CN112994704B - Decoding early termination method, storage medium and electronic equipment - Google Patents
Decoding early termination method, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN112994704B CN112994704B CN202110152057.3A CN202110152057A CN112994704B CN 112994704 B CN112994704 B CN 112994704B CN 202110152057 A CN202110152057 A CN 202110152057A CN 112994704 B CN112994704 B CN 112994704B
- Authority
- CN
- China
- Prior art keywords
- decoding
- information
- early termination
- iteration
- value
- 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
- 238000000034 method Methods 0.000 title claims abstract description 68
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000005259 measurement Methods 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims description 18
- 230000002028 premature Effects 0.000 claims description 9
- 230000007306 turnover Effects 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 description 13
- 230000005540 biological transmission Effects 0.000 description 12
- 238000012937 correction Methods 0.000 description 8
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 208000011580 syndromic disease Diseases 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000010363 phase shift Effects 0.000 description 4
- 238000004088 simulation Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000007476 Maximum Likelihood Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 239000000969 carrier Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000010287 polarization Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
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
- H03M13/11—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 using multiple parity bits
- H03M13/1102—Codes on graphs and decoding on graphs, e.g. low-density parity check [LDPC] codes
- H03M13/1105—Decoding
-
- 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
- Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Error Detection And Correction (AREA)
Abstract
The application relates to a decoding early termination method, a storage medium and electronic equipment, wherein the method comprises the following steps: in the iterative decoding process, calculating information metric values of two iterations before and after; when the information measurement value is determined to meet the preset condition, controlling an early termination counter to count; and stopping iteration when the count value of the early termination counter is larger than a preset count threshold value. Therefore, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of small performance loss, so that the decoding delay is reduced.
Description
Technical Field
The present disclosure relates to the field of decoding technologies, and in particular, to a method for early termination of decoding, a storage medium, and an electronic device.
Background
Low density parity check (Low Density Parity Check, LDPC) codes are widely used in modern digital communication systems as a linear block code whose performance approaches shannon's limit. Belief propagation (Belief Propagation, BP) decoding algorithms are widely used in the decoding of low density parity check codes due to their massively parallel processable nature. The basic idea of belief propagation decoding is that messages are exchanged between variable nodes and check nodes, and specifically, after the variable nodes (check nodes) receive messages of all check nodes (variable nodes) connected with the variable nodes, the messages transmitted to the check nodes (variable nodes) are updated, and the updated messages are transmitted to the check nodes (variable nodes) connected with the variable nodes, so that message transmission is performed iteratively.
In the related art, the judging condition of iteration termination mainly comprises the maximum iteration times, the check result of the decoding result of the transmission bit by the check matrix, the error correction result of the decoding sequence by the error correction code or the mutual combination of the three. However, for a scenario with a low Signal-to-Noise Ratio (SNR) (i.e., a scenario with a very low probability of decoding success), the second and third methods in the related art will fail, and the decoding will always be iterated to the maximum number of iterations until it is terminated, but these iterations have no meaning and cause decoding delay.
Disclosure of Invention
Accordingly, in order to solve the above-mentioned problems, it is necessary to provide a decoding early termination method and apparatus, a storage medium, an electronic device, and a decoder that can effectively reduce decoding delay.
A coding early termination method, comprising:
in the iterative decoding process, calculating information metric values of two iterations before and after;
when the information measurement value is determined to meet the preset condition, controlling an early termination counter to count;
and stopping iteration when the count value of the early termination counter is larger than a preset count threshold value.
A computer readable storage medium having stored thereon a decode early termination program that when executed by a processor implements a decode early termination method as described above.
An electronic device comprises a memory, a processor and a decoding early termination program which is stored in the memory and can run on the processor, wherein the decoding early termination method is realized when the processor executes the decoding early termination program.
A decoder comprises a memory, a processor and a decoding early termination program which is stored in the memory and can run on the processor, wherein the decoding early termination method is realized when the processor executes the decoding early termination program.
A coding early termination device, comprising:
the calculation module is used for calculating the information metric value of the two iterations before and after the iterative decoding process;
the determining module is used for controlling the early termination counter to count when the information measurement value meets the preset condition;
and the termination module is used for terminating iteration when the count value of the early termination counter is larger than a preset count threshold value.
A decoder comprises the decoding early termination device.
