CN112332863B - Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite - Google Patents
Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite Download PDFInfo
- Publication number
- CN112332863B CN112332863B CN202011169347.0A CN202011169347A CN112332863B CN 112332863 B CN112332863 B CN 112332863B CN 202011169347 A CN202011169347 A CN 202011169347A CN 112332863 B CN112332863 B CN 112332863B
- Authority
- CN
- China
- Prior art keywords
- decoding
- network modules
- neural network
- signal
- information
- 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
- 238000013528 artificial neural network Methods 0.000 claims abstract description 61
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 19
- 230000003993 interaction Effects 0.000 claims abstract description 12
- 238000004891 communication Methods 0.000 claims description 15
- 230000010287 polarization Effects 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 11
- 230000005540 biological transmission Effects 0.000 claims description 10
- 239000000654 additive Substances 0.000 claims description 9
- 230000000996 additive effect Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 241000169170 Boreogadus saida Species 0.000 claims 1
- 230000000903 blocking effect Effects 0.000 abstract description 7
- 230000006870 function Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 230000004913 activation Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
- H03M13/03—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
- H03M13/05—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using block codes, i.e. a predetermined number of check bits joined to a predetermined number of information bits
- H03M13/13—Linear codes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Error Detection And Correction (AREA)
Abstract
The invention discloses a polar code decoding algorithm, a receiving end and a system under a low signal-to-noise ratio scene of a low-orbit satellite. The algorithm comprises the following steps: s1, demodulating a received polarized signal to obtain a signal stream, and calculating a likelihood value of the signal stream; s2, inputting the likelihood value into a decoder, and outputting a decoding result by the decoder; the decoder comprises at least one decoding link, each decoding link is provided with M BP decoding network modules and M neural network modules, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged between BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of the decoding links in a blocking mode, and output signals of the neural network modules at the tail ends of all decoding links are combined to obtain decoding results. Compared with the traditional SC algorithm, the method reduces decoding delay, improves decoding accuracy and increases neural network training speed.
Description
Technical Field
The invention relates to the technical field of coding and decoding, in particular to a polar code decoding algorithm, a receiving end and a system under a low signal-to-noise ratio scene of a low-orbit satellite.
Background
Since Arikan proposed polarization codes in 2009, a brand new coding and decoding method of polarization codes has been greatly successful. The polarization code shows considerable advantages in terms of complexity of encoding and decoding and bit error rate control of encoding and decoding.
In the decoding algorithm section, the existing common algorithms are mainly a serial cancellation algorithm (Successive Cancellation Algorithm, hereinafter abbreviated as SC algorithm, derived algorithm such as SCL, etc.) and a belief propagation (Belief Propagation, BP) algorithm. The SC algorithm has the advantage of lower computational complexity (according to the analysis of Arikan, the complexity of the SC decoding algorithm is N log N), but the advantages of the polarization code in terms of channel polarization can be reflected only when the code length reaches a certain degree, and the SC algorithm improves the decoding accuracy of the SC algorithm under short codes at the cost of improving the operational complexity, so that the SC algorithm inevitably has the advantages that the parallel algorithm cannot be adopted due to the characteristics of the SC algorithm, and meanwhile, the algorithm is relatively poor in error correction capability; while parallel computation can be adopted in the BP algorithm, the complexity is high and the ideal error rate cannot be achieved easily under the condition of low iteration number.
In recent years, deep learning algorithms and corresponding artificial neural networks are emerging, and the powerful decision capability of deep learning brings new ideas for decoding algorithms. In the context of polarization codes becoming the standard for 5G wireless communications, the improvement of decoding speed by artificial intelligence neural networks has received attention from the industry. However, the piecewise linear neural network decoding model (Partitioned Linear NN Decoding Model, PLNN) is a more desirable choice in the absence of a sufficiently large amount of training data. Piecewise Linear Neural Networks (PLNNs), i.e., neural networks where the activation function employs an analytical linear function. Typical piecewise linear activation functions include ReLU and ReLU family activation functions, maxOut activation functions.
