CN107508656A - A kind of Spinal joint source-channel decoding methods on BEC channels - Google Patents
A kind of Spinal joint source-channel decoding methods on BEC channels Download PDFInfo
- Publication number
- CN107508656A CN107508656A CN201710606156.8A CN201710606156A CN107508656A CN 107508656 A CN107508656 A CN 107508656A CN 201710606156 A CN201710606156 A CN 201710606156A CN 107508656 A CN107508656 A CN 107508656A
- Authority
- CN
- China
- Prior art keywords
- decoding
- channel
- information
- coding
- bec
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
- H04L1/0014—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the source coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0045—Arrangements at the receiver end
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/004—Arrangements for detecting or preventing errors in the information received by using forward error control
- H04L1/0045—Arrangements at the receiver end
- H04L1/0054—Maximum-likelihood or sequential decoding, e.g. Viterbi, Fano, ZJ algorithms
Abstract
The present invention relates to a kind of Spinal joint source-channel decoding methods on BEC channels, comprise the following steps:S1, using Spinal coding methods to information source data encoding, caused binary bits flow through BEC transmissions to decoding end;S2, for etc. general information source data, decoding end is decoded using maximum likelihood method, for sparse information source, the use of decoding endRow decoding is entered in calculating instead of the Hamming distance in maximum likelihood method, wherein, peFor BEC channel transition probabilities, piFor information source statistical probability, d is Hamming distance (Hamming Distance), if current bit is deleted, then make d=0, k is the length of blockette, compared with prior art, the present invention proposes a kind of Spinal decoding algorithms for sparse information source under BEC channels, can efficient coding fail the data compressed completely in physical layer.
Description
Technical Field
The present invention relates to a decoding method, and more particularly, to a method for jointly decoding a Spinal source channel on a BEC channel.
Background
The traditional digital communication takes shannon's source channel separation coding as a theoretical basis. In order to improve the transmission efficiency, a large amount of data is compressed by adopting source coding which is usually finished at an application layer; in order to improve the transmission quality, channel coding is adopted to correct errors of data in the transmission process, and the channel coding is usually completed in a physical layer. Thus, the physical layer has a common denominator assumption that the data to be transmitted has been compressed without any redundancy. Contrary to this assumption, a large number of applications inject uncompressed data into the network, such as e-mails, web pages, and uncompressed files, and there is a large amount of compressible data in the actual network traffic. Therefore, current physical layer techniques cannot fully utilize redundant information of data to improve system performance.
Disclosure of Invention
The present invention aims to overcome the defects of the prior art and provide a method for jointly decoding a Spinal source channel on a BEC channel, which is suitable for performing source compression, channel error correction protection and seamless code rate adaptation on a binary sparse source in a wired communication network. The transmission spectrum efficiency of the decoding method provided by the invention is close to the Channel capacity of a Binary Erasure Channel (BEC).
The purpose of the invention can be realized by the following technical scheme:
a method for jointly decoding a Spinal source channel on a BEC channel comprises the following steps:
s1, coding data sent by an information source by adopting a Spinal coding method, and transmitting coded binary bits to a decoding end through a BEC channel;
s2, for the binary equal probability information source (namely the probability of occurrence of bit 0 and bit 1 in the binary information source is the same), the decoding end adopts the maximum likelihood method for decoding, and for the sparse information source, the decoding end adoptsInstead of calculation of Hamming distance in maximum likelihood method, where p e For BEC channel transition probability, p i For the source statistical probability, d is the hamming distance (the number of different characters at the corresponding positions of two equal-length character strings), and if the current bit is deleted, d =0, k is the length of the sub-information block.
The step S1 specifically includes the following steps:
s11, coding an information bit sequence M = b with the block length of n bits 1 b 2 ...b n Divided into n/k sub-information blocks in units of k bits, i.e.
