CN111262590B - Underwater acoustic communication information source and channel joint decoding method - Google Patents
Underwater acoustic communication information source and channel joint decoding method Download PDFInfo
- Publication number
- CN111262590B CN111262590B CN202010070406.2A CN202010070406A CN111262590B CN 111262590 B CN111262590 B CN 111262590B CN 202010070406 A CN202010070406 A CN 202010070406A CN 111262590 B CN111262590 B CN 111262590B
- Authority
- CN
- China
- Prior art keywords
- source
- sequence
- information
- symbol
- decoding
- 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
Images
Classifications
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
- H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Error Detection And Correction (AREA)
Abstract
The invention discloses a joint decoding method of an underwater acoustic communication information source channel, which comprises the following steps: acquiring information data sent by a coding end, demodulating the information data, and determining a receiving symbol sequence and an information source contribution degree, wherein the information source contribution degree is the ratio of redundancy of a sending information source sequence to a redundancy mean value; generating a joint decoding tree by using the recursive chaotic model according to the coding parameters of the coding end, and calculating part of coding symbol sequences corresponding to branches in the decoding tree and part of information sequences corresponding to decoding paths in the decoding tree layer by layer; calculating branch metric values of an information sequence of a decoding tree by adopting an information source channel combined maximum posterior probability criterion and combining noise energy, information source prior information, information source contribution degree and a received symbol sequence; and calculating the path metric and the value corresponding to each branch in the decoding tree, and selecting the branch corresponding to the maximum path metric and value to obtain a decoding information sequence. The method of the invention makes full use of the correlation between the information source symbols to correct the channel decoding and improve the decoding performance of the underwater acoustic communication.
Description
Technical Field
The application relates to the technical field of underwater acoustic communication, in particular to an underwater acoustic communication information source and channel joint decoding method.
Background
In a traditional underwater acoustic communication system, a transmitting end carries out information source coding and channel coding on a transmitting information source sequence in sequence and transmits the transmitting information source sequence to a receiving end, and the receiving end carries out channel decoding and information source decoding in sequence according to a receiving symbol sequence to obtain an estimation corresponding to the transmitting information source sequence.
Moreover, due to the characteristics of the underwater acoustic communication system, in actual underwater acoustic communication, a certain correlation generally exists among the information source symbols in the transmission information source sequence, when one information source symbol a is determined, the probability that each information source symbol in the information source symbol set appears as a subsequent symbol of the information source symbol a is distributed in a certain rule, and the distribution depends on the value of the information source symbol a, namely, the correlation among the information source symbols. Redundancy exists in the source sequence due to inter-source correlation, and such redundancy is difficult to remove by conventional source coding methods.
In the prior art, in order to fully utilize the correlation between source sequences, a commonly used decoding method includes: a joint decoding method based on grid and a joint decoding method in iterative form. On one hand, when a large source symbol set is processed, the two decoding methods are difficult to be practically applied due to high algorithm complexity.
On the other hand, because the relative redundancy size contained in the actually transmitted information source sequence is different, the two methods cannot adjust the proportion of the information source prior information in the decoding process, and the performance of the joint decoding method is difficult to guarantee under the condition of transmitting all possible information source sequences.
Disclosure of Invention
The purpose of this application lies in: the correlation between information source symbols is counted by historical sending information, common information and the like, the correlation between the information source symbols is fully utilized in the decoding process of the underwater acoustic communication, channel decoding is corrected, the proportion of the channel transfer probability and the information source transfer probability in decoding measurement is adjusted by utilizing the receiving signal-to-noise ratio and the redundancy of a sending information source sequence, and therefore the decoding performance of the underwater acoustic communication is improved.
In order to achieve the above object, the present invention provides a method for joint decoding of underwater acoustic communication source and channel, wherein the decoding method is suitable for decoding of a communication system in which the source coding adopts fixed-length source coding and the channel coding adopts recursive chaotic codes, and the method comprises:
acquiring information data sent by a coding end, demodulating the information data, and determining a receiving symbol sequence and an information source contribution degree, wherein the information source contribution degree is the ratio of redundancy of a sending information source sequence to a redundancy mean value;
generating a joint decoding tree by using the recursive chaotic model according to the coding parameters of the coding end, and calculating part of coding symbol sequences corresponding to branches in the decoding tree and part of information sequences corresponding to decoding paths in the decoding tree layer by layer;
calculating branch metric values of an information sequence of a decoding tree by adopting an information source channel combined maximum posterior probability criterion and combining noise energy, information source prior information, information source contribution degree and a received symbol sequence;
and calculating the path metric and the value corresponding to each branch in the decoding tree, and selecting the branch corresponding to the maximum path metric and value to obtain a decoding information sequence.
As an improvement of the above method, the method further comprises: calculating a redundancy mean value; the method specifically comprises the following steps:
establishing a source sequence database comprising a plurality of training source sequences, wherein the training source sequences comprise: common information source sequence and historical information source sequence; an information source symbol set is arranged in the information source sequence database, the information source symbol set comprises a plurality of information source symbols, the training information source sequence is formed by combining the information source symbols, and each training information source sequence in the information source sequence database corresponds to a transmission probability;
calculating information source prior information in an information source sequence database;
calculating a redundancy mean value in the information source sequence database;
and storing the information source prior information and the redundancy mean value to an encoding end and a decoding end of a communication system.