Disclosure of Invention
The invention aims at least solving the technical problems in the prior art, and particularly creatively provides a polar code decoding algorithm, a receiving end and a system under a low signal-to-noise ratio scene of a low-orbit satellite.
In order to achieve the above object of the present invention, according to a first aspect of the present invention, there is provided a polar code decoding algorithm in a low signal-to-noise ratio scenario of a low-orbit satellite, comprising: s1, demodulating a received polarized signal to obtain a signal stream, and calculating a likelihood value of the signal stream; s2, inputting the likelihood value into a decoder, and outputting a decoding result by the decoder; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged between BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of the decoding links in a blocking mode, output signals of the neural network modules at the tail ends of all decoding links are combined to obtain decoding results, and M is a positive integer.
The technical scheme is as follows: the method has the advantages that the method reduces decoding delay and improves decoding accuracy compared with the traditional SC algorithm, the neural network module divides data into a plurality of short code blocks, the training speed can not be improved, the requirement of a training set required by training is reduced, the calculation complexity and the number of layers of the neural network can be adjusted, and the network training is very flexible.
In order to achieve the above object of the present invention, according to a second aspect of the present invention, a receiving end is disclosed, which includes a communication module, a likelihood value calculation module, and a decoder; the likelihood value calculation module is respectively connected with the communication module and the decoder; the communication module receives the polarized signals sent by the transmitting end, demodulates the received polarized signals to form signal streams, and transmits the signal streams to the likelihood value calculation module; the likelihood value calculation module calculates likelihood values of the signal streams and outputs the likelihood values to a decoder; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged between BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of the decoding links in a blocking mode, output signals of the neural network modules at the tail ends of all decoding links are combined to obtain decoding results, and M is a positive integer.
The technical scheme is as follows: the receiving end decodes the received polarized signals, each block of the likelihood values is sequentially input into the corresponding decoding links for decoding in the decoding process, the decoder can process the blocks of the likelihood values in parallel, and the M neural network modules sequentially pass through the blocks of the likelihood values in each decoding link for iterative learning to output judgment results.
In order to achieve the above object of the present invention, according to a third aspect of the present invention, there is disclosed a communication system including a transmitting end and a receiving end; the transmitting end carries out linear conversion on N bit channels to obtain a channel with transmission characteristic polarization, carries out coding treatment on information to be transmitted, carries out CQPSK modulation on the information after the coding treatment, and transmits the information to the receiving end through an additive Gaussian white noise channel; the receiving end demodulates the received polarized signals to form signal streams, and likelihood values of the signal streams are calculated; the receiving end inputs the likelihood value into a decoder, and the decoder outputs a decoding result; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged between BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of the decoding links in a blocking mode, output signals of the neural network modules at the tail ends of all decoding links are combined to obtain decoding results, and M is a positive integer.
The technical scheme is as follows: the system determines information bits and fixed bits of a decoder at a transmitting end, generated signals are transmitted to a receiving end through a CQPSK modulation channel, the receiving end of the system decodes the received polarized signals, each block of likelihood values is sequentially input into a corresponding decoding link for decoding in the decoding process, the decoder can process the blocks of the likelihood values in parallel, the blocks of the likelihood values are sequentially subjected to iterative learning through M neural network modules in each decoding link to output judgment results, compared with the traditional SC algorithm, the system adopts a mode of combining BP algorithm and PLNN, decoding delay is reduced, decoding accuracy is improved, the neural network modules divide data into a plurality of short code blocks, training speed cannot be improved, the requirement of training sets required by training is reduced, the calculation complexity and the number of layers of the neural network can be adjusted, and network training is very flexible.
Drawings
FIG. 1 is a schematic diagram of a decoder according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a BP network according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a BP algorithm network element according to an embodiment of the invention;
fig. 4 is a schematic diagram of an implementation flow of the communication system according to an embodiment of the present invention.
Detailed Description
Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative only and are not to be construed as limiting the invention.
In the description of the present invention, it should be understood that the terms "longitudinal," "transverse," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention.
