CN101179279A - Non-rate code coding/decoding method fit for additive white Gaussian noise channel - Google Patents

Abstract

The invention discloses a method for encoding and decoding without rates, which is applicable to an additive white Gaussian noise channel. The method comprises an encoding method and a decoding method. The basic technological thought is to add an accumulator behind an encoder of LT codes in order to ensure that degrees of encoding nodes in a bipartite graph are no longer one, thereby solving the 'error platform' problem of the LT codes in the additive white Gaussian noise channel and adopting a widely used systematic code structure at the same time. Then two methods are provided with easy implementation and better performance for the encoders to select signal nodes for newly added check nodes. A method results in approximately equal distribution of signal node degrees and the other results in limitation of the distribution of signal node degrees at certain rate.

Description

The non-rate code coding/decoding method that is suitable for additive white gaussian noise channels

Technical field

The present invention relates to wireless communication field, be specifically related to a kind of non-rate code coding/decoding method that is suitable for additive white Gaussian noise (AWGN) channel.

Background technology

Block code is widely used in the channel error correction coding.We estimate channel parameter usually earlier, are fixed as (N, K) block code of R=N/K according to code check of this parameter designing.When estimated channel parameter during, though can realize reliable transmission, caused the waste of transmission, because can use the block code of higher code check this moment greater than the channel parameter of reality; When estimated channel parameter during less than the channel parameter of reality, can not realize reliable transmission, need the block code of lower code check this moment.Therefore, do not know that at transmitting terminal channel accurately under the state information situation, the reliable effectively transmission of guarantee information, often need ARQ.How the suitable code check of adaptive selection transmits, and to adapt to different channel parameters, no-rate codes provides a kind of new approaches of dealing with problems for us.

No-rate codes and conventional fixed bit rate coding mode maximum different are that it does not set cbr (constant bit rate) at transmitting terminal, and transmitting terminal can produce encoded packets in some way endlessly and send.Receiving terminal then can receive these encoded packets trial and error decoding then.If decoding failure, receiving terminal can again receive some encoded packets more and continue trial and error decoding then.Receiving terminal will repeat this process up to successfully decoded always.At this moment receiving terminal only need send a very simple feedback signal and inform that transmitting terminal is successfully decoded, and transmitting terminal stops to send then, has so just finished whole transmission course.At this moment, the code check of actual transmissions depends on the encoded packets number of actual transmission, and the encoded packets number that need to send depends on channel conditions at that time, how to make the code check of actual transmissions approach the key issue that at that time channel capacity becomes the no-rate codes design.

Luby has proposed to be called LT sign indicating number (seeing " LTCodes ", Proceedings of the 43rd Annual IEEE Symposium on Foundation ofComputer Science) for binary system removes the no-rate codes of deleting channel (BEC) design.Do not know that at transmitting terminal channel removes when deleting rate, the LT sign indicating number can provide reliable transmission and can approach channel capacity.But the LT sign indicating number also is not suitable for awgn channel.The encoder of LT sign indicating number is at first selected the plurality of data bag, sends with their verification with as encoded packets then.The transmitting terminal of LT sign indicating number can produce encoded packets by the way endlessly and send.The bigraph (bipartite graph) of LT sign indicating number as shown in Figure 1.Two category nodes are arranged among the figure, and circle is represented variable node (variable node), box indicating check-node (checknode).And variable node is divided into two classes, and the left side is information node (information node) representative data bag, and the right is that coding nodes (parity node) is represented encoded packets.The verification of each node that links to each other with each check-node and be 0.If from the viewpoint LT sign indicating number of bigraph (bipartite graph), the number of degrees of its coding nodes are constant to be 1.Because the message of this part variable node is upgraded never, all the time be the initial value that receives, their error probability can not trend towards 0 by the channel conditions decision along with the increase of iterations, inject and influence iterative decoding process thereby form a fixing mistake, this can have a strong impact on decoding algorithm based on bigraph (bipartite graph), and (for example: belief propagation (Blief Propagation (BP)) decoding algorithm) thereby performance causes the generation of " mistake platform " (Error Floor).Simulation result in the article " Rateless Codes on noisy channels " of Ravi Palanki etc. has clearly illustrated this point.Shokrollahi has also designed no-rate codes for BEC, be called the Raptor sign indicating number (see " Raptor Codes ", IEEE Transactions on Information Theory, Vol.52, No.6, June2006)." mistake platform " problem when though it can solve the LT sign indicating number and works in awgn channel, (complexity of coding and decoding all improves for low-density parity-check, LDPC) sign indicating number cascade as ISN and low-density checksum with the LT sign indicating number for it.In addition, systematic code is widely adopted in the practical application, because channel condition does not need decoding when fine, can reduce decoding and consume, and the LT sign indicating number does not have the option of systematic code, and the systematic code option of Raptor sign indicating number will further improve the coding and decoding complexity.