As an improvement of the above method, the calculating the source prior information in the source sequence database specifically includes:
counting any source symbol in the source symbol set as a first symbol s 'of a sample source sequence'1Then, the head symbol s'1Initial probability P ' (s ') in the source symbol set '1) Wherein, the sample signal source sequence is any one of the training signal source sequences;
counting the previous symbol s 'in the sample source sequence'k-1Is a conditional symbol, and is the current symbol s'kFor a transmit symbol, an initial conditional probability P "(s'k|s′k-1) Wherein k is an index of a symbol in the sample source sequence, k is 2, 3, and the conditional symbol and the transmission symbol are any source symbol in the source symbol set;
according to the initial probability P '(s'1) The initial conditional probability P '(s'k|s′k-1) And the code length of the fixed-length source coding is calculated, and the source priori information P (s'1) And P (s'k|s′k-1)：
Where β is an empirically selected parameter, lsAnd the code length of the fixed-length source coding is obtained.
As an improvement of the above method, calculating a redundancy mean value in the source sequence database specifically includes:
specifying a transmission probability of an f-th training source sequence in the source sequence database
Calculating the f-th training information source sequenceRedundancy R of columnsf：
In the formula, KfFor the length of the f-th source sequence,
for the first symbol in the f-th training source sequence,
for the kth source symbol in the f-th training source sequence, K is 2, 3f，
And
information source prior information of a source symbol in the f-th training information source sequence;
according to the redundancy RfAnd transmission probability
Calculating the redundancy mean
Wherein, F is the total number of training source sequences in the source sequence database.
As an improvement of the above method, the calculating a branch metric of an information sequence of a coding tree by using an information source channel combined maximum a posteriori probability criterion in combination with noise energy, information source prior information, information source contribution and a received symbol sequence specifically includes:
calculating an energy mean of the received symbol sequence from the received symbol sequence
Receiving a noise signal, calculating an estimated noise energy E 'of the noise signal'nAnd then estimate noise energy E'nNormalized noise energy E ofnComprises the following steps:
receiving a symbol sequence Y 'according to a part of the received symbol sequence corresponding to a d-layer branch in the coding tree'dAnd the coding symbol sequence X ' corresponds to the part of the coding symbol sequence X ' corresponding to the j branch in the d layer in the coding tree 'd，jAnd said normalized noise energy EnCalculating the branch path channel transition metric L'channel，d，j：
Of formula (II) to'channel，d，jTransferring metrics, x ', for a jth branch path channel in a d-th layer of the coding tree'd，j，iIs the partially encoded symbol sequence X'd，jThe ith element of (1), y'd，iIs the partial receive symbol sequence Y'dMiddle and x'd，j，iA corresponding element; n is a radical ofdReceiving sequence Y 'for a portion of a coding tree corresponding to a d-th branch'dLength of (d);
according to a partial information sequence M 'corresponding to the jth branch in the d layer of the coding tree'd，jCalculating j < th > branch path source transition metric L 'in d < th > layer in the coding tree'source，d，jWeighting by using the information source contribution degree, calculating the sum of the weighted result and the branch path channel transfer measurement, and recording the sum as the branch measurement value L of the j branch in the d layer of the decoding treed，jSaid branch metric value Ld，jThe calculation formula of (2) is as follows:
Ld，j＝L′channel，d，j+αL′source，d，j
where alpha is the contribution of the source, R is the redundancy of the transmitted source sequence,
is the redundancy mean.
As an improvement of the above method, the step of calculating the redundancy of the transmission source sequence specifically includes:
the sending source sequence S ═ (S)1，…，sK)，skThe information source is matched; k is more than or equal to 1 and less than or equal to K; the redundancy R is:
where K is the length of the transmission source sequence.
As an improvement of the above method, the branch path source transition metric L 'of the jth branch in the d layer of the coding tree is calculated'source，d，jThe method specifically comprises the following steps:
judging a partial information sequence M 'corresponding to the jth branch in the d layer in the decoding tree'd，jLength of (1)dWhether the code length is the fixed length information source codingsIf not, setting a jth branch path source transition metric L 'in a d-th layer of the decoding tree'source，d，j0, otherwise, for partial information sequence M'd，jCarrying out information source decoding to obtain a part of information source sequence corresponding to the jth branch in the d layer of the decoding tree
Wherein, td＝Id/lsIs the partial information sequence M'd，jCorresponding partial source sequence S'd，jLength of (2), calculation ofSource transition metric L 'of jth branch path in d-th layer in decoding tree'source，d，j；
Wherein, the j ' th branch path source transition metric L ' in the d layer of the coding tree 'source，d，jThe calculation formula of (2) is as follows:
of formula (II) s'd，j，1Is the partial source sequence S'd，jThe first symbol in (1) is,
and
is a partial source sequence S'd，jThe last two source symbols.