In the description of the present invention, unless otherwise specified and defined, it should be noted that the terms "mounted," "connected," and "coupled" are to be construed broadly, and may be, for example, mechanical or electrical, or may be in communication with each other between two elements, directly or indirectly through intermediaries, as would be understood by those skilled in the art, in view of the specific meaning of the terms described above.
The invention provides a polar code decoding algorithm under a low signal-to-noise ratio scene of a low-orbit satellite, which comprises the following steps in a preferred implementation mode: s1, demodulating a received polarized signal to obtain a signal stream, and calculating a likelihood value of the signal stream; s2, inputting the likelihood value into a decoder, and outputting a decoding result by the decoder; the structure of the decoder is shown in fig. 1, the decoder comprises at least one decoding link, each decoding link is provided with M BP decoding network modules and M neural network modules, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged between the BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of the decoding links in a blocking mode, output signals of the neural network modules at the tail ends of all decoding links are combined to obtain decoding results, and M is a positive integer.
In this embodiment, a first BP decoding network module, a first neural network module, a second BP decoding network module, a second neural network module, … …, an mth BP decoding network module, and an mth neural network module are sequentially provided on each decoding chain from input to output. All decoding network modules with the same serial numbers on the decoding chains consider that the positions of the decoding network modules are corresponding, for example, the positions of the first BP decoding network modules on all decoding chains are corresponding, an integral BP decoding network is formed among all BP decoding network modules corresponding to the positions, the integral decoding network structure is shown in fig. 2, the total bit number of an information stream is assumed to be 8 in fig. 2, each bit of data occupies one layer of network, the information bit is 4, the fixed bit is 4, and data interaction is needed among all layers of networks, as shown in fig. 2, so that a data interaction branch is arranged among BP decoding network modules corresponding to the positions on different decoding chains.
In this embodiment, as shown in fig. 1, on any decoding link, information output by the BP decoding network module enters a neural network module adjacent below to learn, and an output signal on the decoding link is subjected to iterative learning processing by M neural network modules. The neural network module may select an existing neural network structure.
In this embodiment, a schematic diagram of a schematic network element structure of the BP decoding network module is shown in fig. 3, where an information flow in the BP decoding network is divided into two parts, one part is an information flow from right to left (denoted by L), and the other part is an information flow from left to right (denoted by R).
In this embodiment, preferably, the BP algorithm in the BP network module is:
wherein the function->A. B represents two variables respectively; r is R i,j Information flow representing the ith bit of the jth layer in BP network module to the right, L i,j Information flow representing the ith bit to the left of the jth layer in a BP network module, L i,j+1 Information flow representing the i bit left of the j+1th layer in BP network module,/and>i+2 of the j+1 th layer in the BP network module j-1 Information stream with left bit->I+2 of the j-th layer in BP network module j-1 Information stream with left bit->I+2 of the j-th layer in BP network module j-1 Information flow with right bit, R i,j+1 Information flow indicating the i bit right of the j+1th layer in BP network module,/and>i+2 of the j+1 th layer in the BP network module j-1 With bits to the rightThe information flows, i and j, are positive integers.
In this embodiment, in the initial stage of the BP algorithm, as in the BP decoding network structure shown in fig. 2, the initial value of the input layer of the first BP decoding network module on the decoding chain is a partial data bit of the likelihood value, the initial state of the output layer of the first BP decoding network module is 0, the fixed bit is infinity, and the initial value of the node between the input layer and the output layer is 0. It is known which bits are information bits and which bits are fixed bits are transmitted to the receiving end after the transmitting end polarization processing.
In a preferred embodiment, likelihood values LLR (y) for signal stream y are calculated by the following formula;wherein sigma 2 Representing the noise variance of the transmission signal channels of the transmitting end and the receiving end. In the present embodiment, the noise is preferably but not limited to desirably 0, and the variance is σ 2 The transmission signal channels of the transmitting end and the receiving end are additive white gaussian noise (Additive White Gaussian Noise, AWGN) channels.