According to the decoding early termination method and device, the storage medium, the electronic equipment and the decoder, the information metric values of the two iterations before and after are calculated in the iterative decoding process, when the information metric value is determined to meet the preset condition, the early termination counter is controlled to count, and when the count value of the early termination counter is greater than the preset count threshold value, the iteration is terminated. Therefore, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of small performance loss, so that the decoding delay is reduced.
Drawings
FIG. 1 is a schematic diagram of the basic structure of a modern digital communication system;
FIG. 2 is a schematic diagram of a Tanner of a low density parity check code in the related art;
FIG. 3 is a flow chart of a decoding iteration in the related art;
FIG. 4 is a flow chart of a method for early termination of decoding according to an embodiment of the present invention;
FIG. 5 is a graph comparing the performance and average iteration number of the decoding iteration termination method and the early termination method of decoding using soft information based on the soft information metric values of the present application;
FIG. 6 is a graph comparing the performance and average iteration number of the decoding iteration termination method and the early termination method of decoding using the information metric based on hard information according to the present application;
FIG. 7 is a block diagram of an electronic device according to one embodiment of the invention;
FIG. 8 is a block diagram of a decoder according to one embodiment of the present invention;
fig. 9 is a block diagram illustrating a decoding early termination apparatus according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.
It should be noted that, referring to fig. 1, a modern digital communication system may include a source, a channel encoder, a modulator, a channel, a demodulator, a channel decoder, and a sink.
The source is the publisher of the information for generating a binary bit stream of information to be transmitted.
The channel encoder is used for encoding the binary bit stream to improve the ability of the receiving end to identify errors, thereby reducing the error rate to improve the quality of the recovered information. In particular, during transmission of a digital signal on a channel, due to non-ideal digital transmission characteristics of an actual channel and the presence of additive noise, errors may occur at a receiving end, and in order to control errors, an automatic repeat request (Automatic Request, ARQ) error detection technique and a forward error correction (Forward Error Correction, FEC) encoding technique are generally applied to a channel encoder, so as to improve reliability of information transmission, and further reduce an error rate to improve quality of recovered information. The most commonly used forward error correction coding techniques at present are: convolutional codes, TURBO codes, low density parity check codes, polarization codes, and the like.
The modulator is used to map (carry) the encoded binary bit stream onto a carrier to improve spectral efficiency. The modulator generally adopts IQ modulation (specifically, dividing data into two paths, respectively performing carrier modulation, and mutually quadrature two paths of carriers), and corresponding common modulation modes include binary phase shift keying (Binary Phase Shift Keying, BPSK), quadrature phase shift keying (Quadrature Phase Shift Keying, QPSK), quadrature amplitude modulation (Quadrature Amplitude Modulation, QAM), and the like.
A channel is a channel for information transfer, i.e., a medium on which electromagnetic waves propagate, and in a wireless communication system, the channel is free space. When information is transmitted through a channel, noise sources can damage transmitted information, the noise sources can be generally divided into external noise and internal noise according to the sources, wherein the external noise comprises various electromagnetic wave interferences existing in nature, and the internal noise refers to various noises generated by an electronic device.
The demodulator is used for detecting the binary bit sent by the sending end by utilizing a certain criterion according to the modulation mode of the modulator of the sending end. At present, soft demodulation is mainly adopted, so as to obtain soft bits corresponding to the transmission bits, namely log likelihood ratios (Log Likelihood Ratio, LLR) of the transmission bits, which are defined as follows:
wherein LLR (v n ) Representing the transmitted bit v n Corresponding soft bits, p r (y n |v n =a) represents the transmission bit v n Receiving symbol y when =a n A e {0,1}; ln (·) represents a log-taking operation.
The channel decoder is used for decoding by using soft bit information obtained by demodulation according to a forward error correction coding technology adopted by the channel encoder of the transmitting end and adopting a certain decoding criterion to obtain binary bit information transmitted by the transmitting end. Common decoding algorithms include maximum likelihood (Maximum Likelihood, ML) decoding, maximum a posteriori probability (Maximum A Priori Probability, MAP) decoding, belief propagation (Belief Propagation, BP) decoding, and the like.
The information sink is a receiver of the information and is used for receiving the binary bit information obtained by decoding and converting the binary bit information to obtain a message sent by the information source.