The invention has the beneficial effects that:
1. the method comprises the steps of firstly constructing an information source sequence database according to a common information source sequence, a historical transmission sequence and the like, counting information source prior information by using a training sequence database, meanwhile, calculating an average value of redundancy of the information source sequence in the information source sequence database, and storing the average value to an underwater acoustic communication system. At a sending end of a communication system, calculating the contribution degree of an information source according to the redundancy of a sending information source sequence and the average value of the redundancy of the information source sequence in an information source sequence database and sending the contribution degree of the information source to a decoding end; then, the sending information source sequence is sent to an underwater acoustic channel through fixed-length information source coding and recursive chaotic coding in sequence; the decoding end adopts an information source channel combined maximum posterior probability criterion to decode; in the decoding process, along with the expansion of a decoding tree, carrying out source decoding on the information sequence corresponding to each path; meanwhile, the channel transition probability and the information source transition probability, namely information source prior information, are synthesized, and the decoding measurement of each path is calculated; the information source decoding and the channel decoding are combined, information source redundancy is fully utilized to resist channel errors, and the decoding performance of the underwater acoustic communication system is further improved;
2. the method can obtain better error rate performance than the traditional separation decoding under the condition that correlation exists between the information sources, also solves the problem of high algorithm complexity of the existing method, and has higher practical value by utilizing the receiving signal-to-noise ratio and the redundancy of the transmitted information source sequence and adjusting the proportion of the channel transfer probability and the information source transfer probability in the decoding measurement in practical application.
Drawings
FIG. 1 is a schematic flow chart of a method for joint source-channel decoding of underwater acoustic communications according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a joint source-channel decoding process for underwater acoustic communications according to an embodiment of the present invention;
FIG. 3 is a diagram of a decoding tree, according to one embodiment of the present invention;
FIG. 4 is a schematic block diagram of an underwater acoustic communication system according to an embodiment of the present invention;
FIG. 5 is a decoding test simulation diagram according to one embodiment of the present invention.
Detailed Description
In order that the above objects, features and advantages of the present application can be more clearly understood, the present application will be described in further detail with reference to the accompanying drawings and detailed description. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, however, the present application may be practiced in other ways than those described herein, and therefore the scope of the present application is not limited by the specific embodiments disclosed below.
The method of the invention is applied to an underwater acoustic communication system which comprises an encoding end (namely a transmitting end) and a decoding end (namely a receiving end).
As shown in fig. 1, the present invention provides a method for joint decoding of underwater acoustic communication source and channel, wherein the decoding method is suitable for decoding of an underwater acoustic communication system in which the source coding adopts fixed-length source coding and the channel coding adopts recursive chaotic codes, and the decoding method comprises:
specifically, a common information source sequence, a historical transmission information source sequence and the like are utilized to construct an information source sequence database, and meanwhile, the transmission probability corresponding to each training sequence is set according to the experience of actual transmission
The common information source sequence and the historical information source sequence are used as known information source sequences in an information source sequence database and are recorded as training information source sequences. Therefore, for the generated source sequence database, the length of each training source sequence, the source symbol composition and other parameters are known values.
In the present embodiment, the transmission source sequence S is set to (S ═ S)1，s2，s3，s4) Contains 4 information source symbols, which are coded by Unicode and have a fixed length code length ls16, the corresponding transmission information sequence M is 1001011101011110010111100011100001100010101100010110101101001001]。
Further, recording a common information source sequence and a historical information source sequence as training information source sequences to form an information source sequence database, wherein an information source symbol set is arranged in the information source sequence database, the information source symbol set comprises a plurality of information source symbols, the training information source sequence is formed by combining a plurality of information source symbols, each training information source sequence in the information source sequence database corresponds to a transmission probability, and in the step 10, calculating information source prior information in the information source sequence database specifically comprises:
step 11, counting any source symbol in the source symbol set as a head symbol s 'of the sample source sequence'1S 'is the first symbol'1Initial probability P ' (s ') in source symbol set '1) Wherein, the sample signal source sequence is any training signal source sequence;
step 12, counting the previous symbol s 'in the sample source sequence'k-1Is a conditional symbol, and is the current symbol s'kTo transmit a symbol, an initial conditional probability P ' (s ') of the symbol in a source symbol set is transmitted 'k|s′k-1) Wherein k is a label of a symbol in the sample source sequence, and k is 2, 3.. the conditional symbol and the transmission symbol are any source symbol in the source symbol set;
specifically, according to the transmission probability of each sample information source sequence, one sample information source sequence is randomly extracted from the information source sequence database at each time, and each information source symbol is counted as a first information source symbol s'1Initial probability P '(s'1) And determining the previous symbol s'k-1K symbols s 'in time source sequence'kInitial conditional probability P "(s'k|s′k-1)。
The information source sequence database is assumed to comprise three training information source sequences which are sequentially as follows: AA. ABCD and BAD, where the source symbol set is { a, B, C, D }, the initial probability P "(a) ═ 2/3 of the source symbol a, and the initial conditional probability P" (a | B) ═ 1/2, i.e. the probability that the previous letter is B and the next letter is a in the sequence.
And repeating the process until each training information source sequence in the information source sequence database is effectively counted.
Step 13, according to the initial probability P '(s'1) Initial conditional probability P '(s'k|s′k-1) And fixing the code length of the information source code, and calculating the information source prior information, wherein the calculation formula of the information source prior information is as follows:
where β is an empirically selected parameter, lsThe code length of the fixed-length source coding is adopted.
In this embodiment, an empirically estimated selected parameter β is set to 0.001, and a fixed-length code length l is setsThus, the source order can be calculated as 16Source prior information of the column database.