S12, for each sub information blockGenerating a state value S with the length of v bits by using a Hash function i ;
S13, with S i For seeding of the random number generator RNG, the multiple batches of outputs generate a pseudorandom sequence, which is mapped to a 1-bit coded output using a linear mapping function, wherein the RNG function is represented by the following equation:
s14, all state values S of the same batch i Output symbol x 1,j ,x 2,j ,...x i,j ...x n/k,j Constitute a coding channel, where the index i denotes the corresponding state value S i Subscript j represents the serial number of the batch;
and S15, sending the signal on the coding channel to a BEC channel, and after all the symbols of the first channel are sent, continuing coding and sending the symbols of the next channel until the sending end receives the feedback information of the correct decoding of the decoding end or the sending end gives up the information, and stopping sending the symbols.
In step S12, the Hash function is inputted as the sub-information blockAnd the previous state value S i-1 Initial state S 0 Set to 0 as shown in the following formula:
h:{0,1} v ×{0,1} k →{0,1} v
in step S15, the data in each coding channel is divided into a plurality of subchannels to be sent, and the data nodes allocated to each subchannel are not repeated.
In step S2, the maximum likelihood decoding process includes: and using a Hash function, a state initial value and a random number generator which are the same as those of the encoding end to completely reproduce a decoding tree at the decoding end, taking the state initial value as a root node, sequentially considering all possible values of each sub information block, sequentially exhausting, traversing from the root node to a leaf node, calculating the Hamming distance between the received symbol and the encoded symbol generated by all possible information source bits, wherein the path with the minimum Hamming distance is the decoding result.
In the step S2, in the process of decoding the data sent by the sparse information source, starting with the reproduction of a certain level of the decoding tree at the decoding end, only the decoding overhead and the minimum B paths in the node of the level are reserved, and only b.2 is calculated for each subsequent level of expansion k And (3) decoding cost of each child node, continuously reserving the minimum B paths, and so on, and finally reserving only the B paths, wherein the path with the minimum decoding cost is a decoding result.
Compared with the prior art, the invention has the following advantages:
(1) Aiming at physical layer compressible data transmitted in a BEC channel, the physical layer compression is realized based on a Spinal joint source channel decoding algorithm, and compared with the traditional source channel separation coding, the source channel joint coding can obtain higher spectrum efficiency. ( The information source coding is completed in an application layer, and the function of compressing data is realized; the channel coding is completed in the physical layer, and the protection of the data is realized. Thus, the physical layer has a common denominator assumption that the data to be transmitted has been compressed without any redundancy. Contrary to this assumption, a large number of applications inject uncompressed data into the network, such as e-mails, web pages, and uncompressed files, and there is a large amount of compressible data in the actual network traffic, but the method proposed by the present invention can implement both compression and protection at the physical layer. )
(2) Statistical information of sparse sources (i.e. ofP in (1) i ) The side information is transmitted into the decoding end as side information, the occupied resource of the side information can be ignored, and the method is particularly suitable for a wired communication network.
(3) A calculation method of prior Hamming distance is provided, a decoding end carries out decoding by utilizing statistical information to replace calculation of Hamming distance in traditional Spinal coding, and data compression is achieved.
(4) In step S15, the transmitted data is punctured, that is, the data on each coding channel is divided into a plurality of subchannels to be transmitted, and data nodes allocated to each subchannel are not repeated, so that a relatively smooth fine-grained spectrum efficiency can be obtained.
(5) In the process of decoding data sent by a sparse information source, starting from a certain level of a decoding end reproduction decoding tree, only preserving decoding overhead and the maximum B paths in the node of the level, expanding each subsequent level, and only calculating B.2 k The decoding cost of each child node and the maximum B paths are continuously reserved. And the decoding complexity is reduced under the condition of not influencing the decoding performance.
Drawings
FIG. 1 is a Spinal code diagram of the present embodiment;
fig. 2 is a schematic diagram of a puncturing procedure for data transmission in this embodiment;
FIG. 3 is a schematic diagram illustrating a decoding process according to the present embodiment;
FIG. 4 is a flow chart of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
Examples
In order to realize Spinal joint source channel decoding, the invention is realized by the following technical scheme, which mainly comprises the following steps:
1) The core idea of Spinal coding is to use a structure similar to a convolutional code to perform Hash random coding on an input bit sequence by introducing a Hash function. As shown in fig. 1, the specific steps are as follows:
11 Information bit sequence M = b) of length n bits of a coded block 1 b 2 ...b n Divided into n/k sub-information blocks in units of k bits, i.e.
