CN103532666B - Improve the method for data transmission efficiency and LT code performance in distributed transmission - Google Patents
Improve the method for data transmission efficiency and LT code performance in distributed transmission Download PDFInfo
- Publication number
- CN103532666B CN103532666B CN201310465628.4A CN201310465628A CN103532666B CN 103532666 B CN103532666 B CN 103532666B CN 201310465628 A CN201310465628 A CN 201310465628A CN 103532666 B CN103532666 B CN 103532666B
- Authority
- CN
- China
- Prior art keywords
- packet
- code
- degree
- distributed
- data
- 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.)
- Active
Links
Abstract
The invention discloses and a kind of improve the method for data transmission efficiency and LT code performance in distributed transmission, including: cascade protection LT code;By RSD distribution p1,p2,…,pkDegree of choosing d;D the packet chosen is carried out xor operation and finally gives coded data packet xi;Calculate distributed LT code model;The degree of cascade protection is distributed and introduces distributed LT code model, obtain distributed hierarchical protection LT code.The present invention adopts RSD distribution to determine the angle value of coded data packet, when choosing high-priority data, generation degree is chosen when being the coded data packet of 1 or 2 as much as possible in high priority packets, improve the likelihood probability that high priority packets is successfully decoded, it is ensured that what high priority packets was grouped translates probability;After the data message of two information sources wraps in the relaying joint operation that relaying one complexity of employing is very low, the packet forming a LT code integrated transmits, and improves resource utilization and data transmission efficiency, and reduces error probability.
Description
Technical field
The present invention relates to and a kind of improve the method for data transmission efficiency and LT code performance in distributed transmission.
Background technology
LT code (fountain codes) is a kind of linear block codes without code check, and time-limited n original input symbol can be encoded by it, generates unlimited number of coded identification.The cataloged procedure of LT code is: (1) produces random number i according to degree distribution ρ ();(2) i packet is randomly choosed from k raw data packets equal probability;(3) this i packet is carried out XOR, generate the packet after coding, be repeated continuously this process, generate coding groups.Wherein, desirable solitary wave degree distribution ρ () is:It is less desirable that ideality is distributed in erasure channel performance, and the distribution of robust solitary wave is suggested therewith.Consider that probability that desirable solitary wave calibration is 1 is big and diminish therewith the change of k, in order to ensure that number that initial degree is 1 and actual erasure channel condition propose robust solitary wave and be distributedWherein σ allows failed decoding probability N=β k, N to be with the quantity of the packet received needed for the probabilistic decoding success of probability 1-σ when receiving N number of coded data packet when being decoding;S is decoding process moderate is the quantity of the packet of 1.C is any normal number, and in practical application, c is generally less than 1; ρ () is added and can obtain into τ ():
μ (i)=(ρ (i)+τ (i)) β, wherein
Traditional LT code coding method adopts BP decoding algorithm:
(1) first degree of finding is the packet after the coding of 1, if can not find, and decoding failure;
(2) recover with degree be 1 coded data packet be connected raw data packets and it is connected deletion;
(3) coded data packet coupled for the raw data packets recovered is carried out with or and it is connected deletion;
(4) all raw data packets successfully translates then successfully decoded, otherwise repeats step (1)~(3).
Tradition decoding algorithm adopts when the different packet of priority is selected and chooses uniformly, and the likelihood probability that the packet of high priority is successfully decoded is low, and it is relatively low that what high priority packets was grouped translates probability.In model of communication system, two information source s1With information source s2Separate by relay send information to destination, it is assumed that relaying and destination node storage, process information ability be limited, within the given time, a packet can only be stored.Based on this condition, conventional transmission scheme is: the data packets of two information sources each carries out LT code coding respectively, then adopting time-multiplexed strategy by relay transmission the two information sequence, resource utilization is relatively low, and data transmission efficiency is relatively low and error probability is higher.
Summary of the invention
It is an object of the invention to overcome the deficiencies in the prior art, a kind of method of data transmission efficiency and LT code performance in raising distributed transmission selecting mechanism based on suitably degree distribution raw data packets is provided, on the one hand, RSD distribution is adopted to determine the angle value of coded data packet, when choosing high-priority data, generation degree is chosen when being the coded data packet of 1 or 2 as much as possible in high priority packets, coded data packet carries out uniformly choosing of RSD distribution again to some, improve the likelihood probability that high priority packets is successfully decoded, what ensure high priority packets packet translates probability;On the other hand, after the data message of two information sources wraps in the relaying joint operation that relaying one complexity of employing is very low, the packet forming a LT code integrated transmits, and improves resource utilization and data transmission efficiency, and reduces error probability.
