技术领域technical field
本发明属于通信领域，尤其涉及一种基于解码转发协议的误码率近似最优解码的获取方法及系统。The invention belongs to the field of communications, and in particular relates to a method and a system for obtaining an approximate optimal decoding of a bit error rate based on a decoding and forwarding protocol.
背景技术Background technique
作为一种杰出的传输策略，中继辅助通信成为了研究热点，其基本思想是部署一个或多个中继节点来扩展源节点发出信号的覆盖范围。在众多已知的中继协议中，研究最广泛的是放大转发(Amplify-and forward,AF)与解码转发(Decode-and-Forward,DF)协议。作为消除噪声和其他信道受损影响的最经典和实用的中继协议之一，DF有重要的研究意义。在此协议中，中继节点完全解码、重新编码并重新传输源节点的信息。然而，传统的DF协议的实现面临两个问题。As an outstanding transmission strategy, relay-assisted communication has become a research hotspot. Its basic idea is to deploy one or more relay nodes to expand the coverage of the signal sent by the source node. Among the many known relay protocols, the most widely studied are the Amplify-and-forward (AF) and Decode-and-Forward (DF) protocols. As one of the most classical and practical relay protocols to eliminate the effects of noise and other channel impairments, DF has important research significance. In this protocol, the relay node fully decodes, re-encodes and retransmits the source node's information. However, the implementation of the traditional DF protocol faces two problems.
其一，传统的应用在目标节点上的最大比例组合方案假设从源节点至中继节点的传输是完善的，而实际上从源节点至中继节点的链路并不完善，这意味着中继节点有时并不能成功解码。其二，实现DF协议的另一个难点是很难(不可能)在目标节点实现BER最优解码算法。首先，要实现最优解码，必须要获取从源节点至中继节点链路的精确的误差统计信息，然而这在实际系统中很难做到。其次，同最大似然解码(Maximum LikelihoodDecoding,MLD)算法一样，在目标节点的BER最优解码算法不仅需要在目标节点，还要在中继节点上考虑所有可能的符号检测场景。此外，已知的BER/BLER最优解码算法的复杂性与信息快的长度指数相关，并且即便是对于一些未编码的系统，这些算法的实现也非常复杂。First, the traditional maximum ratio combination scheme applied to the target node assumes that the transmission from the source node to the relay node is perfect, but in fact the link from the source node to the relay node is not perfect, which means that the transmission from the source node to the relay node is perfect. Successor nodes sometimes fail to decode successfully. Second, another difficulty in implementing the DF protocol is that it is difficult (impossible) to implement the BER optimal decoding algorithm at the target node. First, to achieve optimal decoding, it is necessary to obtain accurate error statistics of the link from the source node to the relay node, which is difficult to achieve in practical systems. Secondly, like the Maximum Likelihood Decoding (MLD) algorithm, the BER optimal decoding algorithm at the target node needs to consider all possible symbol detection scenarios not only at the target node but also at the relay node. In addition, the complexity of the known BER/BLER optimal decoding algorithms is exponentially related to the length of the information, and the implementation of these algorithms is very complicated even for some uncoded systems.
在更普遍的基于DF协议的单源单终端的复杂中继网络中，每个中继节点都有单输入链路和多输出链路，而目标节点是唯一的具有多输入链路的节点。此外，每个物理上具有多输入链路的中继节点可以在逻辑上表示为具有单输入链路的多个中继节点。虽然实际的网络中的中继节点可能有多个输入链路，但我们的网络图实际上是关于源节点发送的数据包是如何被接收，并由逻辑上的中继节点重新生成再发送给目标节点的图形表示。在此期间，一个实际的具有多输入链路的中继节点可以被表示为具有一个输入链路的多个逻辑中继节点。In the more general single-source single-terminal complex relay network based on the DF protocol, each relay node has a single input link and multiple output links, and the target node is the only node with multiple input links. Furthermore, each relay node with multiple input links physically can be represented logically as multiple relay nodes with a single input link. Although a relay node in an actual network may have multiple input links, our network diagram is actually about how a packet sent by a source node is received, regenerated by a logical relay node and sent to A graphical representation of the target node. During this time, an actual relay node with multiple input links can be represented as multiple logical relay nodes with one input link.
因此，在不完善的源节点至中继节点的链路中，现有的近似最优的解码算法在估计不同中继解码场景的概率的能力较低。Therefore, in imperfect source-to-relay links, existing near-optimal decoding algorithms are less capable of estimating the probabilities of different relay decoding scenarios.
发明内容SUMMARY OF THE INVENTION
本发明所要解决的技术问题在于提供一种基于解码转发协议的误码率近似最优解码的获取方法及系统，旨在解决现有的近似最优的解码算法在估计不同中继解码场景的概率的能力较低的问题。The technical problem to be solved by the present invention is to provide a method and system for obtaining an approximate optimal decoding of bit error rate based on the decoding and forwarding protocol, which aims to solve the problem that the existing approximate optimal decoding algorithm can estimate the probability of different relay decoding scenarios. problems of lower capacity.
本发明是这样实现的，一种误码率近似最优解码的获取方法，包括：The present invention is achieved in this way, a method for obtaining approximately optimal decoding of bit error rate, comprising:
步骤A，建立基于解码转发协议、单源单终端的中继网络模型，所述中继网络模型包括单个源节点、若干中继节点和单个目标节点，每一中继节点均由单输入链路和多输出链路，且所述目标节点为唯一具有多输入链路的节点；Step A, establish a relay network model based on the decoding and forwarding protocol, single source and single terminal, the relay network model includes a single source node, several relay nodes and a single target node, and each relay node is composed of a single input link. and multiple output links, and the target node is the only node with multiple input links;
步骤B，在所述目标节点建立乘积网格结构，所述乘积网格结构表示所有可能的中继节点解码错误的场景，并对中继节点解码所引入的错误比特之间的相关性进行准确建模；In step B, a product grid structure is established at the target node, and the product grid structure represents all possible relay node decoding error scenarios, and the correlation between the error bits introduced by the relay node decoding is accurately carried out. modeling;
步骤C，根据所述乘积网格结构确定分支度量；Step C, determining branch metrics according to the product grid structure;
步骤D，根据确定分支度量的乘积网格结构和BCJR算法对源节点传输的源信息进行计算，得到所述源信息的误码率近似最优解码。In step D, the source information transmitted by the source node is calculated according to the product trellis structure of the determined branch metric and the BCJR algorithm, and the approximate optimal decoding of the bit error rate of the source information is obtained.