Summary of the invention

The purpose of this invention is to provide a kind of non-rate code coding/decoding method that is suitable for awgn channel, we abbreviate this no-rate codes that is suitable for awgn channel as SAR (Systematic AccumulateRateless) sign indicating number.

The non-rate code coding/decoding method that is suitable for awgn channel comprises coding method and interpretation method.Coding method is as follows, and consideration coding transmitting terminal will send m packet, is made up of the plurality of data bit in each packet, and each packet inside comprises that a cyclic redundancy check (CRC) code is used for decoder and judges whether decoding is successful.d ₀, d ₁..., d _j..., d _M-1Represent each packet respectively, subscript j is the numbering of packet.t _iThe presentation code bag, wherein i is the numbering of encoded packets.The coding transmitting terminal at first sends m packet the information bit part that forms systematic code, produces encoded packets t then as follows _i:

1) at first by making the approximate equally distributed information node selection mode of the information node number of degrees or making the information node selection mode generation m dimension two element field vector { G that distributes limited at certain code check place information node number of degrees _Ji, m dimension two element field vector { G _JiIn for the number of the element of " 1 " be a _i

2) m dimension two element field vector { G _JiIn element G _JiValue be " 0 " or " 1 ", element G _JiBe " 1 " that then the packet that is numbered j of its correspondence is with selected, the packet that these are selected makes to obtain after mould 2 adds and be worth s by bit _i, can be expressed as:

$s_i=\underset{j=0}{\overset{m-1}{\mathrm{\Σ}}}{d}_j{G}_{\mathrm{ji}},i=\mathrm{0,1,2}...$

3) by accumulator with s _iWith a last encoded packets t _I-1Make mould 2 and obtain new encoded packets t by bit _i, be expressed as following formula:

t _i=t _I-1+ s _i, t wherein _-1=0

Transmitting terminal produces encoded packets endlessly according to above rule and informs that up to receiving terminal it stops to send.

Interpretation method comprises the steps:

At first receive packet, begin decoding after m packet finishes receiving:

1) decoder utilizes the cyclic redundancy check (CRC) code in each bag to judge that whether m packet be all correct, if all correct, changes step 5) over to; Otherwise change step 2 over to);

2) decoder receives some encoded packets;

3) because receiving terminal is known each encoded packets t _iCorresponding m dimension two element field vector { G _Ji, so decoder can be accurately at the bigraph (bipartite graph) of this sign indicating number of receiving terminal reconstruct;

4) move decoding algorithm on the bigraph (bipartite graph) of reconstruct in step 3), this decoding algorithm is BP algorithm or improved BP algorithm, and then changes operating procedure 1 over to);

5) decoding finishes, and receiving terminal informs that by feedback channel transmitting terminal stops to send.

The described information node number of degrees that make are similar to equally distributed information node selection mode: be newly-increased check-node C _iSelect the step of a information node as follows, wherein i be 0 o'clock from step 2):

1) newly-increased check-node C _iWith a last coding nodes P _I-1Link to each other, thus with original m information node I ₀, I ₁..., I _M-1, (i-1) individual coding nodes P ₀, P ₁..., P _I-1, (i-1) individual check-node C ₀, C ₁..., C _I-1The bigraph (bipartite graph) that constitutes links to each other, and needs to upgrade C _iDistance to each information node;

2) be newly-increased check-node C _iSelect one to be attached thereto, because the bigraph (bipartite graph) change in topology needs to upgrade newly-increased check-node C from its information node apart from maximum _iTo the distance of each information node, repeating step 2) select to finish up to a information node;

3) newly-increased check-node C _iWith newly-increased coding nodes P _iLink to each other, formed one by m information node I ₀, I ₁..., I _M-1, i coding nodes P ₀, P ₁..., P _i, i check-node C ₀, C ₁..., C _iThe new bigraph (bipartite graph) that constitutes.