It is an object of the invention to be achieved through the following technical solutions: improving the method for data transmission efficiency and LT code performance in distributed transmission, it comprises the following steps:
S1: cascade protection LT code, the coding parameter definition of LT code is as follows:
K: the number of primary data information (pdi) bag;
N: the number of the data packets after coding;
H: the number of significant data information bag;
T=hk: the ratio of whole primary data information (pdi) bag number shared by important information data packets number;
p1,p2,…,pk: the discrete probabilistic of the robust solitary wave degree distribution adopted during presentation code, wherein
q1,q2,…,qk: for describing the discrete probabilistic of the robust solitary wave distribution of important information, wherein
Be provided with the high priority packets of h bit, the package number of corresponding low priority be defined as h+1 ..., k}, encoder is output as { x1,x2,...,xn, n is that coded data packet is long;
Degree of hypothesis be the coded data packet number of 1 less than high-priority data quantity h, without loss of generality, defined formula p1K < < h, it will further be assumed that high priority coded data packet number is less than the half of coded data packet total number, wherein p2~0.5, h≤(p can be obtained1+p2) k, t≤p1+p2, p can be obtained1≤t≤p1+p2;
All degree be 1 packet choose and all choose from the data of high priority, be simultaneously introduced a definite positive integer Ω to choose from the data of high priority equally as two raw data packets of coded data packet that degree is 2, remaining degree be 2 and other angle value uniformly choose in whole raw data packets;
S2: initializing count=1, { 1,2,3..., n}, by RSD distribution p for i ∈1,p2,…,pkDegree of choosing d:
If d=1, from high priority packets 1,2,3 ..., h} uniformly chooses d packet;
If d=2, if count≤Ω, from high priority packets 1,2,3 ... h, in uniformly choose d packet, count=count+1;
Otherwise, from packet 1,2,3 ..., k} uniformly chooses d packet;
S3: d the packet chosen is carried out xor operation and finally gives coded data packet xi;
The main thought of cascade protection is: directly affecting the performance of LT code owing to spending the packet that angle value in distribution is 1 and 2, therefore two kinds of packets that degree is 1 and 2 are carried out certain constraint to improve its performance, all of degree is the p of 11K coded data packet is all from high priority packets { 1,2,3, ..., choosing in h}, this is because the decoding of LT code is directly initialize this decoding process by the coded data packet that degree is 1, these packets can participate in decoding the process without waiting other packets at once.Cascade protection strategy can improve the likelihood probability that the packet of high priority is successfully decoded, in like manner by spend be 1 the recovering the packet that Ω degree is 2 and can be further ensured that what the packet of high priority was grouped translates probability of packet.
S4: calculate distributed LT code model;
The step calculating distributed LT code model is: the LT code packet forming integration after the data message of two information sources wraps in the relaying joint operation of relaying one low complex degree of employing is transmitted: two information source s1And s2Sending k coded data packet to relaying, relaying carries out XOR after receiving packet, two information source s in this model simultaneously1And s2Degree be distributed as p (), it is assumed that X1Represent at information source s1XOR d1Individual data message is contracted for fixed output quotas the information after raw coding, X2Represent at information source s2XOR d2Individual data message is contracted for fixed output quotas the data packets after raw coding,Rear degree becomes d1+d2It is added probability density phase convolution according to stochastic variable and can be distributed as p () * p () through the degree after via node, the overall LT code word to obtain on erasure channel link obeys RSD distribution, is obtained degree distribution p () of two information sources by deconvolution.
S5: be distributed by the degree of cascade protection and introduce distributed LT code model, obtains distributed hierarchical protection LT code.
The invention has the beneficial effects as follows:
1) RSD distribution is adopted to determine the angle value of coded data packet, when choosing high-priority data, generation degree is chosen when being the coded data packet of 1 or 2 as much as possible in high priority packets, coded data packet carries out uniformly choosing of RSD distribution again to some, improve the likelihood probability that high priority packets is successfully decoded, it is ensured that what high priority packets was grouped translates probability;
2), after the data message of two information sources wraps in the relaying joint operation that relaying one complexity of employing is very low, the packet forming a LT code integrated transmits, and improves resource utilization and data transmission efficiency, and reduces error probability.
Accompanying drawing explanation
Fig. 1 is the communication system list relay-model of distributed two information source LT codes;
Fig. 2 is the Performance Simulation Results curve chart under LT code, PLT code different parameters;
Fig. 3 is information source total information bag number is 1600, probability of erasure is 0.01 5 kinds of fountain codes Performance Simulation Results curve charts;
Fig. 4 is information source total information bag number is 1600, probability of erasure is 0.06 5 kinds of fountain codes Performance Simulation Results curve charts.