本发明还提供了一种误码率近似最优解码的获取系统，包括：The present invention also provides an acquisition system for approximately optimal decoding of the bit error rate, comprising:
模型构建单元，用于建立基于解码转发协议、单源单终端的中继网络模型，所述中继网络模型包括单个源节点、若干中继节点和单个目标节点，每一中继节点均由单输入链路和多输出链路，且所述目标节点为唯一具有多输入链路的节点；The model building unit is used to build a relay network model based on the decoding and forwarding protocol, single source and single terminal, the relay network model includes a single source node, several relay nodes and a single target node, each relay node is composed of a single source node. input links and multiple output links, and the target node is the only node with multiple input links;
网络构建单元，用于在所述目标节点建立乘积网格结构，所述乘积网格结构表示所有可能的中继节点解码错误的场景，并对中继节点解码所引入的错误比特之间的相关性进行准确建模；A network construction unit for establishing a product grid structure at the target node, the product grid structure representing all possible relay node decoding error scenarios, and for the correlation between the error bits introduced by the relay node decoding sex for accurate modeling;
度量确定单元，用于根据所述乘积网格结构确定分支度量；a metric determining unit, configured to determine a branch metric according to the product grid structure;
信息计算单元，用于根据确定分支度量的乘积网格结构和BCJR算法对源节点传输的源信息进行计算，得到所述源信息的误码率近似最优解码。The information calculation unit is configured to calculate the source information transmitted by the source node according to the product grid structure of the determined branch metric and the BCJR algorithm, and obtain approximately optimal decoding of the bit error rate of the source information.
本发明与现有技术相比，有益效果在于：本发明实施例提供一种基于解码转发协议的在单源单终端的复杂中继网络中的卷积码编码的误码率近似最优解码的获取方法，通过建立基于DF协议的单源单终端的复杂中继网络模型，在目标节点处建立一种新的、合理定义了分支度量的乘积网格结构，该乘积网格结构能够准确地表示所有可能的中继节点解码错误的场景，在此乘积网格结构中BCJR算法对源节点传输的源信息进行计算，得到所述源信息的误码率近似最优解码。通过本发明实施例提供的基于DF协议的卷积码编码的单源单终端的复杂中继网络的误码率近似最优解码的获取方法可以在不同的中继信道场景中实现比现有的解码算法更优的性能。Compared with the prior art, the present invention has the beneficial effects that: the embodiment of the present invention provides a decoding and forwarding protocol-based method for approximate optimal decoding of the bit error rate of convolutional code encoding in a complex relay network with single source and single terminal. The acquisition method is based on the establishment of a complex relay network model with single source and single terminal based on the DF protocol, and a new product grid structure with reasonably defined branch metrics is established at the target node. The product grid structure can accurately represent In all possible scenarios of relay node decoding errors, in this product grid structure, the BCJR algorithm calculates the source information transmitted by the source node, and obtains the approximate optimal decoding of the bit error rate of the source information. The method for obtaining the approximate optimal decoding of the bit error rate of the complex relay network with single source and single terminal based on the convolutional code encoding of the DF protocol provided by the embodiment of the present invention can achieve better decoding than the existing ones in different relay channel scenarios. Better performance of the decoding algorithm.
附图说明Description of drawings
图1是本发明实施例提供的一种误码率近似最优解码的获取方法的流程图；1 is a flowchart of a method for obtaining approximately optimal decoding of a bit error rate provided by an embodiment of the present invention;
图2是本发明实施例提供的基于DF协议的单源单终端的中继网络模型的结构示意图；2 is a schematic structural diagram of a single-source single-terminal relay network model based on a DF protocol provided by an embodiment of the present invention;
图3是本发明实施例提供的NBOD算法的流程图；3 is a flowchart of an NBOD algorithm provided by an embodiment of the present invention;
图4是本发明实施例提供的NBOD算法实例中基于DF协议的中继网络模型的结构示意图；4 is a schematic structural diagram of a relay network model based on the DF protocol in an example of an NBOD algorithm provided by an embodiment of the present invention;
图5是本发明实施例提供的基于DF协议的单源单终端的中继网络模型的5节点模型示意图；5 is a schematic diagram of a 5-node model of a single-source single-terminal relay network model based on a DF protocol provided by an embodiment of the present invention;
图6是本发明实施例提供的NBOD算法(T＝1)与现有技术提供的MRC，C-MRC以及SDF算法与在图5提供的5节点模型的性能比较示意图；6 is a schematic diagram showing the performance comparison between the NBOD algorithm (T=1) provided by the embodiment of the present invention and the MRC, C-MRC and SDF algorithms provided by the prior art and the 5-node model provided in FIG. 5;
图7是本发明实施例提供的一种误码率近似最优解码的获取系统的结构示意图。FIG. 7 is a schematic structural diagram of an acquisition system for approximate optimal decoding of a bit error rate according to an embodiment of the present invention.
具体实施方式Detailed ways
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.
本发明实施例提供了如图1所示的一种误码率近似最优解码的获取方法，包括：An embodiment of the present invention provides a method for obtaining approximate optimal decoding of a bit error rate as shown in FIG. 1 , including:
S101，建立基于解码转发协议、单源单终端的中继网络模型，所述中继网络模型包括单个源节点、若干中继节点和单个目标节点，每一中继节点均由单输入链路和多输出链路，且所述目标节点为唯一具有多输入链路的节点；S101, establish a relay network model based on a decoding and forwarding protocol, a single source and a single terminal, the relay network model includes a single source node, several relay nodes and a single target node, and each relay node is composed of a single input link and a single target node. Multiple output links, and the target node is the only node with multiple input links;
S102，在所述目标节点建立乘积网格结构，所述乘积网格结构表示所有可能的中继节点解码错误的场景，并对中继节点解码所引入的错误比特之间的相关性进行准确建模；S102, establishing a product grid structure at the target node, where the product grid structure represents all possible decoding error scenarios of the relay node, and accurately establishes the correlation between the error bits introduced by the relay node decoding mold;
S103，根据所述乘积网格结构确定分支度量；S103, determining branch metrics according to the product grid structure;
S104，根据确定分支度量的乘积网格结构和BCJR算法对源节点传输的源信息进行计算，得到所述源信息的误码率近似最优解码。S104: Calculate the source information transmitted by the source node according to the product trellis structure of the determined branch metric and the BCJR algorithm to obtain approximately optimal decoding of the bit error rate of the source information.
