On The Decoding And Application Of Low Density Lattice Codes

Due to the poor channel condition in wireless communication,channel coding is required to improve the reliability of the wireless communication system.Lattice codes are Euclidean space analog of linear codes that can achieve the capacity of the additive white Gaussian noise(AWGN)channel.Low density lattice codes(LDLC)are practical lattice codes that can approach the capacity of AWGN channel.Because the codewords of LDLC are real numbers,they are very suitable for the transmission on the actual channel.In this thesis,we mainly focus on the research of LDLC,including the decoding algorithms and the application for two-way relay channel with direct link.The main contents are as follows:First,we have a research on LDLC,including the definition of LDLC,the construction of LDLC and shaping methods.On this basis,the decoding algorithm with quntization and efficient parametric decoding algorithm are deeply studied..Secondly,an improved Gaussian parametric decoding algorithm is proposed based on the existing three/two Gaussian paraamertic decoding algorithm.In three/two Gaussian paramertic decoding algorithm,the growth of the number of Gaussians in each mixture is exponential in forward-backward-type algorithm.In order to reduce the computing complexity,the Gaussian mixture reduction algorithm is applied after each forward and backward recursion step.Simulation results show that the improved Gaussian parametric decoding algorithm has a smaller storage requrements and lower decoding complexity with almost no performance loss for n=100 and n=1000.Finally,we analyze hard decision relaying and soft symbol relaying for two-way relay channel with direct link,and a joint iterative decoding algorithm based on equivalent noise variance combination is proposed.The two-way received messages are decoded with improved Gaussian parametric decoding separately.Then the messages are combined to generate a single message based on equivalent noise variance.The single message is decoded again with improved Gaussian parametric decoding algorithm.Simulation results show that the proposed scheme has a good performance with soft symbol relaying compared with the joint iterative decoding algorithm based on product.Also,the proposed scheme has a lower complexity with soft symbol relaying and hard decision relaying.