Optimized Design Of LDPC Coded BICM Systems

With the great increase in demand for data transmission and the rapid development of in-tegrated circuit technology,the flexible,efficient,and reliable digital communication systems have become mainstream.Most digital communication systems are equipped with coded mo-dulation modules in which the bit-interleaved coded modulation(BICM)technology is most widely applied.The combination of the BICM technology and the popular LDPC codes re-sult in the LDPC-BICM system which is the main concern of this thesis and employed in many communication standards such as WiMAX and WiFi.Moreover,the current 5G standardization meetings have identified LDPC codes as the channel coding schemes for the enhanced mobile broadband scenario and thus the LDPC-BICM system will probably be applied in this 5G sce-nario.The design and optimization of the LDPC-BICM system mainly include three aspects:channel coding,modulation,and the interleaving between coding and modulation.During the last two decades,the technologies of LDPC,modulation,and interleaving keep developing ra-pidly.Thus,the design and optimization of the LDPC-BICM system has been a hot topic in the research community.In this thesis,we mostly use the protograph-based LDPC(P-LDPC)codes for channel coding.The P-LDPC codes are very popular and employed in WiMAX and WiFi standards be-cause of their low memory requirements,high-speed decoding structures,and powerful error-correcting ability.With respect to the LDPC-BICM systems that deploy P-LDPC codes and high-order modulation,the modified PEXIT algorithm is proposed so as to analyze the perfor-mance of such a system.The analysis result is a threshold value and a lower value indicates better performance.Afterwards,we optimize the three aspects of the LDPC-BICM systems,coding,modulation,and interleaving,according to the strategy of minimizing the analysis thres-hold.Compared with conventional designs,the optimized systems can always achieve a signi-ficant gain.In the optimization process,we utilize the differential evolution algorithm to deal with complex optimization problems.The major contributions of this thesis can be summarized as follows:(1)The optimization and design of interleaving.Firstly,we improve the classic VDMM in-terleaving structure and propose a more general GVDMM interleaving structure.The GVDMM structure not only retains the advantages of VDMM such as regular structure and implementa-tion convenience,but also overcomes the main drawback that VDMM has a narrow application scope.Secondly,we make a detailed comparison between our proposed GVDMM structure and the modified VDMM structure presented in other papers.The GVDMM interleaving structure performs much better in most simulation scenarios and the two structures perform similarly in only one exception scenario.Thirdly,regarding the optimization and design of GVDMM inter-leaving,we propose an efficient method which can greatly reduce the complexity of exhaustive search and incurs no performance loss.Moreover,in order to deal with the optimization pro-blem of GVDMM interleaving which might be still very complex after using the above simpli-fied method,we present another optimization method,i.e.,utilizing the differential evolution algorithm to find a well-performed GVDMM-interleaving scheme.Finally,we provide the ap-plication results of the GVDMM interleaving structure in various scenarios.Both threshold analyses and simulated results verify the performance advantage of our proposed GVDMM interleaving structure.(2)The optimization and design of modulation.We apply the strategy of minimizing the analysis threshold in the modulation optimization for visible light communications.In this thesis,we mainly consider the following four visible light communication scenarios:RGB pa-rallel transmission,RGBA parallel transmission,RGB synthesis modulation,and RGBA synt-hesis modulation.In these four scenarios,the proposed optimization results can always obtain a high performance gain in comparison with conventional designs.In addition,we observe point-coincidence phenomenon in the optimization results,i.e.,some points in the constellation coincide into one point.Inspired by such interesting phenomenon,we try to make some points coincide in optimizing the constellations of 8PAM and 16PSK.The corresponding optimization results can significantly reduce the number of constellation points and meanwhile outperform the conventional constellations in the performance.(3)The optimization and design of channel coding.We propose a design method to find good P-LDPC codes with the numbers of nonzero elements in the parity-check matrix and de-coding iterations both strictly limited.Both numerical analyses and simulations verify that the resulting codes exhibit better performance than conventional design results under the limitation of decoding complexity.Furthermore,we utilize this design method in a visible light com-munication system and the resulting code performs better than conventional codes.Lastly,we present an error-correcting coding scheme for data transmission in the enhanced mobile broad-band scenario of 5G.We make a comparison between the coding scheme proposed by Samsung Corporation and ours.The simulated results show that our coding scheme has obvious advan-tage.