下面，结合图2至图6对本发明实施例进行进一步地阐述：Hereinafter, the embodiments of the present invention will be further elaborated with reference to FIGS. 2 to 6 :
本发明实施例提供了一种基于DF协议的，在单源单终端的复杂中继网络中的卷积码编码的误码率(Bit error rate，BER)近似最优解码的获取方法，包括如下步骤：The embodiment of the present invention provides a method for obtaining approximately optimal decoding of the bit error rate (Bit error rate, BER) of convolutional code coding in a complex relay network with single source and single terminal based on the DF protocol, including the following step:
步骤1：建立一种基于DF协议的，单源单终端的中继网络模型；Step 1: establish a relay network model based on DF protocol, single source and single terminal;
其中，本步骤中建立的中继网络模型是一种更普遍的单源单终端的复杂中继网络模型，图2为基于DF协议的单源单终端的中继网络模型的一般结构。在此中继网络模型中，基于DF协议，每个中继节点都有单输入链路和多输出链路，而目标节点是唯一的具有多输入链路的节点。此外，每个物理上具有多输入链路的中继节点可以在逻辑上表示为具有单输入链路的多个中继节点。The relay network model established in this step is a more general single-source and single-terminal complex relay network model, and FIG. 2 shows the general structure of the single-source and single-terminal relay network model based on the DF protocol. In this relay network model, based on the DF protocol, each relay node has a single input link and multiple output links, and the target node is the only node with multiple input links. Furthermore, each relay node with multiple input links physically can be represented logically as multiple relay nodes with a single input link.
需要说明的是，一般的基于DF协议的多跳多分支中继网络通常被认为是上述网络的一个特例，因为包含所有从源节点至中继节点链路的路径是唯一的，所以源节点和所有的中继节点以及它们之间的链路组成了一个以源节点作为根节点，以中继节点作为叶节点，通过其外向链路连接到目标节点的树结构。It should be noted that the general multi-hop multi-branch relay network based on the DF protocol is usually regarded as a special case of the above network, because the path including all links from the source node to the relay node is unique, so the source node and the relay node are unique. All relay nodes and the links between them form a tree structure with the source node as the root node and the relay node as the leaf node, which are connected to the target node through their outgoing links.
其中，从源节点S发出的信息可以通过不同路径到达目标节点D，即：从源节点S直接到目标节点D；从源节点S到中继节点R，再到目标节点D。在本实施例中，假设三条链路(源节点至目标节点链路，源节点至中继节点链路与中继节点至目标节点链路)都是分别具有信噪比(signal-to-noise ratios,SNRs)γsd,γsr和γrd的二进制输入无记忆通道。The information sent from the source node S can reach the target node D through different paths, namely: from the source node S directly to the target node D; from the source node S to the relay node R, and then to the target node D. In this embodiment, it is assumed that the three links (the link from the source node to the target node, the link from the source node to the relay node and the link from the relay node to the target node) have signal-to-noise ratios respectively. ratios, SNRs) γsd , γsr and γrd for binary input memoryless channels.
需要说明的是，γij，i，j∈{s，r，d}，可以定义为：It should be noted that γij , i, j∈{s, r, d} can be defined as:
其中，hij是i节点和j节点之间的信道增益系数，Eb/N0表示信息比特与高斯噪声功率比。Among them, hij is the channel gain coefficient between node i and node j, and Eb /N0 represents the ratio of information bits to Gaussian noise power.
需要说明的是，为上述树结构的所有叶节点的集合。对于其任意的节点v，使用和分别表示其父节点、子节点和以其为根的子树的所有节点的集合。此外，使用|·|表示一个集合的基数。It should be noted, is the set of all leaf nodes of the above tree structure. For any of its nodes v, use and Represents the set of all nodes of its parent node, child node, and subtree rooted with it, respectively. Also, use |·| to denote the cardinality of a set.
假设上述中继网络模型的所有节点为半双工模式，即不能在同一时间传输和接收数据。因此，本发明实施例提供的中继网络模型在每次通信包含有两个不同的时隙。在时隙1中，源节点传输的调制符号为Xs。由于无线通信具有广播的性质，中继节点与目标节点分别用ysr与ysd表示监听到的带有噪声的Xs的观测。It is assumed that all nodes of the above relay network model are in half-duplex mode, that is, they cannot transmit and receive data at the same time. Therefore, the relay network model provided by the embodiment of the present invention includes two different time slots in each communication. In slot 1, the modulation symbol transmitted by the source node is Xs . Due to the broadcast nature of wireless communication, the relay node and the target node use ysr and ysd to denote the observed observations of Xs with noise, respectively.
ysd＝hsdxs+zsdysd =hsd xs +zsd
ysr＝hsrxs+zsrysr =hsr xs +zsr
其中，zsr和zsd是双边功率谱密度为N0/2每维度的零均值高斯噪声。where zsr and zsd are zero-mean Gaussian noise with bilateral power spectral density N0 /2 per dimension.
中继节点对接收到的ysr进行解码并生成cr作为源节点传输码字cs的估计。为了不失一般性，假设使用了二进制相移键控(Binary Phase Shift Keying,BPSK)调制器且所有的传输符号具有标准单位功率。在本实施例中，Xr由cr通过BPSK调制器产生。由于在中继节点处可能存在不成功的解码过程，Xr不一定同于Xs。在时隙2中，中继节点将Xr传送至目标节点，用yrd表示目标节点接收到的具有噪声的xr，则：The relay node decodes the received ysr and generatescr as an estimate of the source node's transmission codeword cs . Without loss of generality, it is assumed that a Binary Phase Shift Keying (BPSK) modulator is used and that all transmitted symbols have standard unit power. In this embodiment, Xr is generated bycr through a BPSK modulator. Since there may be an unsuccessful decoding process at the relay node,Xr is not necessarily the same asXs . In time slot 2, the relay node transmits Xr to the target node, and yrd represents the noisy xr received by the target node, then:
yrd＝hrdxr+zrdyrd =hrd xr +zrd
其中，zrd是双边功率谱密度为N0/2每维度的零均值高斯噪声。基于上述的ysd与yrd，目标节点能够通过特定的解码方案恢复源信息。where zrd is the zero-mean Gaussian noise per dimension with a bilateral power spectral density of N0 /2. Based on the above ysd and yrd , the target node can recover the source information through a specific decoding scheme.