When selecting information node step 2), we will select one of number of degrees minimum when more than one of distance information node farthest, if at this moment during still more than one of the information node of number of degrees minimum, we will select one of them at random.

Described making: if the expectation information node number of degrees reach distribution λ (x) when being distributed in code check for R at the limited information node selection mode of certain code check place information node number of degrees distribution, the coding transmitting terminal will use in advance PEG (Progressive Edge-Growth) algorithm (see " Regular and Irregular Progressive Edge-GrowthTanner Graphs " ', IEEE Transactions On Information Theory, Vol.51, No.1, January2005) generating a code check is R, the information node number of degrees are distributed as the bigraph (bipartite graph) of λ (x), are newly-increased check-node C then _iSelect the method for information node as follows:

1) when code check during more than or equal to R, encoder is newly-increased check-node C according to the annexation of the bigraph (bipartite graph) that generates in advance _iSelect a _iIndividual information node;

2) when code check during less than R, encoder will use and make the approximate equally distributed information node selection mode of the information node number of degrees be newly-increased check-node C _iSelect a information node.

Basic fundamental thought of the present invention is to add an accumulator again in the encoder back of LT sign indicating number, so that the number of degrees of coding nodes no longer are 1 in the bigraph (bipartite graph), thereby solve " mistake platform " problem that the LT sign indicating number works in awgn channel, adopted the systematic code structure that is widely used simultaneously.

Description of drawings

Fig. 1 is the bigraph (bipartite graph) of LT sign indicating number;

Fig. 2 is the coding schematic diagram of SAR sign indicating number;

Fig. 3 is the decoding flow chart of SAR sign indicating number;

Fig. 4 is the bigraph (bipartite graph) of SAR sign indicating number;

Fig. 5 is the approximate bit error rate comparison diagram of equally distributed SAR sign indicating number on each speed point of the LT sign indicating number and the information node number of degrees, signal to noise ratio snr (E _s/ N ₀The 1.9dB of)=-;

Fig. 6 is the approximate bit error rate comparison diagram of equally distributed SAR sign indicating number under each signal to noise ratio of the LT sign indicating number and the information node number of degrees, code check R=0.5;

Fig. 7 is a LT sign indicating number and at the R=0.5 place information node number of degrees limited bit error rate comparison diagram of SAR sign indicating number on each speed point that distribute, signal to noise ratio snr (E _s/ N ₀The 1.9dB of)=-.

Embodiment

The non-rate code coding/decoding method that is suitable for awgn channel comprises coding method and interpretation method.It is characterized in that coding method is as follows, consider that the coding transmitting terminal will send 10000 packets, form by the plurality of data bit in each packet, each packet inside comprises a cyclic redundancy check (CRC) code, this cyclic redundancy check (CRC) code adopts CRC32, is used for decoder and judges whether decoding is successful.d ₀, d ₁..., d _j..., d _M-1Represent each packet respectively, subscript j is the numbering of packet.t _iThe presentation code bag, wherein i is the numbering of encoded packets.The coding transmitting terminal at first sends 10000 packets the information bit part that forms systematic code, produces encoded packets t then as follows _i:

1) at first by making the approximate equally distributed information node selection mode of the information node number of degrees or making the information node selection mode generation 10000 dimension two element field vector { G that distribute limited at certain code check place information node number of degrees _Ji, 10000 dimension two element field vector { G _JiIn for the number of the element of " 1 " be a _i

2) 10000 dimension two element field vector { G _JiIn element G _JiValue be " 0 " or " 1 ", element G _JiBe " 1 " that then the packet that is numbered j of its correspondence is with selected, the packet that these are selected makes to obtain after mould 2 adds and be worth s by bit _i, can be expressed as:

$s_i=\underset{j=0}{\overset{m-1}{\mathrm{\Σ}}}{d}_j{G}_{\mathrm{ji}},i=\mathrm{0,1,2}...$

3) by accumulator with s _iWith a last encoded packets t _I-1Make mould 2 and obtain new encoded packets t by bit _i, be expressed as following formula:

t _i=t _I-1+ s _i, t wherein _-1=0

Transmitting terminal produces encoded packets endlessly according to above rule and informs that up to receiving terminal it stops to send.