In this embodiment, the transmission source sequence S ═ (S)1，s2，s3，s4) The source prior information P of each source symbol is shown in table 1:
TABLE 1
| P(s1) | P(s2|s1) | P(s3|s2) | P(s4|s3) |
| 2.03×10-5 | 8.35×10-4 | 7.51×10-4 | 9.04×10-4 |
Further, in step 10, calculating a redundancy mean value in the source sequence database specifically includes:
when the information source sequence database is generated, the transmission probability of the f-th training information source sequence in the information source sequence database is appointed according to practical application experience
And (3) calculating the redundancy of the f-th training signal source sequence, wherein the redundancy calculation formula is as follows:
in the formula, KfFor the length of the f-th source sequence,
for the first symbol in the f-th training source sequence,
for the kth source symbol in the f-th training source sequence, K is 2, 3f，
And
information source prior information of a source symbol in the f-th training information source sequence;
according to redundancy and transmission probability
And calculating a redundancy mean value, wherein a calculation formula of the redundancy mean value is as follows:
wherein, F is the total number of training source sequences in the source sequence database, and F is 1, 2.
Specifically, through the calculation process, the redundancy of the training information source sequence in the information source sequence database is calculated one by one, and then the transmission probability is combined
The mean value of the redundancy can be calculated
Information source prior information and redundancy mean value
Stored together in a communication systemThe sending end and the receiving end. In this embodiment, the mean of redundancy
Step 20, obtaining information data sent by a coding end, demodulating the information data, and determining a received symbol sequence and an information source contribution degree alpha, wherein the information source contribution degree alpha is a ratio of redundancy of a sent information source sequence to a redundancy mean value;
specifically, in this embodiment, the transmission source sequence S is (S)1，s2，s3，s4) Firstly, fixed-length information source coding is carried out to obtain an information source sequence M to be sent, then the information source sequence M to be sent is subjected to recursive chaotic coding through a recursive chaotic coder, and the length g of an information segment in the recursive chaotic coding process is set to be 1. After the encoding is completed, the encoding symbol sequence X ═ (X) can be obtained1，x2，...，xN) And N is the total number of the code symbols sent by the coding end. Meanwhile, the coding end performs redundancy calculation on the transmission source sequence, and the corresponding calculation formula is as follows:
in this embodiment, the redundancy R of the transmission source sequence is 4.42, and therefore, the source contribution degree
After the encoding end calculates the information source contribution degree alpha, the information source contribution degree alpha is used as side information, the side information and the encoding symbol sequence X are used as information data, the side information and the encoding symbol sequence X are sent to the decoding end, and the decoding end demodulates the side information and the encoding symbol sequence X to obtain the information source contribution degree alpha and the receiving symbol sequence Y.
Step 30, generating a joint decoding tree by using a recursive chaotic model according to the coding parameters of the coding end, and calculating partial codes corresponding to each branch in the decoding tree layer by layerSymbol sequence X'd，jAnd a partial information sequence M 'corresponding to each decoding path in the decoding tree'd，j，
Wherein the symbol sequence is partially encoded
Coding the partial sequence of the symbol sequence X in the coding tree corresponding to the j branch in the d layer, and N'dReceiving a sequence of symbols Y 'for a portion of the coding tree corresponding to a d-th branch'dIs a partial sequence of the received symbol sequence Y;
specifically, as shown in fig. 2, the present embodiment employs fixed-length source coding and conventional recursive chaotic coding. Firstly, an encoding end carries out encoding, after information data of a to-be-encoded end is sent to a decoding end, the decoding end adopts the same encoding method and generates a decoding tree according to the agreed encoding parameters of the encoding end.
As shown in fig. 3, since the length g of the information segment is 1, each leaf node in the coding tree corresponds to the next layer of branches, which includes two branches, 0 and 1.
Decoding layer by using a decoding tree in the recursive chaotic coding, wherein the generation process of the decoding tree and a part of coding symbol sequence X 'corresponding to the jth branch in the d layer of the decoding tree'd，jAnd a partial information sequence M'd，jThe calculation process of (a) is not described herein again, wherein the path indicated by the solid arrow is a decoding path in the decoding tree.
further, in step 40, since each layer of the coding tree contains 1 bit of the information sequence, the coding tree contains D layers of sub-sections in totalThe method comprises the steps of point collection, wherein each layer of sub-node collection comprises J branches, the J (J) th branch in the D (D1, 2,., D) layer of sub-node collection is taken as a current branch, and the current branch is calculated to correspond to a partial information sequence M'd，jBranch metric value of Ld，jThe method specifically comprises the following steps:
step 41, calculating the energy mean value of the received symbol sequence according to the received symbol sequence
Receiving the noise signal, calculating an estimated noise energy E 'of the noise signal'nAnd estimating noise energy E'nNormalized noise energy E ofnNormalized noise energy EnThe calculation formula of (2) is as follows:
wherein, noise energy E 'is estimated'nCan be calculated by the decoding end at idle time, and the energy mean value
And the decoding end calculates after receiving the received symbol sequence, and the decoding end and the receiving symbol sequence adopt a conventional calculation method.