需要说明的是，本发明实施例提供的中继网络模型中总共有G层的中继节点，且在第g层有Mg个中继节点，使用Ri，j来表示在第i个级别的第j个中继节点。此外，源节点和所有的中继节点通过正交信道传输符号，在该中继网络模型中使用BER-最优解码算法进行信息序列us的第i位的解码规则如下：It should be noted that the relay network model provided by the embodiment of the present invention has a total of G-layer relay nodes, and there are Mg relay nodes in the g-th layer, and Ri,j is used to represent the i-th level. The jth relay node of . In addition, the source node and all relay nodes transmit symbols through orthogonal channels. In this relay network model, the BER-optimal decoding algorithm is used to decode thei -th bit of the information sequence us as follows:
其中，ysd和表示目标节点从S-to-D(源节点至目标节点)链路和(中继节点至目标节点)链路所接收的符号。表示中继节点对应的所有叶子节点的集合，向目标节点传输的符号。where, ysd and Represents the destination node from S-to-D (source node to destination node) link and Symbols received by the link (relay node to destination node). Represents all leaf nodes corresponding to the relay node A collection of symbols to transmit to the destination node.
步骤2：在目标节点处建立一种新的乘积网格结构。其中，该网格能准确地表示所有可能的中继节点解码错误的场景，并对中继节点解码所引入的错误比特之间的相关性进行准确地建模；Step 2: Establish a new product grid structure at the target node. Among them, the grid can accurately represent all possible relay node decoding error scenarios, and accurately model the correlation between the error bits introduced by relay node decoding;
具体的，本发明实施例建立一种乘积网格结构用来表示中继节点的解码错误率，并依据合理定义的该乘积网格结构的分支度量结构计算出xs与xr之间的成对错误概率(Pairwise Error Probability,PEP)，作为中继节点发送xr至目标节点，而源节点发送xs的条件概率Pr(xr|xs)的近似。Specifically, the embodiment of the present invention establishes a product grid structure to represent the decoding error rate of the relay node, and calculates the ratio between xs and xr according to the reasonably defined branch metric structure of the product grid structure. For Pairwise Error Probability (PEP), an approximation of the conditional probability Pr(x r| xs ) of the source node sending xs to the target node as a relay node.
具体地，本实施例中引入网格误差表示来描述从xs到xr的可能误差。由于xs与xr之间的误差向量也是一个有效的码字，因此与卷积码相同结构的乘积网格结构可以有效表示所有可能的误差码字，且能精确反映错误信息的相关性。此外，针对条件概率Pr(xr|xs)，构建一个结合xs与xr的乘积网格，将Pr(xr|xs)的计算与乘积网格结构的路径度量相关联。Specifically, a grid error representation is introduced in this embodiment to describe the possible errors from xs to xr . Since the error vector between xs and xr is also an effective codeword, the product trellis structure with the same structure as the convolutional code can effectively represent all possible error codewords, and can accurately reflect the correlation of error information. Furthermore, for the conditional probability Pr (xr |xs ), a product grid combining xs and xr is constructed, and the computation of Pr(xr |xs ) is associated with the path metrics of the product grid structure.
需要说明的是，此乘积网格结构表示可以使NBOD算法有效地解决实现最优译码的两个难点。其一，能够有效计算出每对xs与xr之间的Pr(xr|xs)的PEP近似；其二，基于乘积网格结构，可以实现改进的本领域公开的BCJR算法，通过检查所有可能的(xr，xs)组合得到本领域已知的BER最优解码算法所需要的解码规则并以此来计算源节点的二进制消息包us。It should be noted that this product trellis structure representation can make the NBOD algorithm effectively solve the two difficulties in realizing optimal decoding. First, the PEP approximation of Pr(xr |xs ) between each pair of xs and xr can be effectively calculated; second, based on the product grid structure, the improved BCJR algorithm disclosed in the art can be implemented by Checking all possible (xr , xs ) combinations yields the decoding rules required by BER-optimal decoding algorithms known in the art And use this to calculate the binary message packet us of the source node.
其中，源节点发送Xs,各中继节点分别发送的条件概率可以根据如下公式进行递归分解：Among them, the source node sends Xs , and each relay node sends The conditional probability of It can be recursively decomposed according to the following formula:
其中，Xs表示源节点传输的调制符号，其包含和S1\S2(属于集合S1但不属于集合S2的元素)的联合随机变量。步骤(a)应用了由于链路之间相互独立的噪声所引起的，不同子树中的节点的传输符号是条件独立的性质。步骤(b)遵循了与给出的xs之间独立的事实。随后的步骤以递归方式将类似的参数应用到树的第二层和较低层次上。最后，再考虑了上述树所有的G层后，被分解为在上述乘积网格结构中与所有的链路(除了与目标节点的输入链路)相关的错误概率的乘积。本段阐述了条件概率的计算方法，其中，步骤(a)对应上述公式中的等式步骤(b)对应上述公式中的等式随后的步骤指的是除了等式(a)和等式(b)之后的每一步等式。where Xs represents the modulation symbol transmitted by the source node, which contains and S1 \S2 (elements that belong to set S1 but not to set S2 ) are joint random variables. Step (a) applies the property that the transmitted symbols of nodes in different subtrees are conditionally independent due to the noises between the links being independent of each other. Step (b) follows with given The fact that the xs are independent. Subsequent steps recursively apply similar parameters to the second and lower levels of the tree. Finally, after considering all the G layers of the above tree, is decomposed as the product of the error probabilities associated with all links (except the input link to the target node) in the above-mentioned product trellis structure. This paragraph describes the calculation method of conditional probability, where step (a) corresponds to the equation in the above formula Step (b) corresponds to the equation in the above formula Subsequent steps refer to each step of equations except equation (a) and equation (b).
