication. In the deep space communication channel, multiple sets of coders of different code rates are required to adapt to different states of the channel. Rate Compatible LDPC (RC-LDPC) is an effective way to meet the requirement. Constructing high performance compatible LDPC code has always been an important issue of deep space communication. With focus on the error floor and the construction of RC-LDPC code, this paper studies the following issues based on the description of coding and decoding principle on LDPC code and construction methods:(1) Research on the construction of LDPC code:First, based on the introduction to the LDPC code construction algorithm, a combined PEG-ACE LDPC code construction algorithm is proposed. Construct a basis matrix by PEG algorithm, combined with the improved ACE algorithm to search the circular shift offset, and perform quasi-cyclic extensions on the basis matrix. Simulation indicates that this algorithm has improved compared to the traditional PEG algorithm. Secondly, on the LDPC code of the protograph, an optimized algorithm of protograph LDPC code based on the group replacement is proposed. Replacing the check node of the protograph LDPC code with high performance linear block code not only maintains the excellent performance of the protograph LDPC code, but also constrains the connectivity of the loop, making the error floor lower. Simulation indicates that the LDPC code constructed by this algorithm has better BER performance.(2)LDPC code error floor analysis:First, after analyzing the cause of the error floor which is the trapping set, a LDPC code trapping set elimination algorithm on the basis of graph covers is proposed. The edges of the trapping sets can be damaged by search, and exchange these edges to eliminate the existing trapping sets. Meanwhile in the search and exchange process, depending on the methods of judging edges and freezing edges, avoid generating new trapping set. Simulation indicates that algorithm can eliminate part of the trapping set and reduce LDPC code error floor. Secondly, based on the analysis on the cyclic shift matrix and graph lifting theory, a trapping set elimination algorithm based on the cyclic lifting is proposed. By eliminating trapping sets constituted by the short cycle in the Tanner graph, trapping set with smaller critical value can be eliminated. As the lifting order increases, trapping sets of higher critical value can be eliminated. Simulation verifies cyclic lifting algorithm can eliminate most of the leading trapping set, and thus effectively reduce the LDPC code error floor.(3) Application research on the RC-LDPC code in the deep space communication system: Combined with the characteristics of deep space communication, based on the puncturing algorithm, a RC-LDPC code construction algorithm based on orderly puncturing is propose to achieve a rate compatible LDPC code design. By analyzing the restored tree and optimizing the selection of puncturing nodes, high performance rate compatible LDPC code can be constructed. Simulation verifies that the orderly puncturing algorithm has excellent performance and high bit rate code performance has improved compared with the previous puncturing algorithm with lower error floor, which is suitable for deep space communication system.