The low density parity check code is widely used in modern digital communication systems as a linear block code whose performance approaches shannon's limit, for example: digital video broadcasting systems in europe DVB-S2, WI-FI system, global positioning system (Global Positioning System, GPS)/beidou satellite navigation system and Fifth Generation (5G) mobile communication systems.
The block code is to group the input information sequence with every k symbols, and the channel encoder generates r redundant symbols (called check elements) according to a certain rule for each information group, so as to form a codeword with length of n=k+r, and the block code is generally represented by (n, k). When the relation between the information symbol and the check element of the block code is linear, the block code is called a linear block code. There are two important matrices for linear block codes: generating a matrix G and a check matrix H, for any one of the input information sequences u= (u) 0 ,u 1 ,...u k-1 ) The corresponding codeword of length N is v=u·g and h·v T =0, where v T The representation transposes the vector v.
The low density parity check code is defined as the null space of the check matrix H satisfying the following characteristics: 1) Each row has ρ non-0 elements; 2) Lambda non-0 elements in each column; 3) Compared with the code length N and the number of rows of the check matrix H, the number rho and the number lambda are much smaller, namely the check matrix H has sparse characteristics.
The low density parity check code may be represented by a Tanner graph, which is shown in fig. 2, taking the check matrix H described below as an example.
In FIG. 2, { x 0 ,x 1 ,...,x 7 And is a variable node, i.e., codeword v= (v) 0 ,v 1 ,...,v 7 ) Corresponding transmission symbols, according to the mapping rules x defined in the third generation partnership project (the 3rd Generation Partner Project,3GPP) standard n ＝1-2v n The method comprises the following steps: bit 0 maps to symbol 1; bit 1 maps to symbol-1; { s 0 ,s 1 ,s 2 ,s 3 And the check nodes respectively correspond to four check equations H.v T =0, i.e.:
wherein,,
representing a modulo-2 addition operation. When H (m, n) =1, then the check node s is represented m And variable node x n Connected, i.e. s m The corresponding check equation must include x n Corresponding bit v n 。
Belief propagation decoding algorithms are widely used in decoding of low density parity check codes due to their massively parallel processing characteristics. The basic idea of belief propagation decoding is: soft information (or confidence) is exchanged between the variable nodes and the check nodes, and the exchanged soft information is called a message. The specific implementation is as follows: after receiving all the messages of the check nodes (variable nodes) connected with the variable nodes (check nodes), the variable nodes (check nodes) update the messages transmitted to the check nodes (variable nodes) and transmit the updated messages to the check nodes (variable nodes) connected with the variable nodes, so that the message transmission is performed iteratively, and the message transmission is specifically shown as 3.
The following describes in detail the specific calculation process of each flow in fig. 3:
step 1: initializing.
The input to belief propagation decoding is each transmitted bit v n Channel soft information of (i.e. transmit bit v) n Is denoted by L r (v n ) It is defined as follows:
wherein y is n Representing the transmitted symbol x n Corresponding received symbols, p r (. Cndot.) represents probability values and ln (. Cndot.) represents log-taking operations.
VN2CN-msg in FIG. 3 represents variable node x n Pass to check node s m Variable node x n Passed to the check node s connected thereto m Is recorded as a message of
The definition and calculation formula are as follows:
wherein S is n Representation and variable node x n Connected check node set S n \s m Representation set S n Removing element s m The set of remaining elements that follow up,
representing a received symbol vector (y 0 ,y 1 ,...,y N-1 ) N represents the total number of variable nodes,
representing the ith iteration check node s m' Pass to variable node x n The definition and calculation formula of the message see check node update flow. During the initialization process, p->
Initializing:
step 2): and (5) checking node updating.
Check node s m Receiving variable node x connected with it n Update messages passed to each variable node to which it is connected according to the following:
wherein,,
X m representation and verification node s m Connected variable node set X m \x n Representing a collection X m Removing element x n The remaining element sets after.
Step 3): and updating the variable nodes.
Variable node x n After receiving the message from the check node connected with the message, the message is as follows
Updating the formula updates the message passed to each check node connected thereto:
at the same time, the variable node calculates its transmission bit v n Posterior probability of (2)
The definition and calculation formula are as follows:
step 4): and judging the iteration termination condition.