需要说明的是，可以被近似为PEP，并进一步近似为指数函数的和。当T＝1，也就是使用一个指数函数来近似Q-函数时，可以得到如下：It should be noted, can be approximated as PEP and further as a sum of exponential functions. When T=1, that is, using an exponential function to approximate the Q-function, the following can be obtained:
其中，C是一个常数，dH(·)是计算的位置的数目的函数，是链路的SNR。where C is a constant and dH ( ) is the calculation a function of the number of positions, Yes SNR of the link.
步骤3：合理定义基于步骤2所述的乘积网格结构的分支度量；Step 3: Reasonably define the branch metrics based on the product grid structure described in Step 2;
需要说明的是，上述的条件概率中的每一项可以基于所提出的乘积网格结构表示来计算。由于在此乘积网格结构中总共有个中继节点，通过在目标节点建立一个能够包含Xs和所有的的(Ψ+1)阶的乘积网格结构来应用上述的NBOD算法。(Ψ+1)阶的乘积网格结构上的每条路径都与(Ψ+1)的元祖相关联。(Ψ+1)阶的乘积网格结构上的第一部分路径度量可以用如下表示：It should be noted that each of the above conditional probabilities can be calculated based on the proposed product grid structure representation. Since in this product grid structure there is a total of relay nodes, by establishing a relay node at the target node that can containXs and all (Ψ+1) order product grid structure to apply the above NBOD algorithm. Each path on the product grid structure of order (Ψ+1) is associated with the tuple of (Ψ+1) Associated. The first part of the path metric on the product grid structure of order (Ψ+1) can be expressed as:
其中，考虑所有的链路(从所有的叶子中继节点至目标节点)和S-to-D链路，路径度量的第二、三部分可以被分别指定为和Pr(ysd|xs)。Among them, considering all For links (from all leaf relay nodes to the target node) and S-to-D links, the second and third parts of the path metric can be specified as and Pr(ysd |xs ).
具体的，在第一部分路径度量的表示中，表示为第t个乘积网格的路径度量的第一部分，其定义为：Specifically, in the representation of the first part of the path metric, The first part of the path metric expressed as the t-th product grid, which is defined as:
其中，L是乘积网格结构的总深度，di是的位置的数目(和表示从i-1时刻到i时刻，与状态转移有关联的xs和xr各自的子序列)，是xs和xr之间的汉明距离。上述等式右边的每一项都是相应的状态转移的分支度量。where L is the total depth of the product grid structure anddi is the number of positions ( and represents the respective subsequences of xs and xr associated with the state transition from time i-1 to time i), is the Hamming distance between xs and xr . The above equation Each term on the right is a branch metric for the corresponding state transition.
步骤4：在上述的乘积网格结构中，应用一种新的NBOD算法。该新的NBOD算法的复杂度与信息块的长度呈线性关系，不需要改变现有的采用传统的DF协议的中继节点，且可以扩展到任意的有网格结构的线性分组码上。图3是本发明实施例提供的NBOD算法。Step 4: In the above-mentioned product grid structure, a new NBOD algorithm is applied. The complexity of the new NBOD algorithm has a linear relationship with the length of the information block. It does not need to change the existing relay nodes using the traditional DF protocol, and can be extended to any linear block code with a grid structure. FIG. 3 is an NBOD algorithm provided by an embodiment of the present invention.
具体的，本发明实施例提供的NBOD算法基于在目标节点处构建的乘积网格结构并用BCJR算法实现。具体包括以下步骤：Specifically, the NBOD algorithm provided by the embodiment of the present invention is based on the product grid structure constructed at the target node and implemented by the BCJR algorithm. Specifically include the following steps:
步骤41：计算状态转移因子；Step 41: Calculate the state transition factor;
步骤42：计算前向递归因子；Step 42: Calculate the forward recursion factor;
步骤43：计算后向递归因子；Step 43: Calculate the backward recursion factor;
步骤44：计算源信息的对数似然比(log-likelihood ratio,LLR)；Step 44: Calculate the log-likelihood ratio (LLR) of the source information;
步骤45：得到源信息的硬解码输出。Step 45: Obtain the hard-decoded output of the source information.
需要说明的是，步骤41中状态转移因子为乘积网格结构的三部分路径度量It should be noted that the state transition factor in step 41 is the three-part path metric of the product grid structure
相乘，即：Multiply, that is:
从第t个乘积网格结构的初始化αt，0＝[1，0，...，0]开始，前向递归因子可以根据如下公式递归地进行计算：Starting from the initialization αt,0 = [1, 0, . . . , 0] of the t-th product trellis structure, the forward recursion factor can be calculated recursively according to
其中，Γ是能够在第(l+1)个网格深度中到达状态s′的状态集合。where Γ is the set of states that can reach state s' in the (l+1)th grid depth.
随着初始化βt，h＝[1，0，...，0]，从第t个乘积网格结构到根部的第l网格深度的后向递归因子可以根据如下公式递归地进行计算：With initialization βt,h = [1,0,...,0], the backward recursion factor from the t-th product grid structure to the l-th grid depth of the root can be calculated recursively according to
其中，Γ′为在第l个网格深度下连接到状态s的状态集合。where Γ' is the set of states connected to state s at the lth grid depth.
源信息us，l的对数似然比LLR可以根据如下公式进行计算：The log-likelihood ratio LLR of the source information us, l can be calculated according to the following formula:
其中，Γ+和Γ-是与us，l＝1和us，l＝0在第l个网格深度下分别对应的(sl，sl+1)状态对的集合。where Γ+ and Γ- are the set of (sl , sl+1 ) state pairs corresponding to us,l =1 and us,l =0 at the lth grid depth, respectively.
需要说明的是，本发明实施例提供的NBOD算法的复杂度与信息块的长度呈线性关系，且可以扩展到任意的有网格结构的线性分组码上。不同之处在于，与一个线性分组码相关的是时变网格，而与卷积码相关的是时不变网格。此外，该NBOD算法不用限定源节点和中继节点有相同的卷积码，可以通过跟踪每个卷积码在状态转换期间的相同数量的信息位来构建乘积网格。It should be noted that the complexity of the NBOD algorithm provided by the embodiment of the present invention has a linear relationship with the length of the information block, and can be extended to any linear block code with a trellis structure. The difference is that a linear block code is associated with a time-varying trellis, while a convolutional code is associated with a time-varying trellis. Furthermore, the NBOD algorithm does not require the source node and the relay node to have the same convolutional code, and can build a product grid by keeping track of the same number of information bits for each convolutional code during state transitions.