12 Each sub information blockContinuously applying Hash function to generate corresponding state value S i (also called spin value, length v bits). Wherein S i Sequentially generated by a Hash function, the input of which is a sub-information blockAnd the previous state value S i-1 Initial state S 0 Set to 0 as shown in the following formula:
h:{0,1} v ×{0,1} k →{0,1} v
13 N/k states after information coding, in S i And (i is more than 0 and less than or equal to N/k) is a seed of a Random Number Generator (RNG), a pseudo-random sequence is generated by multi-batch output, and the sequence is mapped into 1-bit coded output by using a linear mapping function. Wherein the RNG function is represented by:
14)x i,j (i denotes the value S at the ith spin value i As seed, j is the batch number of the symbol), all S i Corresponding to the output symbols { x of the same batch 1,j ,x 2,j ,...x i,j ...x n/k,j And forming a coding channel.
15 The coded binary bits are transmitted to a decoding end via a BEC channel for decoding.
16 Spinal) is a code-free code that can continuously generate enough bits to transmit. And when the coding of one channel is finished and sent out, continuing to code and send the next channel until the sending end receives the feedback information of the correct decoding of the decoding end or the sending end gives up the information, and stopping sending the symbols.
17 Spinal) punctures the transmitted data for smoother fine-grained spectral efficiency, i.e., the bits on each code lane are not sent continuously but may be sent every few bits. The specific puncturing scheme is as follows:
171 The general idea of puncturing is: the transmitting end does not continuously transmit bits of each node of one channel but divides each channel into a plurality of sub-channels to transmit.
172 Fig. 2 shows a specific punching process. Each channel is divided into 8 sub-channels. In each sub-channel, only black nodes will be transmitted. The grey nodes represent the bits that have been sent. When the feedback information of the correct decoding at the decoding end is received, the rest sub-channels can not be sent any more.
2) For probable data such as a source, the decoding of the Spinal adopts Maximum Likelihood (ML) decoding, and uses the same initial value S as that of the encoding end 0 The Hash function h and RNG can completely reproduce a decoding tree at the decoding end, and the decoding end is provided with S 0 For root nodes, consideration of order2 of (2) k And (4) exhausting the possible values to n/k layers in sequence, traversing from a root node to a leaf node, calculating the Hamming distance between the received bits and the coded bits generated by all possible source bits, wherein the path with the minimum Hamming distance is a decoding result. For the decoding algorithm of the sparse source, the specific steps are as follows:
21 AdoptThe computation step of hamming distance in the ML decoding process is replaced and called prior hamming distance. Wherein p is e Denotes the BEC channel transition probability, in which d denotes the Hamming distance, but if the current bit is deleted, let d =0,p i Representing the statistical probability of the information source, and the derivation process is as follows:
known as p e d ·(1-p e ) n-d Representing the probability that the Hamming distance is d when n bits are compared, taking the logarithm of the probability to obtain
Since n.ln (1-p) e ) Is constant, and is left alone to reduce the complexity of the operationThe logarithm domain of the prior probability of the information sourceAdding to obtain a calculation formula of the calculation decoding
22 To reduce complexity, starting from a certain level of the decoding tree (set as d level), only the decoding overhead and the minimum B paths in the node of the level are reserved, and only B.2 needs to be calculated for each level of expansion subsequently k And (3) keeping the minimum B paths for the decoding overhead of each child node, and repeating the steps in the same way, and finally keeping only the B paths, wherein the path with the minimum decoding overhead is the decoding result, as shown in fig. 3.
An example of the invention is provided below: the source code length is 256bits, the distribution is unequal, and P (1) =0.1, P (0) =0.9, k =4, v =32, B =256.
1. The encoding process comprises the following specific steps:
step 1, dividing the 256-bits information bit sequence into 64 sub-information blocks with 4 bits as unit.
Step 2, initial state S 0 Set to 0, S 0 Inputting Hash function with first sub-information block, outputting state S 1 Repeating the process to generate S 1 ~S 64 For a total of 64 state values. Each state value is called a spine and has a length of 32bits.