In the related art, the following methods or combinations thereof are used for the iteration termination judgment:
the first way is: setting the maximum iteration number I max Only when the set maximum iteration number I is reached max When the iteration is terminated.
The second way is: using a posterior probability of each transmitted bit calculated during variable node update
Performing hard decision to obtain a decoding result of the transmitted bit sequence, wherein the hard decision is performed according to the following formula:
then, checking, i.e. calculating, by using the check matrix
If->
The iteration is terminated; otherwise, the iteration is continued.
Third mode: if the low density parity check code is combined with an error detection code, such as a cyclic redundancy check (Cyclic Redundancy Check, CRC) code, the error detection code pair may be used to decode the sequence
Detecting error, if no error exists, ending iteration; otherwise, the iteration is continued.
The iteration termination judging method has the following defects: for the scene with low signal-to-noise ratio (i.e. the scene with extremely low probability of decoding success), the second and third modes will fail, and the decoding always iterates to the mostLarge number of iterations I max It is terminated but these iterations have no meaning and result in decoding delays.
In order to solve the technical problems, the invention concept of the application is as follows: and updating the early termination counter by using the information metric values of the previous and the next iterations, thereby terminating the iterations in advance.
In one embodiment, a method for early termination of decoding is provided, which can be applied to the channel decoder shown in fig. 1, and referring to fig. 4, the method for early termination of decoding includes the following steps:
in the iterative decoding process, information metric values of two iterations are calculated, step 402.
For example, referring to fig. 3, in an iterative decoding process, one iteration refers to: the check node and the variable node complete one-time updating, specifically: after receiving the information from the variable nodes connected with the check nodes, the check nodes update the information transferred to each variable node connected with the check nodes according to the formula (7) and send the information to the variable nodes connected with the check nodes, and after receiving the information from the check nodes connected with the variable nodes, the variable nodes update the information transferred to each check node connected with the variable nodes according to the formula (8).
After each iteration is completed, information metric values of two iterations are calculated, and the information metric values are used for representing the difference between the information. In the present application, the information metric value may include two kinds, i.e., a soft information-based information metric value and a hard information-based information metric value, respectively, according to the type of information used.
Taking the information metric based on soft information as an example. Optionally, calculating the information metric value of the two iterations includes: and calculating the increment information quantity of the message transmitted to the variable node by the check node during the two iterations, and calculating the average value of the absolute value of the increment information quantity as an information metric value.
Specifically, after each iteration is completed, the information transmitted by each check node to each variable node connected with each check node in the iteration is obtained
And the message that each check node passed to each variable node to which it was connected at the previous iteration +.>
Specifically, as shown in the above formula (7), the incremental information amount, i.e. the difference value, of the message transmitted by each check node to each variable node connected with each check node in two iterations is calculated
And finally, calculating the average value of absolute values of incremental information amounts of messages transmitted by all check nodes to all variable nodes connected with the check nodes during the previous and subsequent iterations, and taking the average value as an information metric value of the previous and subsequent iterations, wherein the information metric value can be expressed by the following formula:
wherein M represents the total number of check nodes, N m Representation and verification node s m The total number of variable nodes connected, i.e. set X in equation (7) above m Is a number of elements of (a).
Taking as an example an information metric based on hard information. Optionally, calculating the information metric value of the two iterations includes: and calculating the turnover proportion of the decoding bit sequence in the two times of iteration before and after as an information metric value.
The decoding bit sequence is the decoding result of the transmitting bit sequence. Optionally, the decoded bit sequence is obtained according to the following steps: and in the variable node updating process, calculating the posterior probability of each transmitted bit, and performing hard decision on the posterior probability to obtain a decoding bit sequence. Further, calculating the flip ratio of the decoded bit sequence in the two previous and subsequent iterations includes: and performing modulo-2 addition operation on the decoded bit sequences iterated before and after, and calculating the turnover proportion according to the modulo-2 addition operation result.
Specifically, at each timeAfter one iteration is completed, the posterior probability of each transmitted bit in the current iteration is obtained
And posterior probability per transmitted bit at previous iteration +.>
As shown in the above formula (9), and +.>
And->
Making hard decisions to obtain a decoded bit sequence, wherein the hard decisions are made according to the following formula:
then, decoding bit sequence for two iterations
And->
And performing modulo-2 addition operation, and calculating a turnover proportion according to the modulo-2 addition operation result, wherein the turnover proportion is an information metric value of two previous and subsequent iterations, and the information metric value is specifically shown in the following formula:
thus, for different types of information, a corresponding calculation metric mode is given to terminate the iteration in advance.