图4是本发明时实施例提供的NBOD算法示例中的基于DF协议的中继网络模型的结构。其中，R1，1有两个指向R2，1和R2，2的输出链路。因为R2，2与R3，1是叶节点，所以它们分别可以用和来表示。构建一个包含xs和所有中继节点传输符号的5阶乘积网格。其中，路径度量的第一部分可以表示为如下形式：FIG. 4 is a structure of a relay network model based on a DF protocol in an example of an NBOD algorithm provided by an embodiment of the present invention. Among them, R1,1 has two output links to R2,1 and R2,2 . Because R2 , 2 and R3 , 1 are leaf nodes, they can be used separately and To represent. Build a 5th order product grid containing xs and the symbols transmitted by all relay nodes. Among them, the first part of the path metric can be expressed as the following form:
路径度量的第二部分等于第三部分为Pr(ysd|xs)。生成的路径度量可以分解为分支度量的乘积。因此，本发明实施例提供的NBOD算法可以使用在此5阶乘积网格中。具体地，网格表示的分支度量是根据所有中继节点的可能性定义的，在网格表示中，路径度量等于所有沿这条路径的分支度量的乘积。The second part of the path metric is equal to The third part is Pr(ysd |xs ). The resulting path metrics can be decomposed into the product of branch metrics. Therefore, the NBOD algorithm provided by the embodiment of the present invention can be used in the 5th order product grid. Specifically, the branch metric of the trellis representation is defined in terms of the likelihood of all relay nodes, and in the trellis representation, the path metric is equal to the product of all branch metrics along this path.
为了评估本发明实施例提供的基于DF协议的，在单源单终端的复杂中继网络中的卷积码编码的误码率近似最优解码的获取方法，本发明实施例将NBOD算法应用到普遍的基于DF协议的网络中，并与现存的算法进行了BER性能的比较。In order to evaluate the approximate optimal decoding method for obtaining the bit error rate of convolutional code coding in a complex relay network with single source and single terminal based on the DF protocol provided by the embodiment of the present invention, the embodiment of the present invention applies the NBOD algorithm to In the general network based on DF protocol, and compare the BER performance with the existing algorithms.
图5是基于DF协议的单源单终端中继网络模型的5节点模型，所有链路的SNRs都是相同的，在此中继网络模型中共有3个中继节点。假设从S，R1，1，R1，2及R2，1传输的符号分别表示为xs，和则目标节点处可以建立一个同时考虑了xs，和的4阶乘积网格结构。本发明实施例提供的NBOD算法将被应用到此4阶乘积网格中。Figure 5 is a 5-node model of the single-source single-terminal relay network model based on the DF protocol, the SNRs of all links are the same, and there are a total of 3 relay nodes in this relay network model. Assuming that the symbols transmitted from S, R1,1 , R1,2 and R2,1 are denoted as xs , respectively, and Then the target node can establish a simultaneous consideration of xs , and The 4th factor product grid structure. The NBOD algorithm provided by the embodiment of the present invention will be applied to the 4th order product grid.
选用一个具有8进制的多项式生成器(5，7)8的1/2码率4状态的卷积编码器，设定其信息块的长度为K＝256位。图6是本发明实施例提供的NBOD算法(T＝1)与本领域公开的MRC，C-MRC以及SDF算法在图5所述的基于DF协议的5节点中继网络模型上的性能比较，假设每个链路都有同样的SNR。需要注意的是，当SNR比较低时，由于包错误事件发生的概率较大，SDF算法的性能比较差。当BER大约为10-4时，本发明实施例提供的NBOD算法比其他方案好了至少1dB。A 1/2 code rate 4-state convolutional encoder with an octal polynomial generator (5, 7)8 is selected, and the length of its information block is set to K=256 bits. 6 is a performance comparison between the NBOD algorithm (T=1) provided by the embodiment of the present invention and the MRC, C-MRC and SDF algorithms disclosed in the art on the 5-node relay network model based on the DF protocol described in FIG. 5 , It is assumed that each link has the same SNR. It should be noted that when the SNR is relatively low, the performance of the SDF algorithm is relatively poor due to the high probability of packet error events. When the BER is about 10-4, the NBOD algorithm provided by the embodiment of the present invention is at least 1 dB better than other solutions.
综上所述，本发明实施例提供的一种基于DF协议的在单源单终端的复杂中继网络模型中的卷积码编码的BER近似最优解码的获取方法，通过建立基于DF协议的单源单终端的复杂中继网络模型，在目标节点处建立一种新的、合理定义了分支度量的乘积网格结构，该乘积网格结构能够准确地表示所有可能的中继节点解码错误的场景，在此乘积网格结构中使用NBOD算法进行信息序列的位解码。该NBOD算法充分利用了现有的解码算法所忽视的事实，即由中继信道传输的错误数据包仍然是有效码字。此外，该NBOD算法的复杂度与信息块的长度呈线性关系，不需要改变现有的采用传统的DF协议的中继节点，且可以扩展到任意的有网格结构的线性分组码上。通过本发明实施例提供的基于DF协议的卷积码编码的单源单终端的复杂中继网络的BER近似最优解码的获取方法可以在不同的中继信道场景中实现比现有的解码算法更优的性能。To sum up, the embodiment of the present invention provides a method for obtaining the approximate optimal decoding of the BER of convolutional code encoding in a complex relay network model with a single source and single terminal based on the DF protocol. The complex relay network model with single source and single terminal establishes a new product grid structure that reasonably defines the branch metric at the target node. The product grid structure can accurately represent all possible relay node decoding errors. In this scenario, the NBOD algorithm is used for bit decoding of the information sequence in this product trellis structure. This NBOD algorithm takes full advantage of the fact that existing decoding algorithms ignore the fact that erroneous data packets transmitted by the relay channel are still valid codewords. In addition, the complexity of the NBOD algorithm is linearly related to the length of the information block, it does not need to change the existing relay node using the traditional DF protocol, and can be extended to any linear block code with a grid structure. The method for obtaining the BER approximate optimal decoding of a complex relay network with a single source and a single terminal based on the convolutional code encoding of the DF protocol provided by the embodiment of the present invention can achieve better decoding than the existing decoding algorithm in different relay channel scenarios. better performance.