Step 42, receiving symbol sequence Y 'according to the part of the received symbol sequence corresponding to the d-layer branch in the coding tree'dAnd the coding symbol sequence X 'corresponding to the part of the j branch in the d layer in the coding tree'd，jAnd normalized noise energy EnCalculating branch path channel transfer metrics, wherein the calculation formula of the branch path channel transfer metrics is as follows:
of formula (II) to'channel，d，jMetric, x ', for the jth branch path channel transition in the d-th layer of the coding tree'd，j，iIs a partially coded symbol sequence X'd，jThe ith element of (1), y'd，iIs a partial receive symbol sequence Y'dMiddle and x'd，j，iA corresponding element; n is a radical ofdReceiving sequence Y 'for a portion of a coding tree corresponding to a d-th branch'dLength of (d);
step 43, according to the partial information sequence M 'corresponding to the jth branch in the d layer of the decoding tree'd，jCalculating j branch path source transition metric L 'in d layer of decoding tree'source，d，jWeighting by using the information source contribution degree, calculating the sum of the weighted result and the branch path channel transfer measurement, and recording the sum as the branch measurement value L of the j branch in the d layer of the decoding treed，jBranch metric value Ld，jThe calculation formula of (2) is as follows:
Ld，j＝L′channel，d，j+αL′source，d，j
in the formula, α is a source contribution degree calculated by the transmitting end and transmitted to the receiving end.
Further, a branch path source transition metric L 'is calculated'source，d，jThe method specifically comprises the following steps:
judging a partial information sequence M 'corresponding to the jth branch in the d layer in the decoding tree'd，jLength of (1)dCode length l of fixed length source codingsIf not, setting the jth branch path information source transition metric L 'in the d layer of the decoding tree'source，d，jIf 0, for partial information sequence M'd，jCarrying out information source decoding to obtain a part of information source sequence corresponding to the jth branch in the d layer in a decoding tree
Calculating branch path source transition metric L'source，d，j。
Wherein, tdIs partial source sequence S 'corresponding to the current branch'd，jThe number of source symbols involved (sequence length). Since each layer of the coding tree corresponds to one bit in the information sequence, the partial information sequence M'd，jLength of (1)d，jIs equal to the current decoding tree level d, thus td＝d/ls。
Wherein the current branch path source transition metric L'source，d，jThe calculation formula of (2) is as follows:
of formula (II) s'd，j，1Is a partial source sequence S'd，jThe first symbol in (1) is,
and
is a partial source sequence S'd，jMiddle and last two source symbols
Specifically, when the partial information sequence M'd，jIs not the code length l of the fixed-length source codingsInteger multiples of (d ≠ t)dls，t d1, 2, the code length l in this examplesAt this time, 16 ', a new source symbol, that is, a partial information sequence M ' cannot be obtained by source decoding 'd，jDoes not contain a complete, new source symbol.
At this time, no source decoding is performed, and the branch path source transition metric L 'corresponding to the current branch is used'source，d，j0 and calculates the current branch metric value Ld，j(including only branch path channel transfer metrics L'channel，d，j)。
If the partial information sequence M'd，jLength of (1)d，jD is the code length lsAnd integral multiple of the first integer, carrying out source decoding.
When partial information sequence M'd，jLength of (1)d，j＝d＝lsWhen 16, i.e. the decoding tree is extended to layer 16, the length of the partial information sequence is 16 bits, and exactly one complete source symbol s 'can be obtained by source decoding'd，j，1Performing source decoding on part of information sequence corresponding to each path to obtain part of information sequence corresponding to each pathColumn S'd，j＝(s′d，j，1) Calculating branch path channel transition metric L 'by above'channel，d，jAnd source transfer metric L'source，d，jAnd calculating the current branch metric value Ld，j。
When partial information sequence M'd，jLength of (1)d，j＝d＝2lsWhen the signal sequence is 32, namely a decoding tree is expanded to a 32 th layer, a part of the signal sequence obtained by signal source decoding comprises two signal source symbols, namely S'd，j＝(s′d，j，1，s′d，j，2)。
Calculating branch path source transition metric L 'of current branch in a formula'source，d，jAnd a branch path channel transition metric L'channel，d，jTaking the source contribution degree alpha as a branch path source transition metric L'source，d，jAnd calculating the current branch metric value Ld，j。
And step 50, calculating the path metric and the value corresponding to each branch in the decoding tree, and selecting the branch corresponding to the maximum path metric and value to obtain a decoding information sequence.
Specifically, calculating path metrics and values corresponding to each branch in the decoding tree, sorting according to the path metrics and values, reserving the B paths with the maximum path metrics and values, judging whether the number of layers of the decoding tree is equal to the length of a transmission source sequence, if not, continuing to expand the decoding tree for the reserved decoding paths and executing the step 40; if yes, selecting the path with the maximum path metric and value, and outputting the information sequence corresponding to the path as a decoding result.
When performing decoding tree expansion, in order to improve decoding efficiency, reduce unnecessary operations, and also consider the fault-tolerant rate in the decoding process, B (B: 512) branches with the largest path metric sum value are reserved and the branches with smaller path metric sum values are deleted each time the decoding tree is expanded.
Further, step 50 specifically includes:
step 51, measure the branch value L of the current branchd，jBy using an accumulation algorithm, accumulatingThe path measurement sum value of the upper layer branch is sorted, B branches with larger values are selected and reserved to form a maximum branch set, and a part of information sequence M corresponding to each branch in the maximum branch setd，j' recording as a decoding segment, judging whether the current branch is the last layer branch in the decoding tree, if so, executing step 52, if not, performing decoding tree expansion on the reserved maximum branch set, and repeating step 40;
and step 52, selecting the branch corresponding to the maximum path metric sum value in the maximum branch set as a decoding path, and outputting a decoding segment corresponding to the decoding path as a decoding result.