It can be seen from the above formulas (11) and (13) that, based on the calculation metric of the soft information, that is, the absolute average value of the difference value of the two soft information before and after is used as the calculation metric for early termination of the judgment, the complexity of the calculation metric is higher than that based on the hard information, but because the soft information reflects the confidence of the bit judgment, compared with the hard information, more information is contained, and meanwhile, the updated soft information can reflect the iteration condition, so that the early termination of the iteration by adopting the soft information is more effective, accurate and better in performance.
The computation metric based on the hard information, that is, the number of bit flips of two iterations before and after is used as the computation metric, although the computation metric is the computation metric related to the hard information, the metric does not need to carry out matrix multiplication computation, so that the complexity of the algorithm can be effectively reduced compared with the computation metric related to the hard information based on a syndrome. Specifically, the syndrome refers to the product of the correction matrix and the decoded bit vector, when decoding correctly, the syndrome should be zero, the calculation metric related to the hard information based on the syndrome refers to the calculation metric terminated in advance by adopting the non-zero weight value of the syndrome, if the non-zero weight value is too high, the decoding error is larger, and the iteration is terminated in advance. Although both are based on the hard information related computation metrics, the hard information related computation metrics of the present application do not require the computation of the product of the correction matrix and the decoded bit vector, i.e., do not perform the multiplication computation, thus greatly reducing the computational complexity.
And step 404, when the information measurement value is determined to meet the preset condition, controlling the early termination counter to count.
After each iteration is completed and the information metric values of the previous and subsequent iterations are obtained through calculation in the mode, judging whether the information metric values meet preset conditions, if so, indicating that the iteration has no gain, controlling an early termination counter to count, and if so, controlling the early termination counter to increment by 1; and if the preset condition is not met, resetting the early termination counter.
Taking the information metric based on soft information as an example. Optionally, when the average value of the absolute values of the incremental information amounts is far smaller than the average value of the absolute values of the soft information of the channel, it is determined that the information metric value meets a preset condition.
In particular, channelsThe log-likelihood ratio of soft information, i.e. the transmitted bits, is denoted as L r (v n ) The average value of the absolute value of the soft information of the channel is the average value of the absolute values of log likelihood ratios of all transmitted bits in the previous and subsequent iterations, and is shown in the following formula:
further, in the above-described manner, an information metric value η based on soft information is obtained soft When determining the information metric value eta soft I.e. whether the average value of the absolute value of the incremental information quantity is much smaller than the average value of the absolute value of the soft information of the channel
If yes, determining that the information measurement value meets a preset condition, indicating that iteration has no gain, and controlling an early termination counter delta to count at the moment; otherwise, the information metric value does not meet the preset condition, and at this time, the early termination counter delta is cleared, which can be expressed specifically by the following formula:
wherein alpha is soft Representing the threshold value set by the information metric value based on the soft information, can be obtained through simulation.
Taking as an example an information metric based on hard information. Optionally, when the flip ratio is greater than a preset threshold, determining that the information metric value meets a preset condition.
Specifically, in the case of obtaining the information metric value η based on the hard information in the foregoing manner hard When determining the information metric value eta hard The turning proportion is large, if yes, the information measurement value is determined to meet the preset condition, the iteration is not increased, and the counter delta is controlled to count in advance; otherwise, the information metric value does not meet the preset condition, and the early termination counter delta is cleared at this time, specifically by the following stepsThe formula is expressed:
wherein alpha is hard Representing the threshold value set by the information metric value based on the hard information, can be obtained through simulation.
Therefore, the early termination counter is updated based on the information metric values of the previous iteration and the next iteration, so that the iteration can be terminated in advance, the iteration times can be reduced, and the iteration time delay and the power consumption can be reduced.
In step 406, the iteration is terminated when the count value of the early termination counter is greater than the preset count threshold value.
Specifically, if δ > δ th Terminating the iteration and ending the decoding, wherein delta th The set counting threshold value can be obtained through simulation; otherwise, the iteration is continued.
