WO2019056941A1

WO2019056941A1 - Decoding method and device, and decoder

Info

Publication number: WO2019056941A1
Application number: PCT/CN2018/103736
Authority: WO
Inventors: 张朝阳; 王俊; 王桂杰
Original assignee: 华为技术有限公司
Priority date: 2017-09-22
Filing date: 2018-09-03
Publication date: 2019-03-28
Also published as: CN109547034A; CN109547034B

Abstract

A method and device for decoding channel coding in a wireless communication system, and a decoder. The method comprises: determining an information bit set according to information bit quantity K and a target code length N of information to be decoded (S101), and determining a key set according to the information bit set, K and N, wherein the probability that the key set comprises a decoding first error position bit is greater than a first threshold value, and the key set is a subset of the information bit set (S102). According to the key set and a preset rule, polarization decoding is carried out on the information to be decoded, the preset rule being as follows: performing decoding path splitting on information bits in the key set, performing bit-by-bit determination for information bits outside of the key set, generating a decoding path, selecting L paths having the largest PM values when the total number of decoding paths exceeds a preset path width L, saving the same and continuously developing decoding paths, and obtaining a decoding result of each level of decoding bits according to the decoding path of each level of decoding bits (S103). Thus, computing complexity and decoding time delays, which are caused by sequencing, are reduced.

Description

Decoding method and device, decoder

Technical field

The present application relates to the field of communications technologies, and in particular, to a decoding method, device, and decoder.

Background technique

Communication systems usually use channel coding to improve the reliability of data transmission and ensure the quality of communication. The Polar code is the first channel coding method that can theoretically be said to "reach" the channel capacity. The Polar code is a linear block code whose generating matrix is G _N and its encoding process is

Is a binary line vector of length N (ie code length);

Here

B _N is an N×N transposed matrix, such as a bit reverse transposed matrix;

Defined as the Kronecker product of log ₂ N matrices F ₂ , x ₁ ^N is the encoded bit (also called the codeword),

The multiplied by the generator matrix G _N gives the encoded bits, and the process of multiplication is the process of encoding. In the encoding process of the Polar code,

A part of the bits are used to carry information, called information bits, and the set of index bits of information bits is recorded as

The other part of the bit is set to a fixed value pre-agreed by the transceiver, which is called a fixed bit, and the set of indexes is used.

Complement

Said. The fixed bit is usually set to 0, and only needs to be pre-agreed by the transceiver. The fixed bit sequence can be arbitrarily set.

In the decoding method of the Polar code, the decoding process of the existing method of the Successive Cancellation (SC) is: after receiving the information to be decoded (including information bits and fixed bits), Decoding the information bits in the information, calculating the Log Likelihood Ratio (LLR) of each information bit one by one, and performing bit-by-bit decision. If the LLR of the information bit is > 0, the decoding result is 0, if the information bit If the LLR<0, the decoding result is 1, and for the fixed bit in the information to be decoded, no matter how many decoding results of the LLR are set to 0, all the bits are sequentially decoded in order, and the result of the previous decoding bit As an input to the calculation of the latter decoding bit, once the error is judged, the error is spread and there is no chance to recover, so the decoding performance is not high. To solve this problem, in the Successive Cancellation List (SCL), the SCL algorithm saves the decoding results corresponding to 0 and 1 as two branch decoding paths when decoding each information bit ( Referred to as path splitting, Figure 1 is a schematic diagram of the decoding path in the SCL algorithm. As shown in Figure 1, each layer represents one decoding bit. If the decoding result is 0, the path is developed along the left subtree. If the decoding result is 1, the path is developed along the right subtree. When the total number of decoding paths exceeds the preset path width L (generally L=2 ^l ), the path metric (PM) value is selected to be the best. The L paths save and continue to develop the path to decode subsequent decoding bits, wherein the PM value is used to judge whether the path is good or bad, and the PM value is calculated by the LLR. For each level of decoding bits, the PM values of the L paths are sorted from small to large, and the correct path is filtered by the PM value, and so on, until the last bit is translated.

In practical applications, the number of decoding bits is very large. Using the SCL decoding method, for each decoding bit, the PM value of all paths under each decoding bit is calculated, and all paths are based on the PM value. Performing a sort, its computational complexity and decoding delay due to sorting are high.

Summary of the invention

The present application provides a decoding method and apparatus, and a decoder to reduce the computational complexity of the decoding of the Polar code and the decoding delay due to the sorting.

In a first aspect, the present application provides a decoding method, including: determining an information bit set according to a number of information bits K and a target code length N of information to be decoded; determining a key set according to the information bit set, K and N, in a key set The probability of including the first bit of the decoded bit position is greater than the first threshold, and the key set is a subset of the set of information bits; and the information to be decoded is subjected to polarization decoding according to the key set and the preset rule, and the preset rule is: The information bits in the set are split by the decoding path, and the information bits outside the key set are bit-by-bit-determined to generate a decoding path. When the total number of decoding paths exceeds the preset path width L, the path metric PM value is selected. The largest L paths save and continue to develop the decoding path, and the decoding result of each level of decoding bits is obtained according to the decoding path of each level of decoding bits.

According to the decoding method provided by the first aspect, the information bit set is first determined according to the information bit number of the information to be decoded and the target code length, and the key set is determined according to the information bit set, K and N, and the key set is greater than the first threshold. The probability includes decoding the first wrong position bit, performing polarization decoding on the information to be decoded according to the key set during decoding, and decoding the information bit in the key set when the decoding path is generated, outside the key set The information bits are subjected to a bit-by-bit decision. Since the first bit position bit is included in the key set with a high probability, the decoding path will be split with a high probability at the first decoding error position, so the correct path is high probability. It is retained in the current decoding path, so that better decoding performance is obtained, and the number of path splits is greatly reduced, which reduces computational complexity and decoding delay due to sorting. Therefore, the computational complexity and decoding delay of the Polar code decoding are reduced while ensuring better decoding performance.