Specifically, through the above process, the branches in the decoding tree are expanded layer by layer until the decoding tree is expanded to the last layer, the maximum value of the path metric sum value in the branch of the last layer is selected, the path corresponding to the path metric sum value is used as the decoding path, and the decoding segment corresponding to the decoding path is output as the decoding result.
As shown in fig. 4, in order to verify the decoding performance of the decoder in this embodiment, an underwater acoustic communication system, an encoder and a decoder are set up, the encoding end adopts a fixed-length source encoding and recursive chaotic encoding method for encoding, the decoding end adopts a source-channel joint decoding, and compares the bit error rate with the existing underwater acoustic communication system adopting a traditional separation decoding method, as shown in fig. 5, under the condition that the signal-to-noise ratio is the same, the bit error rate (curve 501) corresponding to this embodiment is obviously better than the bit error rate (curve 502) corresponding to the existing method.
The technical scheme of the application is explained in detail in combination with the drawings, and before information transmission is carried out, a common information source sequence, a historical transmission information source sequence and the like are used for constructing an information source sequence database. Modeling the information source into a first-order Markov model, and counting information source prior information by using an information source sequence database. And calculating the redundancy of each information source sequence in the information source sequence database by using an information source sequence redundancy estimation formula, and calculating a mean value of the redundancy of the training sequence. And storing the statistical result of the information source prior information and the redundancy mean value of the training sequence to an underwater acoustic communication system for use when information is transmitted.
The information source coding technology can adopt any existing fixed length coding technology, and meanwhile, the information source contribution degree is calculated. And the information sequence obtained after the information source coding is subjected to recursive chaotic code coding to obtain a coding symbol sequence, and the coding symbol sequence is subjected to transmission sequence planning and signal modulation and then is transmitted to the underwater acoustic channel together with the information source contribution degree.
The decoding end firstly decodes to obtain the information source contribution degree, and then adopts the information source channel joint maximum posterior probability decoding based on the decoding tree. And traversing all possible information sequences bit by bit (layer), reproducing the encoding process to obtain corresponding encoding symbol sequences, comparing the encoding symbol sequences with the received symbol sequences, and selecting the most possible information sequences as decoding results.
The steps in the present application may be sequentially adjusted, combined, and subtracted according to actual requirements.
The units in the device can be merged, divided and deleted according to actual requirements.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and are not limited. Although the present invention has been described in detail with reference to the embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (4)
1. A united decoding method of underwater acoustic communication source channel is suitable for the decoding of a communication system, wherein the source coding adopts fixed-length source coding, and the channel coding adopts recursive chaotic code, and is characterized in that the method comprises the following steps:
acquiring information data sent by a coding end, demodulating the information data, and determining a receiving symbol sequence and an information source contribution degree, wherein the information source contribution degree is the ratio of redundancy of a sending information source sequence to a redundancy mean value;
generating a joint decoding tree by using the recursive chaotic model according to the coding parameters of the coding end, and calculating part of coding symbol sequences corresponding to branches in the decoding tree and part of information sequences corresponding to decoding paths in the decoding tree layer by layer;
calculating branch metric values of an information sequence of a decoding tree by adopting an information source channel combined maximum posterior probability criterion and combining noise energy, information source prior information, information source contribution degree and a received symbol sequence;
calculating path measurement and values corresponding to each branch in a decoding tree, and selecting the branch corresponding to the maximum path measurement and value to obtain a decoding information sequence;
the method further comprises the following steps: calculating a redundancy mean value; the method specifically comprises the following steps:
establishing a source sequence database comprising a plurality of training source sequences, wherein the training source sequences comprise: common information source sequence and historical information source sequence; an information source symbol set is arranged in the information source sequence database, the information source symbol set comprises a plurality of information source symbols, the training information source sequence is formed by combining the information source symbols, and each training information source sequence in the information source sequence database corresponds to a transmission probability;
calculating information source prior information in an information source sequence database;
calculating a redundancy mean value in the information source sequence database;
storing the information source prior information and the redundancy mean value to an encoding end and a decoding end of a communication system;
the calculating of the information source prior information in the information source sequence database specifically includes:
counting any source symbol in the source symbol set as a first symbol s 'of a sample source sequence'1Then, the head symbol s'1Initial probability P ' (s ') in the source symbol set '1) Wherein, the sample signal source sequence is any one of the training signal source sequences;
counting the previous symbol s 'in the sample source sequence'k-1Is a conditional symbol, and is the current symbol s'kWhen transmitting symbols, the transmitting symbols are in the source symbol setIntermediate initial conditional probability P "(s'k|s′k-1) Wherein k is an index of a symbol in the sample source sequence, k is 2, 3, and the conditional symbol and the transmission symbol are any source symbol in the source symbol set;
according to the initial probability P '(s'1) The initial conditional probability P '(s'k|s′k-1) And the code length of the fixed-length source coding is calculated, and the source priori information P (s'1) And P (s'k|s′k-1)：
Where β is an empirically selected parameter, lsThe code length of the fixed-length information source coding is obtained;
calculating a redundancy mean value in the source sequence database, specifically comprising:
specifying a transmission probability of an f-th training source sequence in the source sequence database
Calculating the redundancy R of the f-th training information source sequencef：
In the formula, KfFor the length of the f-th source sequence,
for the first symbol in the f-th training source sequence,
for the kth source symbol in the f-th training source sequence, K is 2, 3f，
And
information source prior information of a source symbol in the f-th training information source sequence;
according to the redundancy RfAnd transmission probability
Calculating the redundancy mean
Wherein, F is the total number of training source sequences in the source sequence database.
