Adaptation Technology Based On LDPC Codes

Efficiency and reliability are two permanent problems in communication systems.Channel coding is an effective way to improve system reliability, LDPC codes are aclass of block codes with near Shannon limit performance that can be realized, theperformance of LDPC codes with long code length can be even better than Turbo code.For efficiency, link adaptation is a well-known technology to enhance the spectralefficiency of communication system and has been widely applied in wirelesscommunication. It's meaningful to investigate the link adaptive system based on LDPCcodes which can make full use of spectral resource while maintaining an acceptabletransmission quality. First of all, we had a thorough research on the construction andtheoretical analysis of LDPC codes, including the construction of parity-check matrix,external information transfer theory and LDPC coded modulation etc., based on theabove, in order to improve the efficiency and reliability at the same time, we combinedLDPC codes with link adaptation technique, and then expanded it from SISO system toMIMO channel. The main work and the innovations are as the following:1,We proposed a new puncturing pattern to generate rate-compatible LDPCcodes. It was done from two aspects: In one respect, we studied various methods toconstructing LDPC codes, including the original method Gallger proposed in 1962,MacKay's method which reinvented LDPC codes 30 years later and PEG algorithmproposed by Hu. Through comparing the above three methods, we chose PEGmethod as our research basis because it is the best one constructing LDPC codeswith moderate code length which not only has better performance than MacKay'scodes but also can realize linear-time encoding. In another respect, we investigatedthe constructing methods of rate-compatible LDPC codes, including randompuncturing and optimal puncturing, we found that for irregular LDPC codes,puncturing variable nodes with lower degree had less impact on the parity checkmatrix, i.e. the configuration of the mother code, and in a weight-increasing parity-check matrix generated by PEG algorithm, variable nodes were arranged according to their degrees, so we proposed a new continuous puncturing which performed similar as random puncturing but more applicable.2 We proposed an improved algorithm to optimize the degree of irregular LDPC codes. EXIT chart was an effective method to analysis the decoding threshold of LDPC codes and optimize the degree distribution, which was simpler than Gaussian approximation. The LDPC decoder could be divided into variable node decoder and check node decoder, and the iterative decoding process could be regarded as the external information transfer between the two component decoders. The decoding threshold could be decided and the degree distribution could be optimized through comparing the output/input of the two decoders. The degrees of check nodes were usually fixed when we optimized the degree distribution of variable nodes, we proposed to consider the check nodes degree when design the variable nodes degree, which could generate LDPC codes with near Shannon limit performance.3 We proposed a new modulation strategy for LDPC coded modulation system, which was suitable for a generalized unequal error protection (UEP). EXIT charts were used to explain the advantages of our scheme and the new mapping can easily be extended to higher code rate. Irregular LDPC codes possessed inherent UEP property that bits with high degree could be more likely decoded and performed better than those with low degree. The UEP could be further enhanced by mapping the coded bits of an irregular LDPC code onto a modulation signal set. In communication system, systematic bits carried far more information than check bits, so we proposed to give them different protection by mapping them to different modulation signal sets. Simulation results showed that compared with the traditional modulation scheme with the same spectral efficiency, the BER performance of our new scheme was improved greatly. Based on computer simulation, we used EXIT chart to give a theoretical analysis from the view of information transfer. And at the same time, we foundd that for regular LDPC codes, the same decoding threshold doesn't mean the same BER performance, and we should also consider the transfer curve of the systematic bits of a codeword.4,We proposed an adaptive modulation and coding system based on LDPC codes, find the way to incorporate the codes in the AMC system, and investigated the impact of channel estimation error on the performance of the whole system. We only considered some adaptive technologies in the physical layer in the above work, but an implied link adaptive technology—HARQ was neglected. In order to optimize the whole communication system through cross-layer design, we proposed an AMC-HARQ system based on rate-compatible LDPC codes, combining the AMC in physical layer and the HARQ in the data link layer. We could choose different transmission schemes roughly by AMC and adapted the scheme deliberately by HARQ. Consequently, we extended the system from SISO channel to MIMO system combined with STBC, the impact of different parameters (including the number of transmit and receive antennas,channel fading parameters,channel estimation error) on the system are investigated thoroughly, through the above research, we had a pilot study of cross-layer design.5,We proposed an adaptive STBC system based on LDPC codes, the code rate and the modulation scheme were fixed, different power were allocated to different transmit antennas according to the current channel conditions, i.e. power control. Firstly, we illustrated the encoding and decoding principle of Alamouti codes systematically, and then demonstrated the principle of adaptive STBC. We allocated the transmit power to each transmit antenna with the help of the weight functions, in a way that should maximize the receiver performance. Then we applied LDPC codes into this adaptive transmit diversity technique under Rayleigh fading channel, the soft output of STBC decoder was the input of the LDPC decoder, and compared the performance of adaptive LDPC-STBC with that of adaptive turbo-STBC.