Interpretation method comprises the steps:

At first receive packet, begin decoding after 10000 packets finish receiving:

1) decoder utilizes the cyclic redundancy check (CRC) code in each bag to judge that whether 10000 packets are all correct, if all correct, change step 5) over to; Otherwise change step 2 over to);

2) decoder receives 500 encoded packets;

3) because receiving terminal is known each encoded packets t _i10000 corresponding dimension two element field vector { G _Ji, so decoder can be accurately at the bigraph (bipartite graph) of this sign indicating number of receiving terminal reconstruct;

4) move decoding algorithm on the bigraph (bipartite graph) of reconstruct in step 3), this decoding algorithm is BP algorithm or improved BP algorithm, and then changes operating procedure 1 over to);

5) decoding finishes, and receiving terminal informs that by feedback channel transmitting terminal stops to send.

When finding that the encoded packets that receives is not enough to correct decoding, receiving terminal need receive 500 encoded packets again.When receiving terminal was received m packet and n encoded packets, corresponding code rate was:

$R=\frac{m}{m+n}$

In order to represent that conveniently, we use the R reciprocal of code check ^-1Portray the variation of code check.After n encoded packets of each many reception Δs, code check is changed to Δ R ^-1:

$\mathrm{\Δ}{R}^{-1}=\frac{\mathrm{\Δn}}{m}$

Δ n is 500 herein, so Δ R ^-1Be 0.05.Table 1 has provided each code check of encoded packets number from 7500 to 10000 correspondences that receiving terminal receives.

n	7500	8000	8500	9000	9500	10000
							R -1	1.75	1.80	1.85	1.90	1.95	2.00

Table 1

According to above description to coding and decoding, the bigraph (bipartite graph) of SAR sign indicating number as shown in Figure 4.Two category nodes are arranged among the figure, and circle is represented variable node (variable node), box indicating check-node (check node).And variable node is divided into two classes, and the left side is information node (information node), and the right is coding nodes (paritynode).One total m is individual for information node, is expressed as I respectively ₀, I ₁..., I _j..., I _M-1, with packet d ₀, d ₁..., d _j..., d _M-1Corresponding one by one.Coding nodes is expressed as P respectively ₀, P ₁..., P _i..., with encoded packets t ₀, t ₁..., t _i... corresponding one by one.Check-node is expressed as C respectively ₀, C ₁..., C _i....The verification of each node that links to each other with a check-node and be 0.The maximum difference of the bigraph (bipartite graph) of SAR sign indicating number and general bigraph (bipartite graph) is that it can constantly enlarge.New encoded packets t of every generation _i, will increase a new coding nodes P among the figure _iWith a new check-node C _iSo along with increasing of encoded packets, this figure can be increasing.Wherein newly-increased check-node C _iFix with the annexation of coding nodes, except first check-node C ₀Only and P ₀Outside linking to each other, C _iAlways with a last coding nodes P _I-1And newly-increased coding nodes P _iLink to each other.

In the bigraph (bipartite graph), the sum that is connected to the limit of certain node is called the number of degrees of this node.The number of degrees that we define information node are distributed as:

$\mathrm{\λ}\left(x\right)=\underset{i}{\mathrm{\Σ}}{\mathrm{\λ}}_i{x}^{i-1}$

λ _iBeing illustrated in all and connecting in check-nodes and all limits of information node, is the shared ratio in limit that the information node of i links to each other with the number of degrees.