Detailed description of the invention
Below in conjunction with accompanying drawing, technical scheme is described in further detail, but protection scope of the present invention is not limited to the following stated.
Improving the method for data transmission efficiency and LT code performance in distributed transmission, it comprises the following steps:
S1: cascade protection LT code, the coding parameter definition of LT code is as follows:
K: the number of primary data information (pdi) bag;
N: the number of the data packets after coding;
H: the number of significant data information bag;
T=hk: the ratio of whole primary data information (pdi) bag number shared by important information data packets number;
p1,p2,…,pk: the discrete probabilistic of the robust solitary wave degree distribution adopted during presentation code, wherein
q1,q2,…,qk: for describing the discrete probabilistic of the robust solitary wave distribution of important information, wherein
Be provided with the high priority packets of h bit, the package number of corresponding low priority be defined as h+1 ..., k}, encoder is output as { x1,x2,...,xn, n is that coded data packet is long;
Degree of hypothesis be the coded data packet number of 1 less than high-priority data quantity h, without loss of generality, defined formula p1K < < h, it will further be assumed that high priority coded data packet number is less than the half of coded data packet total number, wherein p2~0.5, h≤(p can be obtained1+p2) k, t≤p1+p2, p can be obtained1≤t≤p1+p2;
All degree be 1 packet choose and all choose from the data of high priority, be simultaneously introduced a definite positive integer Ω to choose from the data of high priority equally as two raw data packets of coded data packet that degree is 2, remaining degree be 2 and other angle value uniformly choose in whole raw data packets;
S2: initializing count=1, { 1,2,3..., n}, by RSD distribution p for i ∈1,p2,…,pkDegree of choosing d:
If d=1, from high priority packets 1,2,3 ..., h} uniformly chooses d packet;
If d=2, if count≤Ω, from high priority packets 1,2,3 ... h, in uniformly choose d packet, count=count+1;
Otherwise, from packet 1,2,3 ..., k} uniformly chooses d packet;
S3: d the packet chosen is carried out xor operation and finally gives coded data packet xi;
The main thought of cascade protection is: directly affecting the performance of LT code owing to spending the packet that angle value in distribution is 1 and 2, therefore two kinds of packets that degree is 1 and 2 are carried out certain constraint to improve its performance, all of degree is the p of 11K coded data packet is all from high priority packets { 1,2,3, ..., choosing in h}, this is because the decoding of LT code is directly initialize this decoding process by the coded data packet that degree is 1, these packets can participate in decoding the process without waiting other packets at once.Cascade protection strategy can improve the likelihood probability that the packet of high priority is successfully decoded, in like manner by spend be 1 the recovering the packet that Ω degree is 2 and can be further ensured that what the packet of high priority was grouped translates probability of packet.
S4: calculate distributed LT code model;
The step calculating distributed LT code model is: the model of communication system if Fig. 1 is distributed two information source LT codes: two information source s1With information source s2What be independent from send information to destination D by relaying R, it is assumed that relay and the ability of the storage of destination node, process information is limited, within the given time, can only store a packet.Assuming that information source is identical desirable channel to the link 1 and 2 of relaying, it does not have information deletion, and the link 3 being relayed to destination node is erasure channel, all of information deletion, packet loss all occur at link 3.This new coding strategy can improve efficiency of transmission, reduces error probability.The LT code packet forming integration after the data message of two information sources wraps in the relaying joint operation of relaying one low complex degree of employing is transmitted: two information source s1And s2Sending k coded data packet to relaying, relaying carries out XOR after receiving packet, two information source s in this model simultaneously1And s2Degree be distributed as p (), it is assumed that X1Represent at information source s1XOR d1Individual data message is contracted for fixed output quotas the information after raw coding, X2Represent at information source s2XOR d2Individual data message is contracted for fixed output quotas the data packets after raw coding,Rear degree becomes d1+d2It is added probability density phase convolution according to stochastic variable and can be distributed as p () * p () through the degree after via node, the overall LT code word to obtain on erasure channel link obeys RSD distribution, is obtained degree distribution p () of two information sources by deconvolution.
S5: be distributed by the degree of cascade protection and introduce distributed LT code model, obtains distributed hierarchical protection LT code.
The simulation result of this algorithm is as follows:
First, cascade protection LT code being abbreviated as PLT code, distributed LT code is abbreviated as DLT code, and distributed two, four information source LT code division are not abbreviated as 2DLT, 4DLT, and distributed two, four information source cascade protection LT code division are not abbreviated as 2DPLT, 4DPLT code.