本发明实施例还提供了如图7所示的一种误码率近似最优解码的获取系统，包括：The embodiment of the present invention also provides a system for obtaining the approximate optimal decoding of the bit error rate as shown in FIG. 7 , including:
模型构建单元701，用于建立基于解码转发协议、单源单终端的中继网络模型，所述中继网络模型包括单个源节点、若干中继节点和单个目标节点，每一中继节点均由单输入链路和多输出链路，且所述目标节点为唯一具有多输入链路的节点；A model construction unit 701 is used to establish a relay network model based on a decoding and forwarding protocol, a single source and a single terminal, the relay network model includes a single source node, several relay nodes and a single target node, each relay node is composed of a single input link and multiple output links, and the target node is the only node with multiple input links;
网络构建单元702，用于在所述目标节点建立乘积网格结构，所述乘积网格结构表示所有可能的中继节点解码错误的场景，并对中继节点解码所引入的错误比特之间的相关性进行准确建模；The network construction unit 702 is configured to establish a product grid structure at the target node, where the product grid structure represents all possible scenarios of relay node decoding errors, and decodes the error bits introduced by the relay node between the error bits. Correlations are accurately modeled;
度量确定单元703，用于根据所述乘积网格结构确定分支度量；a metric determining unit 703, configured to determine a branch metric according to the product grid structure;
信息计算单元704，用于根据确定分支度量的乘积网格结构和BCJR算法对源节点传输的源信息进行计算，得到所述源信息的误码率近似最优解码。The information calculation unit 704 is configured to calculate the source information transmitted by the source node according to the product trellis structure of the determined branch metric and the BCJR algorithm to obtain approximately optimal decoding of the bit error rate of the source information.
进一步地，所述源节点以S表示，所述中继节点以R表示，所述目标节点以D表示，所述中继网络模型中源节点至目标节点链路、源节点至中继节点链路及中继节点至目标节点链路均为分别具有信噪比γsd、γsr和γrd的二进制输入无记忆通道，γij,i,j∈(s,r,d)，其中：Further, the source node is represented by S, the relay node is represented by R, and the target node is represented by D. In the relay network model, the link from the source node to the target node and the link from the source node to the relay node are represented. The path and the link from the relay node to the target node are binary input memoryless channels with signal-to-noise ratios γsd , γsr and γrd respectively, γij ,i,j∈(s,r,d), where:
hij表示i节点与j节点之间的信道增益系数，Eb/N0表示信息比特与高斯噪声功率比； hij represents the channel gain coefficient between node i and node j, and Eb /N0 represents the ratio of information bits to Gaussian noise power;
所述中继网络模型中的每一节点为不能在同一时间传输和接收数据的半双工模式，在每次通信包含有两个不同的时隙；Each node in the relay network model is a half-duplex mode that cannot transmit and receive data at the same time, and each communication includes two different time slots;
在第一时隙中，源节点传输调制符号Xs，所述中继节点与所述目标节点监听到的带有噪声的Xs的观测分别为ysr和ysr，其中：In the first time slot, the source node transmits the modulation symbol Xs , and the observations of Xs with noise monitored by the relay node and the target node are ysr and ysr , respectively, where:
Zsr和Zsd是双边功率谱密度为N0/2每维度的零均值高斯噪声； Zsr and Zsd are zero-mean Gaussian noise with bilateral power spectral density N0 /2 per dimension;
所述中继节点对接收到的ysr进行解码并生成Cr作为所述源节点传输码字Cs的估计，根据Cr生成传输信号Xr；The relay node decodes the received ysr and generates Cr as an estimate of the source node transmission code word Cs , and generates a transmission signal Xr according to C r;
在第二时隙中，所述中继节点将Xr传送至所述目标节点，以yrd表示所述目标节点接收到的具有噪声的Xr，其中：In the second time slot, the relay node transmits Xr to the target node, denoting the received Xr with noise by the target node by yrd , where:
yrd＝hrdXr+Zrd，Zrd是双边功率谱密度为N0/2每维度的零均值高斯噪声。yrd = hrd Xr + Zrd , where Zrd is zero mean Gaussian noise per dimension with a bilateral power spectral density of N0 /2.
进一步地，在第一时隙，所述传输信号Xr由Cr通过二进制相移键控调制器产生。Further, in the first time slot, the transmission signal Xr is generated by Cr through a binary phase shift keying modulator.
进一步地，所述分支度量包括三个部分：Further, the branch metric includes three parts:
第一部分表示为：The first part is expressed as:
第二部分表示为：The second part is expressed as:
第三部分表示为：Pr(ysd|xs)；The third part is expressed as: Pr(ysd |xs );
其中，表示第t个乘积网格的路径度量的第一部分，其定义为：in, represents the first part of the path metric of the t-th product grid, which is defined as:
L表示乘积网格的总深度，di表示的位置的数目，和表示从i-1时刻到i时刻，与状态转移有关联的xs和xr各自的子序列，表示xs和xr之间的汉明距离，为表示中继网络模型中各节点形成的树结构的所有叶节点的集合，对于其任意的节点v，使用和分别表示其父节点、子节点和以其为根的子树的所有节点的集合。L is the total depth of the product grid, di is the number of positions, and represents the respective subsequences of xs and xr associated with the state transition from time i-1 to time i, represents the Hamming distance between xs and xr , In order to represent the set of all leaf nodes of the tree structure formed by each node in the relay network model, for any node v, use and Represents the set of all nodes of its parent node, child node, and subtree rooted with it, respectively.
进一步地，信息计算单元704具体用于：Further, the information calculation unit 704 is specifically used for:
首先，根据确定分支度量的乘积网格结构计算状态转移因子；First, the state transition factor is calculated according to the product grid structure of the determined branch metrics;
其次，根据确定分支度量的乘积网格结构计算前向递归因子；Second, the forward recursion factor is calculated according to the product grid structure that determines the branch metrics;
接着，根据确定分支度量的乘积网格结构计算后向递归因子；Next, calculate the backward recursion factor according to the product grid structure of the determined branch metrics;
接着，根据所述状态转移因子、所述前向递归因子和所述后向递归因子，计算所述源节点传输的源信息的对数似然比；Next, according to the state transition factor, the forward recursion factor and the backward recursion factor, calculate the log-likelihood ratio of the source information transmitted by the source node;
最后，根据所述源信息的对数似然比进行计算，得到所述源信息的误码率近似最优解码。Finally, calculation is performed according to the log-likelihood ratio of the source information to obtain approximately optimal decoding of the bit error rate of the source information.