From the encoder schematic diagram of Fig. 2 as can be seen, the key issue of SAR sign indicating number design is exactly how to produce { G _JiProblem, just how to select to produce encoded packets t _iA _iThe problem of individual packet.From the viewpoint of graph theory, transmitting terminal produces an encoded packets t _i, increase a check-node C in the corresponding bigraph (bipartite graph) _iWith coding nodes P _i, newly-increased check-node C _iA is arranged _iThe bar limit links to each other with information node, so, selects to produce t _iThe question essence of packet on be exactly to be newly-increased information node C _iSelect a _iThe problem of individual information node, and that different selection modes will make that information node distributes at the number of degrees at each code check place will be different also just makes the performance difference of SAR sign indicating number.The present invention proposes two kinds is easy to realize the information node selection mode of better performances.

The described information node number of degrees that make are similar to equally distributed information node selection mode: a that adopts in this mode _iBe constant a, a is taken as " 4 " herein.This mode will make the number of degrees of information node be distributed in each code check place and be similar to evenly, make the approximate uniform information node selection mode of information node number of degrees distribution so we are called this mode, the SAR sign indicating number that adopts this mode is called the approximate equally distributed SAR sign indicating number of the information node number of degrees.

Need to prove: a node links to each other with another node through some limits in the bigraph (bipartite graph), and these limits form a paths, and the number on limit is the length in this path in the path herein; The length that connects two node shortest paths is the distance of these two nodes, if do not have the path between two nodes, then their distance is infinitely great.As previously mentioned, the decoding algorithm that adopts in the decode procedure of SAR sign indicating number all is based on bigraph (bipartite graph), the length of bigraph (bipartite graph) centre circle will directly influence the performance of decoding algorithm, the big more decoding performance of length of circle is good more, so that the length that the circle in the bigraph (bipartite graph) of the SAR sign indicating number that generates is tried one's best is to select the basic principle of the information node that is attached thereto for newly-increased check-node, this just requires us, and this check-node of eternal chosen distance is farthest when selecting information node.

As no-rate codes, the transmitting terminal of SAR sign indicating number must send encoded packets endlessly, at encoded packets t of every generation _iThe time, increase a check-node C in the bigraph (bipartite graph) _i, transmitting terminal need be newly-increased check-node C _i4 information nodes that selection is attached thereto.Be newly-increased check-node C _iSelect the step of 4 information nodes as follows, wherein i be 0 o'clock from step 2):

1) newly-increased check-node C _iWith a last coding nodes P _I-1Link to each other, thus with original 10000 information node I ₀, I ₁..., I ₉₉₉₉, (i-1) individual coding nodes P ₀, P ₁..., P _I-1, (i-1) individual check-node C ₀, C ₁..., C _I-1The bigraph (bipartite graph) that constitutes links to each other, and needs to upgrade C _iDistance to each information node;

2) be newly-increased check-node C _iSelect one to be attached thereto, because the bigraph (bipartite graph) change in topology needs to upgrade newly-increased check-node C from its information node apart from maximum _iTo the distance of each information node, repeating step 2) select to finish up to 4 information nodes;

3) newly-increased check-node C _iWith newly-increased coding nodes P _iLink to each other, formed one by 10000 information node I ₀, I ₁..., I ₉₉₉₉, i coding nodes P ₀, P ₁..., P _i, i check-node C ₀, C ₁..., C _iThe new bigraph (bipartite graph) that constitutes.

Through above 3 steps, generated new bigraph (bipartite graph), check-node and coding nodes all increase by one than former bigraph (bipartite graph), and the limit is also corresponding have been increased.4 information nodes have been chosen for newly-increased check-node simultaneously.

In step 2) in when selecting information node, we will select one of number of degrees minimum when more than one of distance information node farthest, if at this moment during still more than one of the information node of number of degrees minimum, we will select one of them at random.We always preferentially select the information node of number of degrees minimum, and this can make the number of degrees of information node distribute trend evenly.

Described making: because this selection mode only distributes at the number of degrees of some code check R place restricted information node at the limited information node selection mode of certain code check place information node number of degrees distribution, make at the limited information node selection mode of certain code check place information node number of degrees distribution so we are called this mode, the SAR sign indicating number that adopts this mode is called at the limited SAR sign indicating number of certain code check place information node number of degrees distribution.