Step 3, each spine value is taken as the seed of RNG, and 64 states S are totally obtained after information coding i (0<, i is less than or equal to 64) by S i For seeding of the random number generator RNG, the multi-batch output generates a pseudo-random sequence, which is mapped to a 1-bit coded output using a linear mapping function.
2. Assuming that a sending end transmits 2 coding channels, spinal uses prior information for decoding, and the specific steps are as follows:
step 1, with S 0 For the root node, each sub-information block 2 is considered sequentially 4 The number of possible values, i.e., 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000,1001,1010,1011,1100,1101,1110,1111。
step 2, calculating the decoding cost of the first spine node, wherein the data received by the decoding end on the two channels of the first spine node is y 1,1 And y 1,2 。
Step 3, mixing 0000 and S 0 Inputting Hash function to generate SS 0 ,SS 0 Input RNG function to generate a 1,1 And a 1,2 . Using a formulaThe prior hamming distance is calculated.
And 4, repeating the step 3, and calculating prior Hamming distances of 0001, 0010 \ 82301111, and the like.
And 5, repeating the steps 2, 3 and 4. If the path data after exceeding a certain spine node is larger than 256, 256 paths with the minimum cost are selected from the path data, and other paths are cut off to finally obtain 256 paths.
Claims (6)
1. A method for joint decoding of Spinal source channels on a BEC channel, comprising the steps of:
s1, coding a binary source by adopting a Spinal coding method, and transmitting a coded bit stream to a decoding end through a BEC channel;
s2, for binary equal probability information sources, a decoding end adopts a maximum likelihood method for decoding, and for sparse information sources, a decoding end adoptsInstead of calculation of hamming distance in maximum likelihood method for decoding, wherein p e As BEC channel transition probability, p i To the source statistical probability, d is the hamming distance, let d =0, k be the length of the sub-information block if the current bit is deleted.
2. The method as claimed in claim 1, wherein the step S1 comprises the following steps:
s11, coding an information bit sequence M = b with the block length of n bits 1 b 2 ...b n Divided into n/k sub-information blocks in units of k bits, i.e.
S12, for each sub information blockGenerating a state value S with the length of v bits by using a Hash function i ;
S13, with S i For seeding of the random number generator RNG, the multiple batches of outputs generate a pseudorandom sequence, which is mapped to a 1-bit coded output using a linear mapping function, wherein the RNG function is represented by the following equation:
s14, all state values S of the same batch are calculated i Output symbol x 1,j ,x 2,j ,...x i,j ...x n/k,j I form an encoding channel, where the index i denotes the corresponding state value S i Subscript j represents the serial number of the batch;
and S15, sending the signal on the coding channel to a BEC channel, and after all the symbols of the first channel are sent, continuing coding and sending the symbols of the next channel until the sending end receives the feedback information of the correct decoding of the decoding end or the sending end gives up the information, and stopping sending the symbols.
3. The method as claimed in claim 2, wherein the Hash function is inputted as the sub-information block m in step S12 i And the previous state value S i-1 Initial state S 0 And 0, as shown in the following formula:
h:{0,1} v ×{0,1} k →{0,1} v
4. a method as claimed in claim 2, wherein in step S15, the data in each coding channel is divided into a plurality of sub-channels for transmission, and the data nodes allocated to each sub-channel are not repeated.
5. The method as claimed in claim 1, wherein in step S2, the maximum likelihood decoding process comprises: and using a Hash function, a state initial value and a random number generator which are the same as those of the encoding end to completely reproduce a decoding tree at the decoding end, taking the state initial value as a root node, sequentially considering all possible values of each sub information block, sequentially exhausting, traversing from the root node to a leaf node, calculating the Hamming distance between the received symbol and the encoded symbol generated by all possible information source bits, wherein the path with the minimum Hamming distance is the decoding result.
