可Zigzag解码的前向纠删码编码方法研究

发布时间：2018-03-17 18:39

本文选题：前向纠删码　切入点：可Zigzag解码　出处：《浙江大学》2016年硕士论文　论文类型：学位论文

【摘要】：随着网络技术的飞速发展,接入网络的用户越来越多,基于网络的多媒体应用也越来越丰富多彩。急速增长的网络流量和多样化的业务需求对通信系统的有效性和可靠性要求越来越高。由于网络拥塞、信道衰落等因素影响,网络中的数据传输不可避免地会遇到数据丢包问题。前向纠删编码技术是解决网络丢包的有效手段之而一种实用的前向纠删编码方法需在编解码复杂度与前向纠删性能之间权衡。2013年提出的可Zigzag解码的前向纠删码具有很低的编解码复杂度和最大距离可分特性,但其冗余编码数据包比原始数据包略长。本文以可Zigzag解码的前向纠删编码方法展开研究工作。论文提出了一种基于有限域GF(qp)(q为素数,p≥1)上柯西矩阵的编码系数矩阵构造方法,并证明了根据这种编码系数矩阵构造的编码数据包满足可Zigzag解码的条件。在所提出编码系数矩阵构造方法中,先构造有限域上的柯西矩阵,然后将矩阵元素用本原元表示法表示,并把本原元符号看成编码偏移符号,得到可Zigzag解码的编码系数矩阵。基于GF(qp)上柯西矩阵的构造方法巧妙地将高阶有限域上面向数据包的乘法和加法运算转化为数据包的移位和异或运算,生成的编码数据包不仅具有最大距离可分特性,而且冗余度较小。仿真结果表明,与现有的编码系数矩阵构造方法相比,在生成相同数量的冗余编码数据包时,所提出的基于GF(qp)上柯西矩阵的编码系数矩阵构造方法生成的冗余编码数据包的冗余度更小。针对时变除删信道环境,结合可Zigzag解码的前向纠删码的性质,论文提出了一种可Zigzag解码的无速率码的编码系数矩阵构造方法。无速率码编码时,在一定的范围内随机地选取编码偏移量组成编码系数矢量,编码端根据编码系数矢量将原始数据包做异或运算可以无限地产生新的编码数据包。仿真结果表明,当编码偏移量选取范围的最大值等于254比特(约32字节)时,论文所提出的可Zigzag解码的无速率码的编码数据包具有最大距离可分性质的概率约为99.6%。可Zigzag解码的前向纠删码具有达到前向纠删编码性能限和很低的编解码复杂度两大优点,在网络传输、分布式存储等领域有广泛的应用前景。
[Abstract]:With the rapid development of network technology, more and more users are accessing the network. The multimedia applications based on the network are becoming more and more colorful. The rapid growth of network traffic and diversified service requirements are increasingly demanding the efficiency and reliability of the communication system. Due to network congestion, channel fading and other factors, Data transmission in the network will inevitably encounter the problem of data packet loss. Forward erasure coding is an effective method to solve the problem of packet loss in network. A practical forward erasure coding method should be used in coding and decoding complexity and forward correction. The forward erasure code proposed in 2013, which can be decoded by Zigzag, has the characteristics of low complexity and maximum distance separability. However, the redundant encoded data packet is a little longer than the original packet. In this paper, the forward erasure coding method which can be decoded by Zigzag is used. In this paper, a method of constructing the coding coefficient matrix based on Cauchy matrix over finite field GF(qp)(q is proposed. It is proved that the coded data packets constructed according to the encoding coefficient matrix satisfy the Zigzag decoding condition. In the proposed method of constructing the encoding coefficient matrix, the Cauchy matrix over the finite field is constructed first. Then the matrix elements are represented by primitive representation, and the primitive symbols are regarded as coded offset symbols. The construction method of Cauchy matrix based on Zigzag decode is used to subtly convert the multiplication and addition operations for data packets on high order finite fields into the shift and XOR operations of data packets. The generated coded packets not only have the characteristics of maximum distance separability, but also low redundancy. The simulation results show that, compared with the existing coding coefficient matrix construction methods, when generating the same number of redundant coded packets, The redundancy of the redundant coded data packets generated by the method of constructing the coding coefficient matrix of the Cauchy matrix based on GF-QP) is less. For the time-varying divide-delete channel environment, combining with the properties of forward erasure codes that can be decoded by Zigzag, In this paper, a method of constructing the coding coefficient matrix of rate free code with Zigzag decode is proposed. In the coding of rate code, the encoding offset is randomly selected to form the encoding coefficient vector in a certain range. According to the encoding coefficient vector, the encoding end can generate the new coded data packet infinitely by using the XOR operation of the original data packet. The simulation results show that when the maximum value of the encoding offset is equal to 254 bits (about 32 bytes), In this paper, the probability of maximum distance separability of Zigzag decoded data packets without rate codes is about 99.6. The forward erasure codes that can be decoded by Zigzag have the advantages of achieving the performance limit of forward erasure coding and low encoding and decoding complexity. In network transmission, distributed storage and other fields have a wide range of applications.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.2

【相似文献】