If each packet has only a bit, in fact the SAR sign indicating number that works in certain speed is exactly that IRA (Irregular Repeat Accumulate) sign indicating number (is seen " Irregular Repeat-Accumulate codes, " Proc.2nd Int.Symp.Turbo codes ﹠amp; Related topics, Sep.2000); If n bit arranged in the packet, the SAR sign indicating number that works in certain speed is n parallel independent IRA sign indicating number just.In other words, the SAR sign indicating number is exactly the IRA sign indicating number of rate-compatible in essence.

The channel parameter of awgn channel is portrayed with σ, the standard variance of its expression interchannel noise.The IRA sign indicating number of certain speed, given its information node number of degrees distribution table is shown λ (x), then a channel parameter threshold value of this yard correspondence σ ^*, when channel parameter σ less than this thresholding, can the guarantee information reliable transmission; And work as channel parameter σ greater than this thresholding, information then can not reliable transmission.Under special speed R, optimize λ (x) usually so that IRA sign indicating number thresholding σ ^*As far as possible big also just makes σ ^*It is little that corresponding channel capacity is tried one's best, and then make the channel capacity of approaching that R tries one's best.The optimization problem of λ (x) all was discussed in a lot of documents, has been provided some simultaneously and approached the distribution of channel capacity.

For the IRA sign indicating number, it is very different that the information node number of degrees that approach channel capacity that obtain on each speed distribute, so for the SAR sign indicating number, along with increasing of encoded packets, code check descends, the number of degrees of information node distribute and constantly change, and each code check place all keeps the information node number of degrees to distribute approaching channel capacity being difficult to.We propose a kind of suboptimum but the method that is easy to realize, promptly only guarantee at some special speed R place, the information node number of degrees distribute and reach the distribution that approaches channel capacity, and the number of degrees of other speed place information node distribute relevant with concrete implementation, do not impose any restrictions.This selection mode only distributes at the number of degrees of some code check R place restricted information node.

Requiring the information node number of degrees to be distributed in code check in the enforcement is the distribution λ (x) that can reach in 0.5 o'clock after optimizing in document " IrregularRepeat-Accumulate Codes " table 3, each parameter such as table 2 of this distribution.

All the other parameters	λ 2	λ 3	λ 7	λ 8	λ 18	λ 20	λ 55	λ 58
									0	0.0577128	0.117057	0.2189922	0.0333844	0.2147221	0.0752259	0.0808676	0.202038

Table 2

Transmitting terminal need distribute according to the number of degrees of table 2 and utilize the PEG algorithm to generate a code check in advance to be 0.5 bigraph (bipartite graph), 10000 information nodes to be arranged, 10000 check-nodes, 10000 coding nodes among the figure.Table 3 is the number of the information node of each number of degrees.

The number of degrees	2	3	7	8	18	20	55	58
									Number	2328	3147	2523	337	962	303	119	281

Table 3

Know that by table 2 the information node mean number that links to each other with each check-node is 8.

This mode is newly-increased check-node C _iSelect the method for information node as follows:

1) when code check more than or equal to 0.5 the time, promptly at the encoded packets number that sends during smaller or equal to 10000, encoder is newly-increased check-node C according to the annexation of the bigraph (bipartite graph) that generates in advance _iSelect a _iIndividual information node, a _iDepend on the annexation that generates in the bigraph (bipartite graph);

2) when code check less than 0.5 the time, promptly at the encoded packets number that sends during greater than 10000, encoder will use and make the approximate equally distributed mode of the information node number of degrees for increasing check-node C newly _iSelect a information node, because the information node mean number that links to each other with each check-node in the bigraph (bipartite graph) of Sheng Chenging is 8 in advance, so a gets 8 herein.