6. The method as claimed in claim 1, wherein in step S2, during decoding of the data from the sparse source, a decoding end repeats a certain level of the decoding tree, only the decoding overhead and the minimum B paths in the node of the certain level are reserved, and each subsequent level of expansion calculates only b.2 k And (3) decoding cost of each child node, continuously reserving the minimum B paths, and so on, and finally reserving only the B paths, wherein the path with the minimum decoding cost is a decoding result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710606156.8A CN107508656A (en) | 2017-07-24 | 2017-07-24 | A kind of Spinal joint source-channel decoding methods on BEC channels |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710606156.8A CN107508656A (en) | 2017-07-24 | 2017-07-24 | A kind of Spinal joint source-channel decoding methods on BEC channels |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107508656A true CN107508656A (en) | 2017-12-22 |
Family
ID=60689200
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710606156.8A Pending CN107508656A (en) | 2017-07-24 | 2017-07-24 | A kind of Spinal joint source-channel decoding methods on BEC channels |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107508656A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112333127A (en) * | 2020-10-30 | 2021-02-05 | 中北大学 | Ratioless safety coding method based on Spinal code |
CN113472480A (en) * | 2020-03-31 | 2021-10-01 | 维沃移动通信有限公司 | Transmission processing method and equipment |
CN113783659A (en) * | 2021-08-24 | 2021-12-10 | 湖南遥昇通信技术有限公司 | Data processing method, device and medium based on binary erasure channel |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140006897A1 (en) * | 2012-06-29 | 2014-01-02 | International Business Machines Corporation | Correction of structured burst errors in data |
CN104184554A (en) * | 2014-09-03 | 2014-12-03 | 北京邮电大学 | Undersampling quantification and forwarding method for relay network |
-
2017
- 2017-07-24 CN CN201710606156.8A patent/CN107508656A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140006897A1 (en) * | 2012-06-29 | 2014-01-02 | International Business Machines Corporation | Correction of structured burst errors in data |
CN104184554A (en) * | 2014-09-03 | 2014-12-03 | 北京邮电大学 | Undersampling quantification and forwarding method for relay network |
Non-Patent Citations (2)
Title |
---|
LEE MILLER: ""最小汉明距译码规则"", 《CDMA系统工程手册》 * |
李娟: ""无速率spinal码译码算法研究及复杂度分析"", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113472480A (en) * | 2020-03-31 | 2021-10-01 | 维沃移动通信有限公司 | Transmission processing method and equipment |
CN113472480B (en) * | 2020-03-31 | 2022-09-27 | 维沃移动通信有限公司 | Transmission processing method and equipment |
CN112333127A (en) * | 2020-10-30 | 2021-02-05 | 中北大学 | Ratioless safety coding method based on Spinal code |
CN112333127B (en) * | 2020-10-30 | 2023-09-08 | 中北大学 | Spinal code-based rateless safety coding method |
CN113783659A (en) * | 2021-08-24 | 2021-12-10 | 湖南遥昇通信技术有限公司 | Data processing method, device and medium based on binary erasure channel |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11509414B2 (en) | Method and apparatus for encoding data using a polar code | |
US10666391B2 (en) | Method for encoding information bit sequence in communication network | |
CN101667884A (en) | Method and device for coding channel and method and device for decoding channel | |
WO2012155683A1 (en) | Low-density parity check method and device as the same | |
CN107508656A (en) | A kind of Spinal joint source-channel decoding methods on BEC channels | |
KR20070084151A (en) | Digital signal transmitting apparatus | |
CN111106838B (en) | Communication synchronization method, device and system | |
CN107566084A (en) | A kind of Spinal joint source-channel decoding methods on awgn channel | |
CN113067676A (en) | Novel bit mapping method in polar code high-order modulation system | |
CN108880748B (en) | Coding and decoding method of rateless Spinal code based on Latin square matrix | |
CN113541871A (en) | Method for generating code words and coder-decoder | |
CN112840581B (en) | Channel encoder and method for encoding information words | |
KR20080084178A (en) | Method of low density parity check code decoding using a plurality of layers | |
CN110336644B (en) | Layered coding method under high-dimensional modulation | |
CN117118562A (en) | Coding method and device | |
CN113259057A (en) | Cache-channel joint coding broadcasting method |
Legal Events
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171222 |