Therefore, the early termination counter is updated based on the information metric values of the previous iteration and the next iteration, so that the iteration can be terminated in advance, the iteration times can be reduced, and the iteration time delay and the power consumption can be reduced.
Optionally, the method for terminating decoding in advance further includes: acquiring iteration times; and when the count value of the early termination counter is smaller than or equal to a preset count threshold value, if the iteration number reaches a preset maximum iteration number, terminating the iteration.
That is, when delta > delta th Or the iteration number reaches the preset maximum iteration number I max Terminating iteration and ending decoding; otherwise, continuing iterative decoding to avoid that the iteration cannot be terminated.
The following describes the effects of the early termination method for decoding provided in the present application with reference to fig. 5 to 6.
Fig. 5 shows a related art decoding iteration termination method (a method combining the first and third modes, and the second mode is not used in general because of the large calculation amount due to the matrix multiplication operation) and the soft information based information metric value of the present applicationA comparison of the performance of the decoding early termination method and the average iteration number. Fig. 6 is a graph comparing the performance and average iteration number of the decoding iteration termination method (the method combining the first mode and the third mode, and the second mode is not used because the matrix multiplication operation is needed, and the calculation amount is large) in the related art and the decoding early termination method using the information metric value based on the hard information in the present application. Wherein, the simulation parameters are set as follows: the code block size k=8448, the code rate is 1/3, alpha soft ＝0.16，α hard ＝0.2，δ th ＝3，I max =16. As can be seen from fig. 5 and fig. 6, when the signal-to-noise ratio is low, compared with the decoding iteration termination method in the related art, the average iteration number is obviously reduced, and the block error rate is basically the same, so that when the signal-to-noise ratio is low, the decoding iteration number can be greatly reduced by using the decoding early termination method provided by the application under the condition of extremely low performance loss, thereby reducing decoding delay and power consumption.
In summary, according to the decoding early termination method of the embodiment of the present invention, in the iterative decoding process, the information metric value of the two iterations is calculated, and when the information metric value meets the preset condition, the early termination counter is controlled to count, and when the count value of the early termination counter is greater than the preset count threshold value, the iteration is terminated. Therefore, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of small performance loss, so that the decoding delay and the power consumption are reduced.
In one embodiment, a computer readable storage medium is provided having stored thereon a decode early termination program that when executed by a processor implements the foregoing decode early termination method.
According to the computer readable storage medium, through the decoding early termination method, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of low performance loss, so that the decoding delay and the power consumption are reduced.
In one embodiment, as shown in fig. 7, an electronic device is provided, which includes a memory, a processor, and a premature decoding termination program stored in the memory and capable of running on the processor, where the premature decoding termination method is implemented when the processor executes the premature decoding termination program.
In particular, the memory may include a non-volatile storage medium that may store an operating system, computer programs, databases, and the like, and an internal memory that provides an environment for the operating system and computer programs in the non-volatile storage medium to run. The processor is configured to provide computing and control capabilities that when executed by the computer program implement the previously described early termination method of decoding. It will be appreciated by those skilled in the art that the structure shown in fig. 7 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the electronic device to which the present application is applied, and that a particular electronic device may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
According to the electronic equipment provided by the embodiment of the invention, through the decoding early termination method, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of smaller performance loss, so that the decoding time delay and the power consumption are reduced.
In one embodiment, as shown in fig. 8, a decoder is provided, which includes a memory, a processor, and a premature decoding termination program stored in the memory and capable of running on the processor, where the premature decoding termination method is implemented when the processor executes the premature decoding termination program.
In particular, the memory may include a non-volatile storage medium, which may store computer programs, databases, and the like, and an internal memory, which provides an environment for the execution of the computer programs in the non-volatile storage medium. The processor is configured to provide computing and control capabilities that when executed by the computer program implement the previously described early termination method of decoding. It will be appreciated by those skilled in the art that the structure shown in fig. 8 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the decoder to which the present application is applied, and that a particular decoder may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
According to the decoder provided by the embodiment of the invention, through the decoding early termination method, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of smaller performance loss, so that the decoding time delay and the power consumption are reduced.
In one embodiment, there is provided a decoding early termination apparatus, as shown with reference to fig. 9, including: a calculation module 10, a determination module 20 and a termination module 30.