Claims (3)

1. non-rate code coding/decoding method that is suitable for additive white gaussian noise channels, comprise coding method and interpretation method, it is characterized in that coding method is as follows, consider that the coding transmitting terminal will send m packet, form by the plurality of data bit in each packet, each packet inside comprises that a cyclic redundancy check (CRC) code is used for decoder and judges whether decoding is successful, d ₀, d ₁..., d _j..., d _M-1Represent each packet respectively, subscript j is the numbering of packet, t _iThe presentation code bag, wherein i is the numbering of encoded packets, the coding transmitting terminal at first sends m packet the information bit part that forms systematic code, produces encoded packets t then as follows _i:

1) at first by making the approximate equally distributed information node selection mode of the information node number of degrees or making the information node selection mode generation m dimension two element field vector { G that distributes limited at certain code check place information node number of degrees _Ji, m dimension two element field vector { G _JiIn for the number of the element of " 1 " be a _i

2) m dimension two element field vector { G _JiIn element G _JiValue be " 0 " or " 1 ", element G _JiBe " 1 " that then the packet that is numbered j of its correspondence is with selected, the packet that these are selected makes to obtain after mould 2 adds and be worth s by bit _i, can be expressed as:

$s_i=\underset{j=0}{\overset{m-1}{\mathrm{\Σ}}}{d}_j{G}_{\mathrm{ji}},i=\mathrm{0,1,2}...$

3) by accumulator with s _iWith a last encoded packets t _I-1Make mould 2 and obtain new encoded packets t by bit _i, be expressed as following formula:

t _i=t _I-1+ s _i, t wherein _-1=0

Transmitting terminal produces encoded packets endlessly according to above rule and informs that up to receiving terminal it stops to send;

Interpretation method comprises the steps:

At first receive packet, begin decoding after m packet finishes receiving:

1) decoder utilizes the cyclic redundancy check (CRC) code in each bag to judge that whether m packet be all correct, if all correct, changes step 5) over to; Otherwise change step 2 over to);

2) decoder receives some encoded packets;

3) because receiving terminal is known each encoded packets t _iCorresponding m dimension two element field vector { G _Ji, so decoder can be accurately at the bigraph (bipartite graph) of this sign indicating number of receiving terminal reconstruct;

4) move decoding algorithm on the bigraph (bipartite graph) of reconstruct in step 3), this decoding algorithm is BP algorithm or improved BP algorithm, and then changes operating procedure 1 over to);

5) decoding finishes, and receiving terminal informs that by feedback channel transmitting terminal stops to send.

2. a kind of non-rate code coding/decoding method that is suitable for additive white gaussian noise channels according to claim 1 is characterized in that the described information node number of degrees that make are similar to equally distributed information node selection mode: be newly-increased check-node C _iSelect the step of a information node as follows, wherein i be 0 o'clock from step 2):

1) newly-increased check-node C _iWith a last coding nodes P _I-1Link to each other, thus with original m information node I ₀, I ₁..., I _M-1, (i-1) individual coding nodes P ₀, P ₁..., P _I-1, (i-1) individual check-node C ₀, C ₁..., C _I-1The bigraph (bipartite graph) that constitutes links to each other, and needs to upgrade C _iDistance to each information node;

2) be newly-increased check-node C _iSelect one to be attached thereto, because the bigraph (bipartite graph) change in topology needs to upgrade newly-increased check-node C from its information node apart from maximum _iTo the distance of each information node, repeating step 2) select to finish up to a information node;

3) newly-increased check-node C _iWith newly-increased coding nodes P _iLink to each other, formed one by m information node I ₀, I ₁..., I _M-1, i coding nodes P ₀, P ₁..., P _i, i check-node C ₀, C ₁..., C _iThe new bigraph (bipartite graph) that constitutes;

When selecting information node step 2), we will select one of number of degrees minimum when more than one of distance information node farthest, if at this moment during still more than one of the information node of number of degrees minimum, we will select one of them at random.

3. a kind of non-rate code coding/decoding method that is suitable for Gaussian channel according to claim 1, it is characterized in that described making: if the expectation information node number of degrees reach distribution λ (x) when being distributed in code check for R at the limited information node selection mode of certain code check place information node number of degrees distribution, it is the bigraph (bipartite graph) that R, the information node number of degrees are distributed as λ (x) that the coding transmitting terminal will use the PEG algorithm to generate a code check in advance, is newly-increased check-node C then _iSelect the method for information node as follows:

1) when code check during more than or equal to R, encoder is newly-increased check-node C according to the annexation of the bigraph (bipartite graph) that generates in advance _iSelect a _iIndividual information node;

2) when code check during less than R, encoder will use and make the approximate equally distributed information node selection mode of the information node number of degrees be newly-increased check-node C _iSelect a information node.

Images