本发明实施例还提供了一种终端，包括存储器、处理器及存储在存储器上且在处理器上运行的计算机程序，处理器执行计算机程序时，实现如图1所示的一种基于解码转发协议的误码率近似最优解码的获取方法的各个步骤。An embodiment of the present invention also provides a terminal, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, a decoding-based forwarding as shown in FIG. 1 is implemented. The protocol's bit error rate approximates each step of the optimal decoding method.
本发明实施例中还提供一种可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时，实现如图1所示的一种基于解码转发协议的误码率近似最优解码的获取方法的各个步骤。Embodiments of the present invention further provide a readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, implements a decoding and forwarding protocol-based approximate maximum bit error rate as shown in FIG. 1 . Each step of the method for obtaining the optimal decoding.
另外，在本发明各个实施例中的各功能模块可以集成在一个处理模块中，也可以是各个模块单独物理存在，也可以两个或两个以上模块集成在一个模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically alone, or two or more modules may be integrated into one module. The above-mentioned integrated modules can be implemented in the form of hardware, and can also be implemented in the form of software function modules.
所述集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-OnlyMemory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: U disk, removable hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes.
以上所述仅为本发明的较佳实施例而已，并不用以限制本发明，凡在本发明的精神和原则之内所作的任何修改、等同替换和改进等，均应包含在本发明的保护范围之内。The above descriptions are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention shall be included in the protection of the present invention. within the range.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811328693.1ACN109302269A (en) | 2018-11-09 | 2018-11-09 | A method and system for obtaining approximate optimal decoding of bit error rate |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811328693.1ACN109302269A (en) | 2018-11-09 | 2018-11-09 | A method and system for obtaining approximate optimal decoding of bit error rate |
| Publication Number | Publication Date |
|---|---|
| CN109302269Atrue CN109302269A (en) | 2019-02-01 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811328693.1APendingCN109302269A (en) | 2018-11-09 | 2018-11-09 | A method and system for obtaining approximate optimal decoding of bit error rate |
| Country | Link |
|---|---|
| CN (1) | CN109302269A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105144598A (en)* | 2012-12-03 | 2015-12-09 | 数字无线功率有限公司 | Systems and methods for advanced iterative decoding and channel estimation of concatenated coding systems |
| CN106982106A (en)* | 2015-12-28 | 2017-07-25 | 法国矿业电信学校联盟 | Recurrence sub-block is decoded |
| WO2017219718A1 (en)* | 2016-06-22 | 2017-12-28 | Huawei Technologies Co., Ltd. | Phase noise estimation and cancellation |
| CN108370253A (en)* | 2015-12-24 | 2018-08-03 | 英特尔公司 | Mixed schedule for ldpc decoding and the assembly line based on latch |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105144598A (en)* | 2012-12-03 | 2015-12-09 | 数字无线功率有限公司 | Systems and methods for advanced iterative decoding and channel estimation of concatenated coding systems |
| CN108370253A (en)* | 2015-12-24 | 2018-08-03 | 英特尔公司 | Mixed schedule for ldpc decoding and the assembly line based on latch |
| CN106982106A (en)* | 2015-12-28 | 2017-07-25 | 法国矿业电信学校联盟 | Recurrence sub-block is decoded |
| WO2017219718A1 (en)* | 2016-06-22 | 2017-12-28 | Huawei Technologies Co., Ltd. | Phase noise estimation and cancellation |
| Title |
|---|
| BIN QIAN: ""A Near BER-Optimal Decoding Algorithm for Convolutionally Coded Relay Channels With the Decode-and-Forward Protocol"", 《IEEE》* |
| Publication | Publication Date | Title |
|---|---|---|
| US10833792B2 (en) | Overlapped multiplexing-based modulation and demodulation method and device | |
| CN110336567B (en) | Joint iterative decoding method applied to G-LDPC coding cooperation | |
| Williamson et al. | Variable-length convolutional coding for short blocklengths with decision feedback | |
| Xiao et al. | Optimal decoding and performance analysis of a noisy channel network with network coding | |
| Zhang et al. | Finite block-length analog fountain codes for ultra-reliable low latency communications | |
| US8125938B2 (en) | Relay station and method for enabling reliable digital communications between two nodes in a wireless relay based network | |
| Yadav et al. | Commitment over compound binary symmetric channels | |
| Tian et al. | A scalable graph neural network decoder for short block codes | |
| Bourduge et al. | Bit interleaved chirp spread spectrum coded modulations with iterative decoding based on LDPC codes for coherent and non-coherent regimes | |
| US12432019B2 (en) | OMAMRC transmission method and system with variation in the number of uses of the channel | |
| Hernaez et al. | On the design of a novel joint network-channel coding scheme for the multiple access relay channel | |
| Li et al. | Robust delimitation of aggregated packets based on Viterbi decoding with natural redundancy in protocols | |
| You et al. | A simple near‐optimal path selection scheme for multi‐hop wireless relay networks based on Viterbi algorithm | |
| Qian et al. | A near BER-optimal decoding algorithm for convolutionally coded relay channels with the decode-and-forward protocol | |
| CN109302269A (en) | A method and system for obtaining approximate optimal decoding of bit error rate | |
| CN114337691B (en) | A Zipper code soft decoding method based on Chase-Pyndiah algorithm | |
| Deng et al. | Joint source channel anytime coding | |
| CN113965292B (en) | A low-complexity polar code SC decoding method based on aggregation construction | |
| Schlegel et al. | Generalized superposition modulation and iterative demodulation: A capacity investigation | |
| Aggarwal et al. | Distributed computing with channel noise | |
| Donev et al. | Polar-coded pulse position modulation for the poisson channel | |
| Jiang et al. | Iterative demodulation and decoding of polar coded BICM with bit-flipping | |
| Insausti et al. | Distributed pseudo-gossip algorithm and finite-length computational codes for efficient in-network subspace projection | |
| Ebrahimi et al. | Two-layer coded channel access with collision resolution: Design and analysis | |
| CN109861777A (en) | A kind of acquisition method and acquisition system for approximate optimal decoding of bit error rate |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | Application publication date:20190201 |