The calculation module 10 is used for calculating information metric values of two iterations before and after the iterative decoding process; the determining module 20 is configured to control the early termination counter to count when the information metric value meets a preset condition; the termination module 30 is configured to terminate the iteration when the count value of the early termination counter is greater than a preset count threshold value.
In one embodiment, the calculation module 10 is further configured to calculate an incremental information amount of the message transmitted from the check node to the variable node in two iterations, and calculate an average value of absolute values of the incremental information amounts as the information metric value.
In one embodiment, the calculating module 10 is further configured to calculate the flip ratio of the decoded bit sequence in the two previous and subsequent iterations as the information metric value.
In one embodiment, the calculating module 10 is further configured to perform a modulo-2 addition on the decoded bit sequence of the two previous and subsequent iterations, and calculate the flip ratio according to the modulo-2 addition result.
In one embodiment, the calculating module 10 is further configured to calculate a posterior probability of each transmitted bit in the variable node update process, and make a hard decision on the posterior probability to obtain a decoded bit sequence.
In one embodiment, the determining module 20 is further configured to determine that the information metric value meets the preset condition when the average value of the absolute values of the incremental information amounts is far smaller than the average value of the absolute values of the soft information of the channel.
In one embodiment, the determining module 20 is further configured to determine that the information metric value meets a preset condition when the flip ratio is greater than a preset threshold value.
In one embodiment, the decoding early termination device further includes an obtaining module (not shown in the figure) for obtaining the iteration number; the termination module 30 is further configured to terminate the iteration if the iteration number reaches the preset maximum iteration number when the count value of the early termination counter is less than or equal to the preset count threshold value.
For specific limitation of the decoding early termination device, reference may be made to the limitation of the decoding early termination method hereinabove, and the description thereof will not be repeated here. The modules in the early termination device may be implemented in whole or in part by software, hardware, or a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
According to the decoding early termination device provided by the embodiment of the invention, the information metric values of the two iterations before and after are calculated in the iterative decoding process, when the information metric value is determined to meet the preset condition, the early termination counter is controlled to count, and when the count value of the early termination counter is greater than the preset count threshold value, the iteration is terminated. Therefore, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of small performance loss, so that the decoding delay and the power consumption are reduced.
In one embodiment, a decoder is provided, including the decoding early termination device described above.
According to the decoder provided by the embodiment of the invention, through the decoding early termination device, when the signal-to-noise ratio is low, the iteration times can be greatly reduced under the condition of smaller performance loss, so that the decoding time delay and the power consumption are reduced.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the various embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The above examples merely represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.
Claims (7)
1. A method for early termination of decoding, comprising:
in the iterative decoding process, calculating information metric values of two iterations before and after;
when the information measurement value is determined to meet the preset condition, controlling an early termination counter to count;
terminating the iteration when the count value of the early termination counter is larger than a preset count threshold value;
calculating information metric values of two iterations before and after the calculation, comprising:
calculating the turnover proportion of a decoding bit sequence in two iterations before and after as the information metric value, wherein the decoding bit sequence is obtained by performing hard decision on posterior probability of a sending bit sequence;
when the turnover proportion is larger than a preset threshold value, determining that the information measurement value meets a preset condition;
calculating the turnover proportion of the decoding bit sequence in the two times of iteration comprises the following steps:
and carrying out modulo-2 addition operation on the decoded bit sequences iterated before and after, and calculating the turnover proportion according to the modulo-2 addition operation result.
2. The method of claim 1, wherein calculating the information metric value for two iterations before and after comprises:
and calculating the increment information quantity of the message transmitted to the variable node by the check node during the two iterations, and calculating the average value of the absolute values of the increment information quantity to be used as the information metric value.
3. The decoding early termination method of claim 1, wherein the sequence of decoded bits is obtained according to the steps of:
and in the variable node updating process, calculating the posterior probability of each transmitted bit, and performing hard decision on the posterior probability to obtain the decoding bit sequence.
4. The decoding early termination method of claim 2, wherein the information metric value is determined to satisfy a preset condition when an average value of absolute values of the incremental information amounts is far smaller than an average value of absolute values of channel soft information.
5. The early termination method of decoding of claim 1, further comprising:
acquiring iteration times;
and when the count value of the early termination counter is smaller than or equal to a preset count threshold value, if the iteration number reaches a preset maximum iteration number, terminating the iteration.