2. The method as claimed in claim 1, wherein the step of calculating branch metric values of the information sequence of the decoding tree by using the maximum a posteriori probability criterion of the source channel combination and combining the noise energy, the source prior information, the source contribution degree and the received symbol sequence comprises:
calculating an energy mean of the received symbol sequence from the received symbol sequence
Receiving a noise signal, calculating an estimated noise energy E 'of the noise signal'nAnd then estimate noise energy E'nNormalized noise energy E ofnComprises the following steps:
receiving a symbol sequence Y 'according to a part of the received symbol sequence corresponding to a d-layer branch in the coding tree'dAnd the coding symbol sequence X ' corresponds to the part of the coding symbol sequence X ' corresponding to the j branch in the d layer in the coding tree 'd，jAnd said normalized noise energy EnCalculating a branch path channel transition metric L'channel，d，j：
Of formula (II) to'channel，d，jTransferring metrics, x ', for a jth branch path channel in a d-th layer of the coding tree'd，j，iIs the partially encoded symbol sequence X'd，jThe ith element of (1), y'd，iIs the partial receive symbol sequence Y'dMiddle and x'd，j，iA corresponding element; n is a radical ofdReceiving sequence Y 'for a portion of a coding tree corresponding to a d-th branch'dLength of (d);
according to a partial information sequence M 'corresponding to the jth branch in the d layer of the coding tree'd，jCalculating j < th > branch path source transition metric L 'in d < th > layer in the coding tree'source，d，jWeighting by using the information source contribution degree, calculating the sum of the weighted result and the branch path channel transfer measurement, and recording the sum as a branch measurement value L of the j branch in the d layer in the decoding treed，jSaid branch metric value Ld，jThe calculation formula of (2) is as follows:
Ld，j＝L′channel，d，j+αL′source，d，j
where alpha is the contribution of the source, R is the redundancy of the transmitted source sequence,
is the redundancy mean.
3. The method as claimed in claim 2, wherein the step of calculating the redundancy of the transmission source sequence comprises:
the sending source sequence S ═ (S)1，…，sK)，skThe information source is matched; k is more than or equal to 1 and less than or equal to K; the redundancy R is:
where K is the length of the transmission source sequence.
4. The method of claim 3, wherein the calculating the branch path source transfer metric L 'of the jth branch in the d-th layer of the decoding tree'source，d，jThe method specifically comprises the following steps:
judging a partial information sequence M 'corresponding to the jth branch in the d layer in the decoding tree'd，jLength of (1)dWhether the code length is the fixed length information source codingsIf not, setting a jth branch path source transition metric L 'in a d-th layer of the decoding tree'source，d，j0, otherwise, for partial information sequence M'd，jCarrying out information source decoding to obtain a part of information source sequence corresponding to the jth branch in the d layer of the decoding tree
Wherein, td＝Id/lsIs the partial information sequence M'd，jCorresponding partial source sequence S'd，jCalculating a jth branch path source transition metric L 'in a d-th layer of the coding tree'source，d，j；
Wherein, in the coding treeJth branch path source transition metric L 'in d-th layer'source，d，jThe calculation formula of (2) is as follows:
of formula (II) s'd，j，1Is the partial source sequence S'd，jThe first symbol in (1) is,
and
is a partial source sequence S'd，jThe last two source symbols.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010070406.2A CN111262590B (en) | 2020-01-21 | 2020-01-21 | Underwater acoustic communication information source and channel joint decoding method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010070406.2A CN111262590B (en) | 2020-01-21 | 2020-01-21 | Underwater acoustic communication information source and channel joint decoding method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111262590A CN111262590A (en) | 2020-06-09 |
| CN111262590B true CN111262590B (en) | 2020-11-06 |
Family
ID=70954679
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010070406.2A Active CN111262590B (en) | 2020-01-21 | 2020-01-21 | Underwater acoustic communication information source and channel joint decoding method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111262590B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5559757A (en) * | 1991-12-18 | 1996-09-24 | Catipovic; Josko A. | Spatial diversity processing for underwater acoustic telemetry |
| US6819630B1 (en) * | 2002-09-20 | 2004-11-16 | The United States Of America As Represented By The Secretary Of The Navy | Iterative decision feedback adaptive equalizer |
| CN101252409A (en) * | 2007-04-12 | 2008-08-27 | 中国科学院研究生院 | New algorithm of combined signal source channel decoding based on symbol level superlattice picture |
| CN102739322A (en) * | 2012-06-19 | 2012-10-17 | 哈尔滨工程大学 | Remote underwater acoustic communication method based on soft-demodulation soft-decoding joint iteration |
| CN107181533A (en) * | 2016-03-09 | 2017-09-19 | 中国科学院声学研究所 | A kind of interpretation method of fountain codes and the underwater sound communication transmission method of fountain codes |
| CN109981226A (en) * | 2019-04-16 | 2019-07-05 | 厦门大学 | A kind of combined dispatching decoding algorithm of joint source-channel coding system |
| CN110166386A (en) * | 2019-05-30 | 2019-08-23 | 中国科学院声学研究所 | A kind of underwater sound communication equilibrium interpretation method based on recurrence Constructing Chaotic Code |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5568530B2 (en) * | 2011-09-06 | 2014-08-06 | 日本電信電話株式会社 | Sound source separation device, method and program thereof |
| CN106303582A (en) * | 2016-08-20 | 2017-01-04 | 航天恒星科技有限公司 | A kind of Joint Source Channel decoding method and system |
| DE102017130591B4 (en) * | 2016-12-20 | 2022-05-25 | Hyperstone Gmbh | Method and device for error correction coding based on data compression |
| US10313056B2 (en) * | 2017-02-06 | 2019-06-04 | Mitsubishi Electric Research Laboratories, Inc. | Irregular polar code encoding |
| CN109257136A (en) * | 2017-07-12 | 2019-01-22 | 中国科学院大学 | The two-way coding and decoding method of the JPEG2000 arithmetic code of combined signal source channel and safety |
| CN108155972A (en) * | 2017-12-26 | 2018-06-12 | 厦门大学 | The decoding optimization method of distributed associating signal source and channel system |
-
2020
- 2020-01-21 CN CN202010070406.2A patent/CN111262590B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5559757A (en) * | 1991-12-18 | 1996-09-24 | Catipovic; Josko A. | Spatial diversity processing for underwater acoustic telemetry |
| US6819630B1 (en) * | 2002-09-20 | 2004-11-16 | The United States Of America As Represented By The Secretary Of The Navy | Iterative decision feedback adaptive equalizer |
| CN101252409A (en) * | 2007-04-12 | 2008-08-27 | 中国科学院研究生院 | New algorithm of combined signal source channel decoding based on symbol level superlattice picture |
| CN102739322A (en) * | 2012-06-19 | 2012-10-17 | 哈尔滨工程大学 | Remote underwater acoustic communication method based on soft-demodulation soft-decoding joint iteration |
| CN107181533A (en) * | 2016-03-09 | 2017-09-19 | 中国科学院声学研究所 | A kind of interpretation method of fountain codes and the underwater sound communication transmission method of fountain codes |
| CN109981226A (en) * | 2019-04-16 | 2019-07-05 | 厦门大学 | A kind of combined dispatching decoding algorithm of joint source-channel coding system |
| CN110166386A (en) * | 2019-05-30 | 2019-08-23 | 中国科学院声学研究所 | A kind of underwater sound communication equilibrium interpretation method based on recurrence Constructing Chaotic Code |
Non-Patent Citations (4)
| Title |
|---|
| Hypothesis feedback equalization for M-ary parallel combinatory communication in deep water;Xiujun Shu;《2014 12th International Conference on Signal Processing (ICSP)》;20150122;1649-1653 * |
| 基于信源残留冗余的水声通信联合译码方法;胡承昊;《声学技术》;20190430;第38卷(第2期);313-315 * |
| 水声通信低复杂度最大似然联合均衡译码方法;台玉朋;《声学学报》;20181130;第43卷(第6期);894-904 * |
| 水声通信技术研究进展与技术水平现状;王海斌;《信号处理》;20190930;第35卷(第9期);1441-1449 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111262590A (en) | 2020-06-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106877884B (en) | Polar code decoding method for reducing decoding path splitting | |
| CN1132320C (en) | Optimal soft-output decoder for tail-biting trellis codes | |
| CN1101997C (en) | Method and apparatus for rate determination in communication system | |
| TWI384766B (en) | Method and apparatus for channel encoding and decoding in a communication system using low-density parity-check codes | |
| CN110995278B (en) | Improved polarity code serial elimination list bit overturning decoding method and system | |
| CN109347487B (en) | Bit freezing auxiliary-based polar code SCL decoding method | |
| CN107040262A (en) | A kind of method of the List predicted values of calculating polar code SCL+ CRC decodings | |
| JPH10135850A (en) | Method for converting information bit with error correcting code, encoder and decoder for executing this method | |
| JPH10135849A (en) | Method for converting information bit with error correcting code, encoder and decoder for executing this method | |
| CN107809303B (en) | Polarization code decoding method based on AWGN subchannel optimization | |
| US8081719B2 (en) | Method and system for improving reception in wired and wireless receivers through redundancy and iterative processing | |
| WO2018021925A1 (en) | Polar code encoding with puncturing, shortening and extending | |
| CN114157309B (en) | Polar code decoding method, device and system | |
| CN110336639B (en) | Polar code multilayer coding modulation method based on capacity distribution and application | |
| CN112332864B (en) | Polarization code decoding method and system for self-adaptive ordered mobile pruning list | |
| CN108833052B (en) | Channel polarization decoding path metric value sorting method | |
| CN110061803B (en) | Low-complexity polar code bit interleaving coding modulation method | |
| CN116318185A (en) | Polar code coding and decoding method and low-frequency wireless communication system using same | |
| CN115473537A (en) | Improved ES-SCL decoding method in polarization code | |
| CN111262590B (en) | Underwater acoustic communication information source and channel joint decoding method | |
| CN106788458B (en) | Hard decision-oriented forward and backward estimation method for insertion deletion and substitution errors | |
| CN112803954A (en) | Improved BP List decoding algorithm based on CRC (cyclic redundancy check) segmentation processing | |
| CN116683919A (en) | OSD decoding method based on ordered TEPs of LDPC code | |
| CN110166386B (en) | Underwater acoustic communication balanced decoding method based on recursive chaotic code | |
| CN112039537A (en) | Polarization code decoding method based on right-direction information processing and implementation device |
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 |