First h value performances different for PLT is emulated by we, emulation takes h=50,150,300, Ω=h2, simulation parameter is: k=1000, c=0.05, σ=0.5, q=0.05.Simulation result is as shown in Figure 2, can be seen that from simulation result, as long as parameter h takes suitable value, when destination node receives less encoded digital information bag, PLT code and LT code performance are almost equal, but along with, when receiving coded data packet increase, PLT code decoding success rate is higher than gradually, is better than LT code.
Tradition LT code, distributed LT code, distributed hierarchical are protected the performance simulation of LT code to compare by Fig. 3 to 4, and LT, 2DLT, 4DLT, 2DPLT, 4DPLT are carried out Comparative Simulation.Former data packets number during emulation: k=1600 and 2400, degree distributed constant: c=0.05, σ=0.5, probability of erasure: q takes 0.01 and 0.06, to choosing of the parameter h of 2DPLT, 4DPLT code, by the Multi simulation running to different h values, choose the h value making performance best as far as possible, 2DPLT, 4DPLT code of occurrence: k=1600, h=780, Ω=390;2DPLT, 4DPLT code of k=2400, h=1180, Ω=590.Analogous diagram is as shown in Figs. 3-4.From above-mentioned simulation result it can be seen that the distributed hierarchical protection fountain codes DPLT after improving is when parameter h and Ω value takes suitable value, performance increases.
Claims (1)
1. improve the method for data transmission efficiency and LT code performance in distributed transmission, it is characterised in that: it comprises the following steps:
S1: cascade protection LT code, the coding parameter definition of LT code is as follows:
K: the number of primary data information (pdi) bag;
N: the number of the data packets after coding;
H: the number of significant data information bag;
T=h/k: the ratio of whole primary data information (pdi) bag number shared by important information data packets number;
p1,p2,…,pk: the discrete probabilistic of the robust solitary wave degree distribution adopted during presentation code, wherein
q1,q2,…,qk: for describing the discrete probabilistic of the robust solitary wave distribution of important information, wherein
Be provided with the high priority packets of h bit, the package number of corresponding low priority be defined as h+1 ..., k}, encoder is output as { x1,x2,...,xn, n is that coded data packet is long;
Degree of hypothesis be the coded data packet number of 1 less than high-priority data quantity h, without loss of generality, defined formula p1K < < h, it will further be assumed that high priority coded data packet number is less than the half of coded data packet total number, wherein p2~0.5, h≤(p can be obtained1≤p2) k, t≤p1+p2, p can be obtained1≤t≤p1+p2;
All degree be 1 packet choose and all choose from the data of high priority, be simultaneously introduced a definite positive integer Ω to choose from the data of high priority equally as two raw data packets of coded data packet that degree is 2, remaining degree be 2 and other angle value uniformly choose in whole raw data packets;
S2: initializing count=1, { 1,2,3..., n}, by RSD distribution p for i ∈1,p2,…,pkDegree of choosing d:
If d=1, from high priority packets 1,2,3 ..., h} uniformly chooses d packet;
If d=2, if count≤Ω, from high priority packets 1,2,3 ..., h} uniformly chooses d packet, count=count+1;
Otherwise, from packet 1,2,3 ..., k} uniformly chooses d packet;
RSD distribution is adopted to determine the angle value of coded data packet, when choosing high-priority data, generation degree is chosen when being the coded data packet of 1 or 2 as much as possible in high priority packets, coded data packet carries out uniformly choosing of RSD distribution again to some, improve the likelihood probability that high priority packets is successfully decoded, it is ensured that what high priority packets was grouped translates probability;
S3: d the packet chosen is carried out xor operation and finally gives coded data packet xi;
S4: calculate distributed LT code model, the step calculating distributed LT code model is: the LT code packet forming integration after the data message of two information sources wraps in the relaying joint operation of relaying one low complex degree of employing is transmitted: two information source s1And s2Sending k coded data packet to relaying, relaying carries out XOR after receiving packet, two information source s in this model simultaneously1And s2Degree be distributed as p (), it is assumed that X1Represent at information source s1XOR d1Individual data message is contracted for fixed output quotas the information after raw coding, X2Represent at information source s2XOR d2Individual data message is contracted for fixed output quotas the data packets after raw coding,Rear degree becomes d1+ d2It is added probability density phase convolution according to stochastic variable and can be distributed as p () * p () through the degree after via node, the overall LT code word to obtain on erasure channel link obeys RSD distribution, is obtained degree distribution p () of two information sources by deconvolution;