6. A computer-readable storage medium, having stored thereon a decoding early termination program that, when executed by a processor, implements the decoding early termination method of any of claims 1-5.
7. An electronic device comprising a memory, a processor, and a premature decoding termination program stored in the memory and executable on the processor, wherein the premature decoding termination method according to any one of claims 1-5 is implemented when the premature decoding termination program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110152057.3A CN112994704B (en) | 2021-02-03 | 2021-02-03 | Decoding early termination method, storage medium and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110152057.3A CN112994704B (en) | 2021-02-03 | 2021-02-03 | Decoding early termination method, storage medium and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112994704A CN112994704A (en) | 2021-06-18 |
| CN112994704B true CN112994704B (en) | 2023-06-16 |
Family
ID=76346587
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110152057.3A Active CN112994704B (en) | 2021-02-03 | 2021-02-03 | Decoding early termination method, storage medium and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112994704B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111327332A (en) * | 2020-03-31 | 2020-06-23 | 东南大学 | Noise gradient reduction multi-bit flip decoding early-stop method of LDPC code |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1905374A (en) * | 2005-07-25 | 2007-01-31 | 松下电器产业株式会社 | Coding and decoding method of asymmetric low dense check code |
| CN101345602B (en) * | 2008-08-21 | 2013-01-16 | 上海交通大学 | Early termination method of low density check code iteration decoding |
| US8661326B1 (en) * | 2010-08-04 | 2014-02-25 | Marvell International Ltd. | Non-binary LDPC code decoding early termination |
| CN103905063B (en) * | 2014-04-23 | 2017-02-15 | 湖南国科微电子股份有限公司 | Decoding method for LDPC coder |
-
2021
- 2021-02-03 CN CN202110152057.3A patent/CN112994704B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111327332A (en) * | 2020-03-31 | 2020-06-23 | 东南大学 | Noise gradient reduction multi-bit flip decoding early-stop method of LDPC code |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112994704A (en) | 2021-06-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108282264B (en) | Polar code decoding method based on bit flipping serial elimination list algorithm | |
| US6518892B2 (en) | Stopping criteria for iterative decoding | |
| US9385753B2 (en) | Systems and methods for bit flipping decoding with reliability inputs | |
| US9323611B2 (en) | Systems and methods for multi-stage soft input decoding | |
| US9559873B2 (en) | Signal receiving apparatus based on faster than nyquist and signal decoding method thereof | |
| US20090222711A1 (en) | Generalized Multi-Threshold Decoder for Low-Density Parity Check Codes | |
| US10547328B1 (en) | Implementation of LLR biasing method in non-binary iterative decoding | |
| US8650451B2 (en) | Stochastic stream decoding of binary LDPC codes | |
| US20140258809A1 (en) | Systems and methods for decoding with late reliability information | |
| US9009578B1 (en) | Methodology for improved bit-flipping decoder in 1-read and 2-read scenarios | |
| CN112953569B (en) | Decoding method and device, storage medium, electronic equipment and decoder | |
| US9614548B1 (en) | Systems and methods for hybrid message passing and bit flipping decoding of LDPC codes | |
| CN106656209B (en) | Cascade code method for correcting synchronous errors by adopting iterative decoding | |
| US9432054B2 (en) | Method for decoding a correcting code with message passing, in particular for decoding LDPC codes or turbo codes | |
| CN107026655B (en) | Method for decoding a codeword and decoder | |
| CN106788458B (en) | Hard decision-oriented forward and backward estimation method for insertion deletion and substitution errors | |
| CN112994704B (en) | Decoding early termination method, storage medium and electronic equipment | |
| CN114430279B (en) | List Viterbi decoding method, device, decoder and storage medium | |
| CN102291198B (en) | Channel decoding method and device | |
| US10903855B2 (en) | Convolutional LDPC decoding method and apparatus, decoder, and system | |
| Geldmacher et al. | EXIT-optimized index assignments for turbo decoders with unreliable LLR transfer | |
| CN114499548B (en) | Decoding method, device and storage medium | |
| CN118054797B (en) | Coding and decoding method, device and equipment | |
| CN113824452B (en) | Decoding method based on grid graph, component decoder and channel decoder | |
| CN114448448B (en) | CA-SCL-based polarization code encoding and decoding method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |