Research On Polar Code And Its MBM-concatenated Technology

Channel coding is one of the most important technologies in wireless communication systems to combat interference and fading.Polar codes are a newly proposed channel coding which can theoretically reach the Shannon limit with low complexity of encoding and decoding,and thus can be widely applied.As a new type of modulation,Media-Based Modulation(MBM)is based on embedding part of information in the channel states,so as to improve the efficiency of data transmission and enhance security at the physical layer.The concatenation of polar codes and MBM can improve the system performance with low complexity.In this thesis,we mainly study the decoding algorithm of polar codes,the concatenated scheme of polar codes and MBM as well as the detection and decoding of concatenated systems.Firstly,the principle and characteristics of polar codes and MBM are introduced.For polar codes,polarization phenomenon is presented and the density evolution and Gauss approximation are used to measure the channel reliability.Then,the encoding and decoding algorithm of polar codes are described.In addition,the performance of successive cancellation(SC)algorithm and belief propagation(BP)algorithm are compared.As to MBM,we introduce the principle of MBM,apply it to the multiple-input multiple-output(MIMO)system,and compare the performance of different multi-antenna schemes with the same transmission rate.Then,we propose an improved double thresholding strategy to prune paths and a new decoding algorithm called dynamic path extention SCL(DPE-SCL)algorithm.The proposed pruning strategy changes the previous threshold and reduces the complexity from O(L log L)to O(L).DPE-SCL algorithm relies on the splitting rule and pruning strategy.This decoding algorithm does not restrict the candidate paths to a fixed number,but allows its dynamic change.The number of paths is controlled by the splitting rule and the pruning strategy.Simulation results show that our DPE-SCL algorithm has better performance in low signal to noise ratio(SNR)at the expense of the increased number of paths while the performance is poor in high SNR with lower complexity.Finally,the concatenated scheme of polar codes and MBM called polar coded media-bsed modulation(PC-MBM)is studied.The novelty is that the frozen bits in polar codes are embedded in channel states(MBM message)while the free bits are mapped to source-based modultaion(SBM)message.As the frozen bits are preset sequence,the channel index can be known at the receiver with no error and be applied to SBM detection directly.Therefore,the detection complexity will decrease significantly.Then we transform the PC-MBM channel into parallel independent sub-channels,which is used to construct the polar code by computing the channel capacity of sub-channels.At last,two kinds of low-complexity maximum likelihood(ML)detection methods based on soft-decision algorithm are proposed.Firstly,the search space is limited to a certain extent by using the idea of sphere decoding.Secondly,we propose a region-restricted ML detection method,which can further decrease the search space and reduce the computation complexity.