In a possible design, the information bits in the key set are split by the coding path, and the information bits outside the key set are determined bit by bit, including:

Where u _i is the information bit in the information bit set, S is the key set, L(u _i ) is the log likelihood ratio LLR of u _i , h(·) is a hard decision function,

In one possible design, a key set is determined based on the set of information bits, K and N, including:

Constructing a full binary tree of depth n according to the information bit set, K and N, n=log ₂ N, and the information bits and fixed bits in the full binary tree are leaf nodes;

All sub-blocks with a ratio of 1 are determined from the full binary tree, the sub-blocks with a ratio of 1 are sub-trees with all the leaf nodes being information bits, and the sub-blocks with a ratio of 1 contain the number of leaf nodes greater than or equal to 1;

The elements constituting the key set are determined from all sub-blocks having a ratio of 1, and the elements constituting the key set are the first information bits in each sub-block of ratio 1.

With the decoding method provided by this embodiment, a full binary tree of depth n is constructed according to the information bit set, K and N, and all sub-blocks with a ratio of 1 are determined from the full binary tree, and are determined from all sub-blocks with a ratio of 1. Forming the elements of the key set, thereby determining a key set including the first bit of the decoded bit position with high probability, and finally performing polarization decoding on the information to be decoded according to the key set and the preset rule, and generating the decoding path into the key set The information bits are split by the decoding path, and the information bits outside the key set are bit-by-bit-determined. Since the first wrong position bit is included in the key set with high probability, the decoding path will be translated at the first probability with high probability. The position of the code error is split, so the correct path is retained in the current decoding path with high probability, so that better decoding performance is obtained, and the number of path splits is greatly reduced, the computational complexity is reduced, and the ordering band is reduced. The decoding delay. Therefore, the computational complexity and decoding delay of the Polar code decoding are reduced while ensuring better decoding performance.

In one possible design, the elements that make up the key set further include at least one of the remaining information bits in all of the sub-blocks of ratio 1.

With the decoding method provided by this embodiment, the probability that the key set contains the bits of the decoded first error position can be further improved.

A key set corresponding to the information bit set, K and N is queried from the stored key set table, and the key set table stores the information bit set, the correspondence between K and N and the key set.

The decoding method provided by the embodiment, after pre-storing the key set table, determining the information bit set according to the number of information bits of the information to be decoded and the target code length, and querying the information bit set from the stored key set table, A key set corresponding to K and N, thereby determining a key set including the first bit of the decoded bit position with high probability, and finally performing polarization decoding on the information to be decoded according to the key set and the preset rule, without constructing a full binary tree to determine the key The set can further reduce the decoding delay.

In one possible design, the first threshold is 99%.

In a second aspect, the present application provides a decoding apparatus, including: a first determining module, configured to determine an information bit set according to a number of information bits K and a target code length N of information to be decoded; and a second determining module, configured to The information bit set, K and N determine a key set, the key set includes a probability that the first bit of the decoded bit position is greater than the first threshold, the key set is a subset of the information bit set, and the decoding module is configured to The preset rule performs polarization decoding on the information to be decoded. The preset rule is: splitting the information path of the information bits in the key set, and performing bit-by-bit decision on the information bits outside the key set to generate a decoding path. When the total number of decoding paths exceeds the preset path width L, the L paths with the largest path metric PM value are selected to be saved and continue to develop the decoding path, and each level of decoding is obtained according to the decoding path of each level of decoding bits. Bit decoding result.

In a possible design, the second determining module is configured to: construct a full binary tree with a depth n according to the information bit set, K and N, n=log ₂ N, and the information bits and the fixed bits in the full binary tree are leaf nodes; All sub-blocks with a ratio of 1 are determined from the full binary tree. The sub-block with a ratio of 1 is a sub-tree with all leaf nodes being information bits, and the sub-block with a ratio of 1 contains the number of leaf nodes greater than or equal to 1; The elements constituting the key set are determined among all the sub-blocks having a ratio of 1, and the elements constituting the key set are the first information bits in each of the sub-blocks having a ratio of 1.

In a possible design, the second determining module is configured to: query a key set corresponding to the information bit set, K and N from the stored key set table, as a key set, and store the information bit set in the key set table, K And the correspondence between N and the key set.

In one possible design, the first threshold is 99%.

The beneficial effects of the decoding apparatus provided by the above second aspect and the possible designs of the above second aspect can be seen in the beneficial effects brought by the above first aspect and the possible designs of the first aspect, and no longer Narration.

In a third aspect, the application provides a decoder, including: a memory and a processor;

The memory is used to store program instructions;

The processor is operative to invoke program instructions in the memory to perform the first aspect and the decoding method in any of the possible designs of the first aspect.

In a fourth aspect, the present application provides a readable storage medium having stored therein an execution instruction, where the decoding apparatus performs the first aspect and the first aspect when at least one processor of the decoding apparatus executes the execution instruction A coding method in any of the possible designs.

In a fifth aspect, the present application provides a program product comprising an execution instruction stored in a readable storage medium. At least one processor of the decoding device can read the execution instructions from a readable storage medium, the at least one processor executing the execution instructions such that the decoding device implements the first aspect and the decoding in any of the possible designs of the first aspect method.

DRAWINGS

1 is a schematic diagram of a decoding path in an SCL algorithm;

2 is a schematic structural diagram of a system of a sending device and a receiving device provided by the present application;

3 is a schematic diagram of a process of constructing a full binary tree when N=4;

4 is a flowchart of an embodiment of a decoding method provided by the present application;

Figure 5 is a full binary tree constructed with N = 16;

FIG. 6 is a flowchart of another embodiment of a decoding method provided by the present application;

FIG. 7 is a flowchart of another embodiment of a decoding method provided by the present application;

8 is a comparison diagram of the SNR required for CA-SCL decoding to achieve 0.1% BLER using the Split-reduced decoding method and the performance of the conventional CA-SCL decoding method;

9 is a schematic diagram of delay and complexity gain at different code rates using a Split-reduced decoding method;

10a-10f are comparison diagrams of BLER performances under different code length code rate configurations using a Split-reduced decoding method and a conventional CA-SCL decoding method;

FIG. 11 is a schematic structural diagram of an embodiment of a decoding device provided by the present application;

12 is a schematic structural diagram of a decoder provided by the present application;

FIG. 13 is a schematic structural diagram of another decoder provided by the present application.

Detailed ways

The embodiments of the present application can be applied to a wireless communication system. It should be noted that the wireless communication system mentioned in the embodiments of the present application includes but is not limited to: Narrow Band-Internet of Things (NB-IoT), global mobile Global System for Mobile Communications (GSM), Enhanced Data Rate for GSM Evolution (EDGE), Wideband Code Division Multiple Access (WCDMA), Code Division Multiple Access (CDMA) 2000 System (Code Division Multiple Access, CDMA2000), Time Division-Synchronization Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), and Next Generation 5G Mobile Communication System The three major application scenarios: Enhanced Mobile Broad Band (eMBB), Ultra-reliable and low-latency communications (uRLLC), and Massive Machine-Type Communications (Massive Machine-Type Communications, mMTC).

The communication device involved in the present application mainly includes a network device or a terminal device. The decoding method provided by the present application may be implemented by software or hardware on the network device side/terminal device side. If the sending device in the present application is a network device, the receiving device is a terminal device; if the sending device in the present application is a terminal device , the receiving device is a network device.

In the embodiment of the present application, the terminal device includes, but is not limited to, a mobile station (MS, Mobile Station), a mobile terminal (Mobile Terminal), a mobile telephone (Mobile Telephone), a mobile phone (handset), and a portable device (portable equipment). And so on, the terminal device can communicate with one or more core networks via a Radio Access Network (RAN), for example, the terminal device can be a mobile phone (or "cellular" phone), with wireless communication Functional computers, etc., terminal devices can also be portable, pocket-sized, handheld, computer-integrated or in-vehicle mobile devices or devices.

In the embodiment of the present application, the network device may be a device for communicating with the terminal device, for example, may be a base station (Base Transceiver Station, BTS) in a GSM system or a CDMA system, or a base station in a WCDMA system (NodeB) , NB), may also be an evolved base station (Evolved Node B, eNB or eNodeB) in the LTE system, or the network device may be a relay station, an access point, an in-vehicle device, a wearable device, and a network side in a future 5G network. Equipment or network equipment in the future evolution of the Public Land Mobile Network (PLMN).

The wireless communication system of the present application may include a transmitting device and a receiving device. FIG. 1 is a schematic diagram of a system architecture of a sending device and a receiving device provided by the present application. As shown in FIG. 1 , the sending device is an encoding side, which may be used for Encoding and outputting the encoded information, the encoded information is transmitted to the decoding side on the channel; the receiving device is the decoding side, and can be used to receive the encoded information sent by the transmitting device, and decode the encoded information.

2 is a schematic flow chart of a wireless communication system. As shown in FIG. 2, on the transmitting device side, the source is sequentially sent after source coding, channel coding, rate matching, and digital modulation. On the receiving device side, the destination is sequentially output by digital demodulation, de-rate matching, channel decoding, and source decoding. The channel coding code may use a Polar code, and the channel decoding may use the decoding method provided by the present application.

The decoding method provided by the present application mainly relates to the decoding process of the Polar code. In order to obtain better decoding performance, in the existing decoding method SCL, when decoding each information bit, the translation corresponding to 0 and 1 is performed. The code results are saved as two branch decoding paths. When the total number of decoding paths exceeds the preset path width L, the L paths with the best PM value are selected to be saved and the path is further developed to decode the subsequent decoding bits. For each decoding bit, calculate the PM value of all the paths under each decoding bit, and sort all the paths according to the PM value, the computational complexity and the decoding delay due to the sorting. In the text, the decoding delay is high. To solve the above problem, the present application provides a decoding method, which determines a key set according to a set of information bits of information to be decoded, a number of information bits K of information to be decoded, and a target code length N, and the key set includes a decoding first error. The probability of the position bit (the first decoding error bit) is greater than the first threshold value, and the information bits in the key set are split by the coding path at the time of decoding, and information other than the information bits in the key set is decoded. The bit uses a hard decision function for bit-by-bit decision. Since the first bit position bit is included in the key set with high probability, the decoding path will split at the first decoding error position with high probability, so the correct path is high. The probability is preserved in the current decoding path, so that better decoding performance is obtained, and the number of path splits is greatly reduced, thereby reducing the total number of decoding paths, thus reducing the number of paths for which the PM value is to be calculated and participating. The number of sorted paths reduces computational complexity and decoding delay due to sorting, thus achieving a reduction in decoding performance while ensuring better decoding performance Polar code decoding computational complexity and decoding delay. The decoding method, device and decoder provided by the present application are described in detail below with reference to the accompanying drawings.

For the sake of convenience, the definitions of the parameters that may be used in this application are first given, as shown in Table 1:

Table I

The following is a brief introduction to the full binary tree, the sub-block with the ratio of 1, and the sub-block with the ratio of 0 in the present application. The polarization code of N=4, K=2, and R=2/4 is taken as an example. Information bit set

First, a full binary tree with a depth of n=log ₂ (N)=2 is constructed according to N. FIG. 3 is a schematic diagram of a process of constructing a full binary tree when N=4, as shown in the left side of FIG. 3, a full binary tree, 4 The leaf nodes {D, E, F, G} correspond to the information bits {u ₁ , u ₂ , u ₃ , u ₄ }, respectively, and a full binary tree as shown in the middle of FIG. 3, for belonging to the information bit set

Bits, leaf nodes are represented by black, belonging to a fixed set of bits

The leaf nodes are represented by white nodes, and then the nodes in the full binary tree are dyed. The coloring rule is: for non-leaf nodes (such as A, B, C), if the left and right child nodes have the same color, the node color and its children The color of the nodes is the same, otherwise the node is dyed as a gray node, and the dyed full binary tree is a full binary tree as shown on the right side of Figure 3, which is the final constructed full binary tree.

Based on the constructed full binary tree (a full binary tree shown on the right side of Figure 3), the root node corresponding to the all black binary tree such as node C in the full binary tree is called a subblock of ratio 1 (because its leaf nodes are all information) Bit), that is, a sub-block of ratio 1 is a subtree in which all leaf nodes are information bits. The root node of an all-white binary tree such as Node B is a sub-block with a ratio of 0 (because its leaf nodes are fixed bits). A sub-block is a component polar code.

4 is a flowchart of an embodiment of a decoding method provided by the present application. The executor of this embodiment may be a network device or a terminal device as a sending device. As shown in FIG. 4, the method in this embodiment may include:

S101. Determine a set of information bits according to the number of information bits K of the information to be decoded and the target code length N.

Specifically, after receiving the information to be decoded, the target code lengths N and R of the information to be decoded may be determined, and the target code length is also referred to as the code length of the polarization code, and the value of K may be based on the code rate R and the code. After the sequence length M is determined, K=M*R, M passes

determine.

Wherein, when determining the information bit set, the existing information bit set may be determined by any existing method for constructing the information bit set, for example, by Gaussian Approximation (GA), Density Evolution (DE), and polarization weight. (Polar Weight, PW) and other methods to determine, taking the Gaussian approximation method as an example, specifically calculating the reliability of the polarized channel according to the Gaussian approximation method, and ranking the high reliability according to the polarization channel reliability in ascending or descending order. The polarized channel is used as a subchannel of the information bit, and the index of the subchannel of the information bit is the index of the information bit, that is, the information bit set is determined. Other information bit set construction methods are not enumerated in this embodiment.

S102. Determine a key set according to the information bit set, K and N. The probability that the key set includes the first bit of the decoded bit position is greater than the first threshold, and the key set is a subset of the information bit set.

The decoding first error location bit may specifically be the first wrong location bit in the SC decoding process, and the first threshold is such that the key set includes the decoding first error location bit with a higher probability, and the first threshold may be It is 99%.

Specifically, as a first implementation manner, determining a key set according to the information bit set, K, and N may specifically include:

S1021: Construct a full binary tree with a depth n according to the information bit set, K and N, n=log ₂ N, and the information bits and the fixed bits in the full binary tree are leaf nodes.

The information bits and the positions of the fixed bits are determined according to the information bit set determined in S101, for example, N=2 ⁴ and K=9 of the information to be decoded, and the information bit set is determined according to S101.

The fixed set of bits corresponds to:

A full binary tree with a depth of n=log ₂ (N)=4 is constructed with information bits and fixed bits as leaf nodes, and a full binary tree constructed with N=16 is shown in FIG. 5 , and the full binary tree shown in FIG. 5 is according to the above. The dyeing rules are dyed. The leaf nodes are black with information bits, and the leaf nodes are white for fixed bits. For non-leaf nodes, if the left and right child nodes have the same color, the node color is the same as its child node color. Otherwise, the node It is colored as a gray node.

S1022: determining all sub-blocks with a ratio of 1 from a full binary tree, the sub-block with a ratio of 1 is a sub-tree with all leaf nodes being information bits, and the sub-block with a ratio of 1 includes a number of leaf nodes greater than or equal to 1. .

Specifically, taking the full binary tree shown in FIG. 5 as an example, all sub-blocks with a ratio of 1 are determined from the full binary tree, and the sub-blocks with a ratio of 1 are sub-trees in which all leaf nodes are information bits, as shown in FIG. 5 . In the A, B, C, and D subtrees shown, each subblock is a component polarization code, but only contains information bits. For example, subtree A is a component code of length 4, which contains information bits {u ₁₃ , u ₁₄ , u ₁₅ , u ₁₆ }, subtree B is a component code of length 2, containing information bits {u ₁₁ , u ₁₂ }, and subtree C is a component code of length 2, including Information bits {u ₇ , u ₈ }, subtree D is a component code of length 1 containing the information bits {u ₆ }.

S1023. Determine, from all the sub-blocks with a ratio of 1, an element that constitutes a key set, and the elements that make up the key set are the first information bits in each sub-block with a ratio of 1.

Specifically, taking the full binary tree shown in FIG. 5 as an example, the sub-block with a ratio of 1 includes A, B, C, and D sub-trees, and the elements constituting the key set are determined from the sub-trees of A, B, C, and D, respectively. The first information bit is taken into the key set, and the first information bits in the subtrees A, B, C, and D are u ₆ , u ₇ , u ₁₁ , and u _{13 respectively} , so the full binary tree shown in FIG. 5 corresponds to The key set S = {u ₆ , u ₇ , u ₁₁ , u ₁₃ }.

Further, as a second implementable manner, in order to improve the probability that the key set includes the first bit of the decoded first bit position, the elements constituting the key set may further include: at least all of the remaining information bits in the sub-blocks of all ratios 1 An information bit, or taking the full binary tree shown in Figure 5 as an example, after taking the first information bit from the A, B, C, and D subtrees, since the D subtree contains only one information bit, from A and B. And all the remaining information contained in the C subtree, and at least one information bit may be taken out as an element constituting the key set, for example, u ₈ is extracted from the C subtree, and the key set S={u ₆ , u ₇ , u ₁₁ , u ₁₃ , u ₈ }; or take u ₁₂ from the B subtree, the key set S={u ₆ , u ₇ , u ₁₁ , u ₁₃ , u ₁₂ }; or, take u ₁₄ from the A subtree The key set S={u ₆ , u ₇ , u ₁₁ , u ₁₃ , u ₁₄ }, or one information bit is taken from each subtree. It should be noted that the number of path splits is reduced in the decoding process. The key set S is a subset of the set of information bits.

As a third implementation manner, determining a key set according to the information bit set, K, and N may specifically include:

Specifically, the first set of implementable manners and the second implementable manner may be used to determine different sets of information bits, and key sets corresponding to K and N, which are pre-stored in the key set table and receive information to be decoded. After that, the information bit set is determined according to the information bit number K of the information to be decoded and the target code length N, and then the key set corresponding to the information bit set, K and N can be directly queried from the key set table, and the method can further Reduce the decoding delay.

S103. Perform polarization decoding on the information to be decoded according to the key set and the preset rule. The preset rule is: performing coding path splitting on the information bits in the key set, and performing bit-by-bit decision on the information bits outside the key set. Generating a decoding path. When the total number of decoding paths exceeds a preset path width L, the L paths with the largest PM value are selected to be saved and continue to develop the decoding path, and each decoding bit of each decoding bit is obtained. The decoding result of the primary decoding bit.

Wherein, the information bits in the key set are split by the decoding path, and the path splitting is to save the decoding results corresponding to 0 and 1 as the two branch decoding paths, and the information bits except the information bits in the key set. The bit-by-bit decision is used to generate a decoding path, and the fixed bit is also subjected to bit-by-bit decision. The bit-by-bit decision result of the fixed bit is set to 0. When the total number of decoding paths exceeds the preset path width L, the path metric is selected. The L paths with the largest PM value are saved and continue to develop the decoding path. After the decoding path is generated, the SCL decoding method or the Cyclic redundancy check Aided-SCL (CA-SCL) decoding method can be used. The same process determines the decoding path of each level of decoding bits, for example, using the same process as the SCL decoding method, that is, calculating the PM values of all decoding paths of each level of decoding bits and sorting them, PM The one decoding path with the largest value is used as the decoding path of the decoding bits, and finally the decoding result of each level of decoding bits is obtained according to the decoding path of each level of decoding bits. It should be noted that the calculation of the PM value may be performed by using an existing calculation method, which is not limited in this application. For example, the same process as the CA-SCL decoding method is used, that is, the PM values of all decoding paths of each level of decoding bits are calculated and sorted, and a decoding path having the largest PM value is used as decoding of decoding bits. The path is verified by Cyclic Redundancy Check (CRC) to verify the correctness of the decoding path. Finally, the decoding result of each level of decoding bits is obtained according to the decoding path of each level of decoding bits.

Decoding the information bits in the key set, and performing bit-by-bit decision on the information bits outside the key set, that is, when generating the decoding path, the next level of information bits in the key set is two. The next level of information bits outside the key set is one, as an implementable method, specifically:

After receiving the information to be decoded, the N LLRs corresponding to the information to be decoded are obtained by calculation, and the bit-by-bit decision result of the fixed bits is 0 according to the LLR and the hard decision function.

In the decoding method of this embodiment, since the key set includes decoding the first mis-location bit (the first decoding error bit) with a high probability, the information bits in the key set are translated at the time of decoding. The code path is split, and the information bits except the information bits in the key set are bit-by-bit-determined. When the total number of decoding paths exceeds the preset path width L, the L paths with the largest PM value are selected and saved. Decoding the path, and finally obtaining the decoding result of each level of decoding bits according to the decoding path of each level of decoding bits. Since the first bit position bit is included in the key set with high probability, the decoding path will be high. The probability splits at the position of the first decoding error, so the correct path is retained in the current decoding path with a high probability, thereby obtaining better decoding performance, and the number of path splitting is greatly reduced, thereby reducing the number of paths. The total number of decoding paths, thus reducing the number of paths for which the PM value is to be calculated and the number of paths participating in the ordering, thus reducing computational complexity and decoding delay due to sequencing.

The decoding method provided in this embodiment first determines an information bit set according to the number of information bits of the information to be decoded and the target code length, and determines a key set according to the information bit set, K and N, and the key set is greater than the first threshold. The probability includes decoding the first wrong position bit, performing polarization decoding on the information to be decoded according to the key set during decoding, and decoding the information bit in the key set when the decoding path is generated, outside the key set The information bits are bit-by-bit, and since the first mis-location bit is included in the key set with a high probability, the decoding path will split at the first decoding error position with a high probability, so the correct path is reserved with high probability. In the current decoding path, better decoding performance is obtained, and the number of path splits is greatly reduced, which reduces computational complexity and decoding delay due to sorting. Therefore, the computational complexity and decoding delay of the Polar code decoding are reduced while ensuring better decoding performance.

The technical solutions of the method embodiment shown in FIG. 4 are described in detail below by using several specific embodiments.

FIG. 6 is a flowchart of another embodiment of a decoding method provided by the present application. The executor of this embodiment may be a network device or a terminal device as a sending device. In this embodiment, the number of information bits is K=9. The target code length is N=2 ⁴ as an example. In this embodiment, a key set is determined by constructing a full binary tree. As shown in FIG. 6, the method in this embodiment may include:

S201. Determine a set of information bits according to the number of information bits K of the information to be decoded and the target code length N.

In this embodiment, K=9, N=2 ⁴ =16, and the information bit set determined according to K and N is, for example,

The fixed bit set is accordingly:

S202. Construct a full binary tree with a depth n according to the information bit set, K and N. The information bits and the fixed bits in the full binary tree are leaf nodes.

With the information bits and fixed bits as leaf nodes, n=log ₂ (N)=4, construct a full binary tree with a depth of 4, as shown in FIG. 5, the full binary tree shown in FIG. 5 is performed according to the above dyeing rule. Dyeing, the black leaf node is the information bit, and the white leaf node is the fixed bit. For the non-leaf node, if the color of the left and right child nodes is the same, the color of the node is the same as the color of the child node, otherwise the node is colored as a gray node.

S203. Determine, from the full binary tree, all sub-blocks with a ratio of 1.

The sub-block with a ratio of 1 is a sub-tree with all leaf nodes being information bits, and the sub-block with a ratio of 1 includes a number of leaf nodes greater than or equal to 1.

S204. Determine, from all sub-blocks with a ratio of 1, an element that constitutes a key set, and the elements that make up the key set are the first information bits in each sub-block of ratio 1.

Specifically, taking the full binary tree shown in FIG. 5 as an example, the sub-block with a ratio of 1 includes A, B, C, and D sub-trees, and the elements constituting the key set are determined from the sub-trees of A, B, C, and D, respectively. The first information bit is taken into the key set, and the key set S={u ₆ , u ₇ , u ₁₁ , u ₁₃ }.

Further, as a second implementation manner, in order to improve the probability that the key set includes the bit of the first error location, the step may also be:

S204', determining elements constituting the key set from all sub-blocks having a ratio of 1, the elements constituting the key set are the first information bits in each sub-block of ratio 1 and all sub-blocks of ratio 1 At least one of the remaining information bits.

Specifically, after the first information bits are respectively taken out from the A, B, C, and D subtrees, since the D subtree contains only one information bit, from all the remaining information contained in the A, B, and C subtrees, At least one information bit can be taken out as an element constituting a key set.

S205. Perform polarization decoding on the information to be decoded according to the key set and the preset rule.

The preset rule is: performing coding path splitting on information bits in the key set, performing bit-by-bit decision on information bits outside the key set, and generating a decoding path, when the total number of decoding paths exceeds a preset path width L The L paths with the largest PM value are selected to save and continue to develop the decoding path, and the decoding result of each level of decoding bits is obtained according to the decoding path of each level of decoding bits.

The decoding method provided in this embodiment first determines an information bit set according to the number of information bits of the information to be decoded and the target code length, and constructs a full binary tree with a depth n according to the information bit set, K and N, and determines from the full binary tree. All sub-blocks with a ratio of 1 determine the elements that make up the key set from all sub-blocks with a ratio of 1, thereby determining the key set including the first bit of the decoded bit with high probability, and finally treating according to the key set and the preset rule. The decoding information is subjected to polarization decoding, and when the decoding path is generated, the information bits in the key set are split by the decoding path, and the information bits outside the key set are determined bit by bit, because the first wrong position bit is included with high probability In the key set, the decoding path will split at the first decoding error position with high probability, so the correct path is retained in the current decoding path with high probability, thus obtaining better decoding performance. However, the number of path splits is greatly reduced, which reduces computational complexity and decoding delay due to sorting. Therefore, the computational complexity and decoding delay of the Polar code decoding are reduced while ensuring better decoding performance.

FIG. 7 is a flowchart of another embodiment of a decoding method provided by the present application. The executor of the present embodiment may be a network device or a terminal device that is a sending device. In this embodiment, a key set is determined by querying a key set table. The key set can be determined in the manner shown in FIG. 6 and stored in the key set table in advance. The key set cannot be determined by constructing a full binary tree on the line, and the decoding delay can be further reduced. As shown in FIG. 7, the method in this embodiment may include :

S301. Determine a set of information bits according to the number of information bits K of the information to be decoded and the target code length N.

S302. Query a key set corresponding to the information bit set, K and N from the stored key set table, as a key set, and store the information bit set, the correspondence relationship between K and N and the key set in the key set table.

S303. Perform polarization decoding on the information to be decoded according to the key set and the preset rule.

The decoding method provided in this embodiment, after pre-storing the key set table, determining the information bit set according to the number of information bits of the information to be decoded and the target code length, and querying the information bit set from the stored key set table, K And a key set corresponding to N, thereby determining a key set including decoding the first error position bit with high probability, and finally performing polarization decoding on the information to be decoded according to the key set and the preset rule, comparing the embodiment shown in FIG. The method provided can further reduce the decoding delay without constructing a full binary tree online to determine a key set.

In the embodiment of the present application, in order to analyze the capability of the key set obtained by the methods shown in S202 to S204 to include the first decoding error bit in the SC decoding process, the first in the SC decoding process is counted by Monte Carlo simulation. The probability that a decoding error bit falls into the key set S. The specific simulation results are shown in Table 2. In Table 2, the polarization code length used in the simulation is N=1024, K=512, and the simulation times are T=10 ⁶ . The “falling into the set S” represents the first wrong position. The bits belong to the key set. The number of occurrences of S, "number of error frames" is the total number of frames in which the SC decoding process always fails, and "probability" is the ratio between "falling into the set S" and "number of error frames", "S size" The number of information bits included in the key set S.

Table II

It can be seen from Table 2 that even at low signal-to-noise ratio (E _b /N ₀ ), the probability that the key set of the SC decoding process contains the first bit of the decoded bit position is very close to 100%, with the signal-to-noise ratio The increase, the ability of the key set to contain the bits of the first wrong position is also increased, even the probability reaches 100%, and the key set contains only 25% of the information bits compared to the information bit K, which means Compared with the traditional CA-SCL decoder, the decoding method provided by the present application will reduce the path splitting number by 75%, thereby greatly reducing the decoding delay caused by sorting by CA-SCL decoding. And complexity.

The following describes the delay and complexity of the SCL decoding method and the decoding method provided by the present application. Compared with SC decoding, SCL decoding has a serial decoding delay inherent in SC decoding. The sorting delay (Sorting Latency) generated by filtering the 2L decoding paths, and L is the preset path width. For an N-long polarization code, the decoding delay due to the serial decoding feature is 2N-2. When the SC decoder is translated to the i-th information bit, the serial decoding delay due to serial decoding can be approximated by:

Latency _SC (i)≈2×i-2,i∈[1,N]

Assuming that the SCL decoder decodes the i-th information bit and all L decoding paths are processed in parallel at the same time, the decoding delay of the SCL decoder is:

It can be seen from the above equation that the decoding delay of the SCL decoding method is related to the information bit position when the maximum number of paths is reached, and the corresponding decoding delay can be reduced by extending the sequence number of the decoding information bits corresponding to the maximum number of paths. For a general SCL decoder, 2L paths are sorted when decoding to the log ₂ L information bits, for a polarization code with a code length of N and a code rate of R=K/N. The total decoding delay can be calculated using 2N-2+(K-log ₂ (L))·Latency _sort (L). Therefore, the decoding delay of the SCL decoder relative to the SC decoder, that is, the delay due to the ordering is

Latency _sorting =(K-log ₂ (L))·Latency _sort (L) (Equation 1)

The Latency _sort varies depending on the sorting algorithm used, and is generally a function of the maximum number of paths.

The computational complexity of the SCL decoding method can be calculated using the following formula:

Complexity _SCL-based (N,K,L)=L·(Nlog ₂ N)·Q+(K-log ₂ (L))·CAS(L)

The one-time f operation and the one-time g operation in SC decoding include five addition operations and one comparison operation. In the above formula, Q is the average complexity of one-time f calculation or one-time g operation. CAS(L) is the Compare-And-Select complexity of the sorting algorithm used, which varies according to the sorting network selected. The complexity due to sorting can be calculated by:

Complexity _sorting (N, K, L) = (K-log ₂ (L)) · CAS (L) (Formula 2)

Among them, the CAS corresponding to the bitonic sorter and the simplified bubble sorter is:

Unlike the conventional SCL decoding method, the decoding method provided by the present application (hereinafter referred to as Split-reduced SCL) will split the decoding path if and only if the information bits belong to the key set S, that is, the maximum number of Lists is reached. The time will be delayed, and the sorting delay of the Split-reduced SCL decoder is:

Latency _sorting (N,K,L)=(|S|-log ₂ (L))·Latency _sort (L) (Equation 3)

The decoding complexity due to sorting is:

Complexity _sorting =(|S|-log ₂ (L))·CAS(L) (Equation 4)

Combining (Equation 1)-(Equation 4), the delay and complexity gain of Split-reduced SCL decoding can be calculated by:

Where S is the number of information bits of the key set.

The performance of the Split-reduced SCL decoding method provided by the present application is described below in conjunction with the simulation comparison diagram. The specific parameters used in the simulation are shown in Table 3:

Table 3

Figure 8 shows the signal-to-noise ratio (SNR) required for CA-SCL decoding with Split-reduced decoding method to achieve 0.1% error block rate (BLER) and the performance of traditional CA-SCL decoding method. The comparison chart, the traditional CA-SCL decoding method is represented by Direct splitting in Figure 8. From the simulation results in Figure 8, it can be seen that the Split-reduced decoding method is compared to the traditional CA-SCL decoding method in BLER performance. With almost no loss, a decoding performance similar to that of the CA-SCL decoding method can be obtained. FIG. 9 is a schematic diagram of delay and complexity gain at different code rates by using a Split-reduced decoding method. It can be seen that the Split-reduced decoding method of the present application has a relative code rate and a code length configuration. The conventional SCL decoder has lower decoding delay and complexity.

10a-10f are BLER performance comparison diagrams of a different code length code rate configuration using a Split-reduced decoding method and a conventional CA-SCL decoding method. The conventional CA-SCL decoding method uses Direct in FIG. 10a-10f. Splitting indicates that the information bits K in Figures 10a-10f are 32, 48, 64, 80, 120, and 200, respectively. It can be seen from the simulation results that the Split-reduced decoding of the present application is performed under an arbitrary code length code rate configuration. The method can obtain the BLER decoding performance similar to the traditional CA-SCL decoding method, and has lower decoding delay and decoding complexity than the traditional CA-SCL decoder in decoding delay and complexity. degree.

FIG. 11 is a schematic structural diagram of an embodiment of a decoding device provided by the present application. As shown in FIG. 11, the apparatus in this embodiment may include: a first determining module 11, a second determining module 12, and a decoding module 13, where a determining module 11 is configured to determine an information bit set according to the information bit number K and the target code length N of the information to be decoded, and the second determining module 12 is configured to determine a key set according to the information bit set, K and N, the key The probability that the set includes the first bit of the decoded bit position is greater than the first threshold. Optionally, the first threshold is 99%. The key set is a subset of the information bit set, and the decoding module 13 uses And performing polarization decoding on the information to be decoded according to the key set and a preset rule, where the preset rule is: performing coding path splitting on information bits in the key set, for the key set The information bits outside are subjected to a bit-by-bit decision to generate a decoding path. When the total number of decoding paths exceeds the preset path width L, the L paths with the largest PM value are selected and saved, and the decoding path is continued, according to each Level decoding Patent obtain a decoding result of the decoding paths each decoding a bit.

The information path in the key set is split by a coding path, and the information bits outside the key set are determined bit by bit. Specifically, the information may be:

Optionally, the second determining module 12 is configured to:

Constructing a full binary tree of depth n according to the information bit set, K and N, n=log ₂ N, wherein the information bits and the fixed bits in the full binary tree are leaf nodes, and all ratios are determined from the full binary tree A sub-block of 1, a sub-block of ratio 1 is a sub-tree in which all leaf nodes are information bits, and a sub-block having a ratio of 1 includes a number of leaf nodes greater than or equal to 1, and is determined from all sub-blocks having a ratio of 1. The elements constituting the key set, the elements constituting the key set are the first information bits in each of the sub-blocks having a ratio of 1.

Optionally, the elements constituting the key set further include: at least one information bit of the remaining information bits in all the sub-blocks with a ratio of 1.

Optionally, the second determining module 12 is configured to: query, from the stored key set table, a key set corresponding to the information bit set, K and N, as the key set, where the information set bits are stored in the key set table The correspondence between sets, K, and N and key sets.

The device in this embodiment may be used to implement the technical solution of the method embodiment shown in FIG. 4, and the implementation principle and technical effects are similar, and details are not described herein again.

The application may divide the function module into the decoding device according to the above method example. For example, each function module may be divided according to each function, or two or more functions may be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of the modules in the embodiments of the present application is schematic, and is only a logical function division. In actual implementation, there may be another division manner.

FIG. 12 is a schematic structural diagram of a decoder provided by the present application, where the decoder 700 includes:

The memory 701 is configured to store program instructions, and the memory may be a flash memory.

The processor 702 is configured to call and execute program instructions in the memory to implement various steps in the decoding method shown in FIG. 4. For details, refer to the related description in the foregoing method embodiments.

Alternatively, the memory 701 may be independent or as shown in FIG. 13. FIG. 13 is a schematic structural diagram of another decoder provided by the present application. The memory 701 is integrated with the processor 702.

The decoder of Figures 12 and 13 also includes a transceiver (not shown) for transceiving signals through the antenna.

The decoder can be used to perform various steps and/or processes corresponding to the decoding device in the above method embodiments.

The present application also provides a readable storage medium having stored therein an execution instruction, when the at least one processor of the decoding device executes the execution instruction, the decoding device performs the decoding provided by the various embodiments described above. method.

The application also provides a program product comprising an execution instruction stored in a readable storage medium. At least one processor of the decoding device can read the execution instructions from a readable storage medium, and the at least one processor executes the execution instructions such that the decoding device implements the decoding methods provided by the various embodiments described above.

It will be understood by those skilled in the art that in the above embodiments, it may be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented in software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in accordance with embodiments of the present application are generated in whole or in part. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be from a website site, computer, server or data center Transfer to another website site, computer, server, or data center by wire (eg, coaxial cable, fiber optic, digital subscriber line (DSL), or wireless (eg, infrared, wireless, microwave, etc.). The computer readable storage medium can be any available media that can be accessed by a computer or a data storage device such as a server, data center, or the like that includes one or more available media. The usable medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a DVD), or a semiconductor medium (such as a solid state disk (SSD)).

Claims

A decoding method, comprising:

Determining an information bit set according to the information bit number K of the information to be decoded and the target code length N;

Determining, according to the information bit set, K and N, a key set, wherein the probability of decoding the first error location bit in the key set is greater than a first threshold, and the key set is a subset of the information bit set;

And performing polarization decoding on the to-be-decoded information according to the key set and the preset rule, where the preset rule is: performing coding path splitting on information bits in the key set, where the key set is The outer information bits are bit-by-bit-determined to generate a decoding path. When the total number of decoding paths exceeds the preset path width L, the L paths with the largest path metric PM value are selected and saved, and the decoding path is continued, according to each The decoding path of the primary decoding bits results in the decoding result of each level of decoding bits.
The method according to claim 1, wherein said performing coding path splitting on information bits in said key set, and performing bit-by-bit decision on information bits outside said key set comprises:

Where u i is the information bit in the information bit set, S is the key set, L(u i ) is the log likelihood ratio LLR of u i , h(·) is a hard decision function,
The method according to claim 1, wherein said determining a key set according to said set of information bits, K and N comprises:

Constructing a full binary tree of depth n according to the information bit set, K and N, n=log 2 N, wherein the information bits and the fixed bits in the full binary tree are leaf nodes;

All sub-blocks with a ratio of 1 are determined from the full binary tree, and a sub-block with a ratio of 1 is a sub-tree with all leaf nodes being information bits, and a sub-block with a ratio of 1 includes a number of leaf nodes greater than or equal to 1. ;

The elements constituting the key set are determined from all sub-blocks having a ratio of 1, and the elements constituting the key set are the first information bits in each sub-block of ratio 1.
The method of claim 3, wherein the elements constituting the key set further comprise: at least one of the remaining information bits in all sub-blocks of ratio 1.
The method according to claim 1, wherein said determining a key set according to said set of information bits, K and N comprises:

As the key set, the key set table corresponding to the information bit set, K and N is queried as the key set, and the key set table stores the information bit set, the correspondence relationship between K and N and the key set.
The method according to any one of claims 1 to 5, wherein the first threshold value is 99%.
A decoding device, comprising:

a first determining module, configured to determine an information bit set according to the information bit number K of the information to be decoded and the target code length N;

a second determining module, configured to determine, according to the information bit set, K and N, a key set, where the probability of decoding the first error location bit in the key set is greater than a first threshold, and the key set is the information a subset of the set of bits;

a decoding module, configured to perform polarization decoding on the information to be decoded according to the key set and a preset rule, where the preset rule is: performing coding path splitting on information bits in the key set, Performing a bit-by-bit decision on the information bits outside the key set to generate a decoding path. When the total number of decoding paths exceeds a preset path width L, the L paths with the largest path metric PM value are selected and saved. The decoding path obtains the decoding result of each level of decoding bits according to the decoding path of each level of decoding bits.
The device according to claim 7, wherein the decoding of the information bits in the key set is performed, and the information bits outside the key set are determined bit by bit, including:

Where u i is the information bit in the information bit set, S is the key set, L(u i ) is the log likelihood ratio LLR of u i , h(·) is a hard decision function,
The device according to claim 7, wherein the second determining module is configured to:

Constructing a full binary tree of depth n according to the information bit set, K and N, n=log 2 N, wherein the information bits and the fixed bits in the full binary tree are leaf nodes;

All sub-blocks with a ratio of 1 are determined from the full binary tree, and a sub-block with a ratio of 1 is a sub-tree with all leaf nodes being information bits, and a sub-block with a ratio of 1 includes a number of leaf nodes greater than or equal to 1. ;

The elements constituting the key set are determined from all sub-blocks having a ratio of 1, and the elements constituting the key set are the first information bits in each sub-block of ratio 1.
The apparatus of claim 9, wherein the elements constituting the key set further comprise: at least one of the remaining information bits in all of the sub-blocks having a ratio of one.
The device according to claim 7, wherein the second determining module is configured to:

As the key set, the key set table corresponding to the information bit set, K and N is queried as the key set, and the key set table stores the information bit set, the correspondence relationship between K and N and the key set.
Apparatus according to any one of claims 7 to 11, wherein said first threshold value is 99%.
A decoder, comprising:

a memory for storing program instructions;

And a processor, configured to invoke and execute program instructions in the memory to implement the decoding method according to any one of claims 1 to 6.
A computer readable storage medium for storing a computer program for performing the method of any of claims 1-6 when the computer is running.
A computer program product, comprising: a computer program for performing the method of any of claims 1-6 when the computer is running.