In a preferred embodiment, the neural network module includes an input layer, T hidden layers, and an output layer, and the entire neural network can be expressed as: { L 0 ，L 1 ，L 2 ，...，L T ，L T+1 }，L 1 To L T To conceal layer L 0 And L T+1 Respectively an input layer and an output layer; the t-th layer output vector is: o (o) t ＝f t (o t-1 )＝φ t (C t )；Wherein, the variable C t ＝o t-1 ω t +b t ，ω t Weight matrix representing layer t network, b t Bias matrix of layer t network, o t-1 Representing the input vector, t.epsilon.1, T+1]T is a positive integer.
In a preferred embodiment, the weight matrix ω of the layer t network of the neural network module t And bias matrix b t Obtained through back propagation training, the loss function L in the back propagation training is:
wherein k represents the number of samples of the training set of the back propagation training; b i' Signal value representing the i 'th information bit of a sample, i.e. the i' th information bit output from a preceding BP network module,/>Soft decision result representing the i' th information bit of the sample,/and a method for determining the same>i' is a positive integer. Because the length, the fixed position and the like of each neural network module are changed, and the length is shorter, each module can be independently trained, and the training is more beneficial.
In a preferred embodiment, in the decoding link, the output signal of the BP network module is input to the next adjacent neural network module and returned to the input end of the BP network module, the next adjacent neural network module performs learning processing on the input signal to obtain a soft judgment result, and the neural network module inputs the soft judgment result to the next adjacent BP network module.
In a preferred embodiment, the method further comprises step S0: the step S0 is as follows: as shown in fig. 4, the transmitting end performs linear conversion on N bit channels to obtain a channel with polarized transmission characteristics, performs coding processing on information to be transmitted, performs CQPSK modulation on the information after the coding processing, and then transmits the information to the receiving end through an additive white gaussian noise channel, where N is a positive integer.
In a preferred embodiment, the information to be transmitted is encoded using the following formula:wherein x represents information after encoding processing, u represents information to be transmitted, < >> Represents the n-th order kronecker product of G.
In the present embodiment, the principle of the polarization code is a channel polarization theory, and a channel with polarized transmission characteristics can be obtained by performing specific linear conversion on N bit channels. Wherein, the virtual channel with higher transmission quality is used as information bit, the rest bits do not transmit information, and a fixed bit stream is sent. The polarization code structure is expressed in terms of p (N, K), where N is the code length, and N is typically the power of 2 to N, i.e., n=2 n K represents the length occupied by the information bit, the fixed information bit is N-K, the fixed bit is processed by 0 or 1, the code rate of the polarization code isAfter CQPSK modulation, the signal is transmitted to a receiving end through an additive Gaussian white noise (Additive White Gaussian Noise, AWGN) channel for decoding, and the received signal is denoted as y= (1-2 x) +z, wherein z is expected to be 0 variance sigma 2 Is a gaussian white noise of (c). The transmitting end firstly completes source coding, determines information bits and fixed bits of the decoder, and the generated signals are transmitted to the decoder through the AWGN channel by CQPSK modulation.
The invention also discloses a receiving end, in a preferred implementation mode, the receiving end comprises a communication module, a likelihood value calculation module and a decoder; the likelihood value calculation module is respectively connected with the communication module and the decoder; the communication module receives the polarized signals sent by the transmitting end, demodulates the received polarized signals to form signal streams, and transmits the signal streams to the likelihood value calculation module; the likelihood value calculating module calculates likelihood values of the signal stream and outputs the likelihood values to the decoder; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged among BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of all decoding links in a blocking mode, output signals of all decoding link tail end neural network modules are combined to obtain decoding results, and M is a positive integer.
The invention also discloses a communication system, in a preferred embodiment, the system comprises a transmitting end and a receiving end; the transmitting end carries out linear conversion on N bit channels to obtain a channel with polarized transmission characteristics, carries out coding processing on information to be transmitted, carries out CQPSK modulation on the information after the coding processing, and transmits the information to the receiving end through an additive Gaussian white noise channel; the receiving end demodulates the received polarized signals to form signal streams, and likelihood values of the signal streams are calculated; the receiving end inputs the likelihood value into a decoder, and the decoder outputs a decoding result; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, data interaction branches are arranged among BP decoding network modules corresponding to positions on different decoding links, likelihood values are input into the BP decoding network modules at the starting ends of all decoding links in a blocking mode, output signals of all decoding link tail end neural network modules are combined to obtain decoding results, and M is a positive integer.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that: many changes, modifications, substitutions and variations may be made to the embodiments without departing from the spirit and principles of the invention, the scope of which is defined by the claims and their equivalents.
Claims (10)
1. A polar code decoding method under a low signal-to-noise ratio scene of a low-orbit satellite is characterized by comprising the following steps:
s1, demodulating a received polarized signal to obtain a signal stream, and calculating a likelihood value of the signal stream;
s2, inputting the likelihood value into a decoder, and outputting a decoding result by the decoder;
the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, all the decoding network modules with the same serial numbers on the decoding links consider that the positions of the BP decoding network modules are corresponding to each other, data interaction branches are arranged between the BP decoding network modules corresponding to the positions on different decoding links, likelihood value blocks are input into the BP decoding network modules at the starting ends of all the decoding links, the output signals of the neural network modules at the tail ends of all the decoding links are combined to obtain decoding results, and M is a positive integer.
2. The method for decoding a polar code in a low signal-to-noise scenario of a low-orbit satellite according to claim 1, wherein the likelihood value LLR (y) of the signal stream y is calculated by the following formula;
wherein sigma 2 Representing the noise variance of the transmission signal channels of the transmitting end and the receiving end.
3. The method for decoding the polarization code in the low signal-to-noise ratio scene of the low-orbit satellite according to claim 1, wherein the BP algorithm in the BP network module is:
wherein the function isA. B represents two variables respectively; r is R i,j Information flow representing the ith bit of the jth layer in BP network module to the right, L i,j Information flow representing the ith bit to the left of the jth layer in a BP network module, L i,j+1 Information flow representing the i bit left of the j+1th layer in BP network module,/and>i+2 of the j+1 th layer in the BP network module j-1 Information stream with left bit->I+2 of the j-th layer in BP network module j-1 Information stream with left bit->I+2 of the j-th layer in BP network module j-1 Information flow with right bit, R i,j+1 Information flow indicating the i bit right of the j+1th layer in BP network module,/and>i+2 of the j+1 th layer in the BP network module j-1 And the information flow with right bit is positive integer.
4. The method for decoding a polar code in a low signal-to-noise ratio scenario of a low-orbit satellite according to claim 1, wherein the neural network module comprises an input layer, T hidden layers and an output layer, and the output vector of the T th layer is:
o t ＝φ t (C t )；
wherein the variables areC t ＝o t-1 ω t +b t ,ω t Weight matrix representing layer t network, b t Bias matrix of layer t network, o t-1 Representing the input vector, t.epsilon.1, T+1]T is a positive integer.
5. The method for decoding polar codes in a low signal-to-noise ratio scenario in a low-orbit satellite according to claim 4, wherein a weight matrix ω of a layer t network of the neural network module t And bias matrix b t Obtained through back propagation training, the loss function L in the back propagation training is:
wherein k represents the number of samples of the training set of the back propagation training; b i' A signal value representing the i 'th information bit of the sample, i.e. the i' th information bit output from the previous BP network module, b i' ∈{0,1}；Soft decision result representing the i' th information bit of the sample,/and a method for determining the same>i' is a positive integer.
6. The method for decoding a polar code in a low signal-to-noise ratio scenario of a low-orbit satellite according to claim 1, wherein in a decoding link, an output signal of a BP network module is input to a next adjacent neural network module and simultaneously returned to an input end of the BP network module, the next adjacent neural network module learns the input signal to obtain a soft judgment result, and the neural network module inputs the soft judgment result to the next adjacent BP network module.
7. The method for decoding a polar code in a low signal-to-noise scenario of a low-earth-orbit satellite according to claim 1, further comprising step S0:
the step S0 is as follows:
the transmitting end carries out linear conversion on N bit channels to obtain a channel with polarized transmission characteristics, carries out coding processing on information to be transmitted, carries out CQPSK modulation on the information after the coding processing, and then transmits the information to the receiving end through an additive Gaussian white noise channel, wherein N is a positive integer.
8. The method for decoding a polar code in a low signal-to-noise ratio scenario for a low-earth-orbit satellite according to claim 7,
encoding the information to be transmitted by using the following formula:
wherein x represents information after encoding processing, u represents information to be transmitted,represents the n-th order kronecker product of G.
9. The receiving end is characterized by comprising a communication module, a likelihood value calculation module and a decoder; the likelihood value calculation module is respectively connected with the communication module and the decoder;
the communication module receives the polarized signals sent by the transmitting end, demodulates the received polarized signals to form signal streams, and transmits the signal streams to the likelihood value calculation module;
the likelihood value calculation module calculates likelihood values of the signal streams and outputs the likelihood values to a decoder;
the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, all the decoding network modules with the same serial numbers on the decoding links consider that the positions of the BP decoding network modules are corresponding to each other, data interaction branches are arranged between the BP decoding network modules corresponding to the positions on different decoding links, likelihood value blocks are input into the BP decoding network modules at the starting ends of all the decoding links, the output signals of the neural network modules at the tail ends of all the decoding links are combined to obtain decoding results, and M is a positive integer.
10. A communication system comprising a transmitting end and a receiving end;
the transmitting end carries out linear conversion on N bit channels to obtain a channel with transmission characteristic polarization, carries out coding treatment on information to be transmitted, carries out CQPSK modulation on the information after the coding treatment, and transmits the information to the receiving end through an additive Gaussian white noise channel;
the receiving end demodulates the received polarized signals to form signal streams, and likelihood values of the signal streams are calculated; the receiving end inputs the likelihood value into a decoder, and the decoder outputs a decoding result; the decoder comprises at least one decoding link, M BP decoding network modules and M neural network modules are arranged on each decoding link, the BP decoding network modules and the neural network modules are alternately arranged, all the decoding network modules with the same serial numbers on the decoding links consider that the positions of the BP decoding network modules are corresponding to each other, data interaction branches are arranged between the BP decoding network modules corresponding to the positions on different decoding links, likelihood value blocks are input into the BP decoding network modules at the starting ends of all the decoding links, the output signals of the neural network modules at the tail ends of all the decoding links are combined to obtain decoding results, and M is a positive integer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011169347.0A CN112332863B (en) | 2020-10-27 | 2020-10-27 | Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011169347.0A CN112332863B (en) | 2020-10-27 | 2020-10-27 | Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112332863A CN112332863A (en) | 2021-02-05 |
| CN112332863B true CN112332863B (en) | 2023-09-05 |
Family
ID=74296944
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011169347.0A Active CN112332863B (en) | 2020-10-27 | 2020-10-27 | Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112332863B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113162632B (en) * | 2021-04-29 | 2022-08-09 | 东方红卫星移动通信有限公司 | Intelligent QC-LDPC decoding method, decoder and low-orbit satellite communication system |
| CN113316216B (en) * | 2021-05-26 | 2022-04-08 | 电子科技大学 | Routing method for micro-nano satellite network |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107241106A (en) * | 2017-05-24 | 2017-10-10 | 东南大学 | Polarization code decoding algorithm based on deep learning |
| CN109495116A (en) * | 2018-10-19 | 2019-03-19 | 东南大学 | The SC-BP hybrid decoding method of polarization code and its adjustable hardware structure |
| CN109586730A (en) * | 2018-12-06 | 2019-04-05 | 电子科技大学 | It is a kind of based on the polarization code BP decoding algorithm intelligently post-processed |
| CN110212922A (en) * | 2019-06-03 | 2019-09-06 | 南京宁麒智能计算芯片研究院有限公司 | A kind of polarization code adaptive decoding method and system |
| CN110798228A (en) * | 2019-10-29 | 2020-02-14 | 南京宁麒智能计算芯片研究院有限公司 | Polarization code turning decoding method and system based on deep learning |
| CN111106839A (en) * | 2019-12-19 | 2020-05-05 | 北京邮电大学 | Polarization code decoding method and device based on neural network |
| CN111711455A (en) * | 2020-05-26 | 2020-09-25 | 北京邮电大学 | Polarization code BP decoding method based on neural network |
-
2020
- 2020-10-27 CN CN202011169347.0A patent/CN112332863B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107241106A (en) * | 2017-05-24 | 2017-10-10 | 东南大学 | Polarization code decoding algorithm based on deep learning |
| CN109495116A (en) * | 2018-10-19 | 2019-03-19 | 东南大学 | The SC-BP hybrid decoding method of polarization code and its adjustable hardware structure |
| CN109586730A (en) * | 2018-12-06 | 2019-04-05 | 电子科技大学 | It is a kind of based on the polarization code BP decoding algorithm intelligently post-processed |
| CN110212922A (en) * | 2019-06-03 | 2019-09-06 | 南京宁麒智能计算芯片研究院有限公司 | A kind of polarization code adaptive decoding method and system |
| CN110798228A (en) * | 2019-10-29 | 2020-02-14 | 南京宁麒智能计算芯片研究院有限公司 | Polarization code turning decoding method and system based on deep learning |
| CN111106839A (en) * | 2019-12-19 | 2020-05-05 | 北京邮电大学 | Polarization code decoding method and device based on neural network |
| CN111711455A (en) * | 2020-05-26 | 2020-09-25 | 北京邮电大学 | Polarization code BP decoding method based on neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112332863A (en) | 2021-02-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109586730B (en) | Polarization code BP decoding algorithm based on intelligent post-processing | |
| CN112332863B (en) | Polar code decoding algorithm, receiving end and system under low signal-to-noise ratio scene of low orbit satellite | |
| CN110278001B (en) | Polarization code partition decoding method based on deep learning | |
| US20080063104A1 (en) | Search efficient mimo trellis decoder | |
| CN109921803B (en) | High-density linear block code decoding method based on neural network | |
| CN110932734B (en) | Deep learning channel decoding method based on alternative direction multiplier method | |
| CN110113057A (en) | A kind of polarization code decoder using deep learning | |
| CN101119177B (en) | Bit-symbol signal processing method for coherent communication machine | |
| CN114157309A (en) | Polar code decoding method, device and system | |
| CN107896137A (en) | A kind of sort method suitable for the division of polarization code decoding path | |
| CN101420291A (en) | Combined decoding method for network and channel code in relay system | |
| CN110430018B (en) | Bals code transmission method for sliding window with balanced protection | |
| JPH1075271A (en) | Transmitting device and receiving device | |
| CN112332864A (en) | Polar code decoding method and system for self-adaptive ordered mobile pruning list | |
| KR100515472B1 (en) | Channel coding and decoding method and multiple-antenna communication systems performing the same | |
| CN116436567A (en) | Semantic communication method based on deep neural network | |
| CN115483936A (en) | LDPC code decoding method based on neural network in optical fiber communication system | |
| CN113381799A (en) | Low-orbit satellite-ground link end-to-end sparse code multiple access method based on convolutional neural network | |
| CN109194336B (en) | Method, system and device for coding and decoding cascade Spinal code | |
| CN110995277A (en) | Multi-layer neural network assisted penalty dual decomposition channel decoding method | |
| CN107094026B (en) | Graph merging detection decoding method for NB-LDPC coding | |
| CN110601699B (en) | Method for realizing code rate dynamic variable multi-element LDPC code | |
| CN115037315B (en) | Multi-stage flexible self-adaptive SCL pruning method based on Polar codes | |
| CN116418395A (en) | Relay laser communication coding method for FSO system | |
| CN107682122B (en) | Iterative demodulation decoding method of wireless optical communication multilevel coding modulation system |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 401123 39-1, Building 1, No. 64, Middle Huangshan Avenue, Yubei District, Chongqing Patentee after: China Star Network Application Co.,Ltd. Address before: 618 Liangjiang Avenue, Longxing Town, Yubei District, Chongqing Patentee before: Dongfanghong Satellite Mobile Communication Co.,Ltd. |