After the data message of two information sources wraps in the relaying joint operation that relaying one complexity of employing is very low, the packet forming a LT code integrated transmits, and improves resource utilization and data transmission efficiency, and reduces error probability;
S5: be distributed by the degree of cascade protection and introduce distributed LT code model, obtains distributed hierarchical protection LT code.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310465628.4A CN103532666B (en) | 2013-06-06 | 2013-10-08 | Improve the method for data transmission efficiency and LT code performance in distributed transmission |
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013102223298 | 2013-06-06 | ||
CN201310222329.8 | 2013-06-06 | ||
CN201310222329 | 2013-06-06 | ||
CN201310465628.4A CN103532666B (en) | 2013-06-06 | 2013-10-08 | Improve the method for data transmission efficiency and LT code performance in distributed transmission |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103532666A CN103532666A (en) | 2014-01-22 |
CN103532666B true CN103532666B (en) | 2016-07-20 |
Family
ID=49934383
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310465628.4A Active CN103532666B (en) | 2013-06-06 | 2013-10-08 | Improve the method for data transmission efficiency and LT code performance in distributed transmission |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103532666B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106533614B (en) * | 2016-06-28 | 2019-04-09 | 东南大学 | A kind of distributed rupee transform coding method based on partial information feedback |
WO2022016315A1 (en) * | 2020-07-20 | 2022-01-27 | Qualcomm Incorporated | Using degree information for lower layer transmission and reception |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101902296A (en) * | 2010-06-23 | 2010-12-01 | 中兴通讯股份有限公司 | Coding/decoding method and device for fountain codes |
CN102227103A (en) * | 2011-06-23 | 2011-10-26 | 天津大学 | Channel coding transmission method based on unequal mistake protection fountain code |
CN102664639A (en) * | 2012-04-19 | 2012-09-12 | 哈尔滨工业大学深圳研究生院 | Encoding method of distributed LT code |
-
2013
- 2013-10-08 CN CN201310465628.4A patent/CN103532666B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101902296A (en) * | 2010-06-23 | 2010-12-01 | 中兴通讯股份有限公司 | Coding/decoding method and device for fountain codes |
CN102227103A (en) * | 2011-06-23 | 2011-10-26 | 天津大学 | Channel coding transmission method based on unequal mistake protection fountain code |
CN102664639A (en) * | 2012-04-19 | 2012-09-12 | 哈尔滨工业大学深圳研究生院 | Encoding method of distributed LT code |
Also Published As
Publication number | Publication date |
---|---|
CN103532666A (en) | 2014-01-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101453297B (en) | Encoding method and apparatus for low density generation matrix code, and decoding method and apparatus | |
CN101459430B (en) | Encoding method and apparatus for low density generation matrix code | |
EP3193472B1 (en) | Data packet processing method and device | |
CN101425871B (en) | Multi-element error correcting code transmitting and receiving apparatus, data communication system and related method | |
CN105991227B (en) | Data coding method and device | |
CN103250463A (en) | Subset coding for communication systems | |
US9444494B2 (en) | Systems and methods for network coding using convolutional codes | |
Hof et al. | Polar coding for reliable communications over parallel channels | |
CN106254030B (en) | Two-way coding and decoding method without rate Spinal code | |
CN101582744A (en) | Encoding and decoding method of RS fountain codes based on iterative approach | |
CN107124251A (en) | A kind of polarization code encoding method based on arbitrary kernel | |
CN103944676A (en) | MLT code coding and decoding method based on deep space communication environment | |
CN103152652A (en) | Video frame data network transmission method based on Cauchy reed-solomon (RS) code | |
CN101286745B (en) | Coding method and device by interleaving | |
CN103532666B (en) | Improve the method for data transmission efficiency and LT code performance in distributed transmission | |
US20240048157A1 (en) | Encoding method, decoding method, electronic device and storage medium | |
KR101643039B1 (en) | Methods for optimizing degree distribution of luby-transform code | |
CN102195743B (en) | Coding scheme of dynamic real-time fountain code | |
CN103812615A (en) | Coding method for Luby transform code in distributed communication system, and relay communication method | |
CN107222294A (en) | A kind of improved fountain codes degree Distribution Algorithm | |
WO2017041232A1 (en) | Encoding and decoding framework for binary cyclic code | |
CN103346859B (en) | The encoding and decoding method of distributed unequal error protection LT code | |
CN107257244A (en) | A kind of fountain code encoding method based under broadcast environment | |
CN108631931A (en) | A kind of method and device of construction polarization code sequence | |
CN102916707B (en) | Compatible convolutional code generator polynomial defining method, coding method and encoder |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |