LDPC (Low Density Parity Check Code) decoding method for reducing error floor based on column layering

Abstract

The invention discloses an LDPC (Low Density Parity Check Code) decoding method for reducing an error floor based on column layering, and belongs to the technical field of channel coding and decoding. The decoding of the LDPC code is easily influenced by a trap set in a high signal-to-noise ratio region, and an information transmission path is blocked in the decoding process, so that the decoding fails, the performance curve declines gently, and the application of the LDPC code in some scenes requiring a low error flat layer is limited. According to the method, all variable nodes in the decoding process are grouped according to columns, all variable nodes in one group are processed in parallel, after decoding fails due to the fact that decoding enters a trap set, the post-processing process is added, the weight of information transmitted to the variable nodes by check nodes is modified according to the check results of the check nodes, and decoding is carried out again to obtain a final result. According to the method, the decoding process can be effectively enabled to jump out of a trap set, the performance of the LDPC code in a high signal-to-noise ratio region is obviously improved, and an error floor is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of channel coding in the field of communication, and in particular relates to an LDPC code decoding method based on column stratification for reducing error levels. Background Art

[0002] In modern communication systems, channel coding is one of the core technologies to improve the reliability of data transmission. With the rapid development of communication technology, the data transmission rate continues to increase. Signals are easily affected by factors such as noise and interference during transmission, resulting in data errors. Channel coding introduces redundant information into the information, allowing the receiver to correct errors when receiving signals containing errors, thereby effectively improving the reliability of data. The continuous innovation and optimization of channel coding technology is the basis for ensuring the efficient and stable operation of modern communication systems.

[0003] LDPC code is a commonly used channel coding scheme. Through iterative decoding methods such as the belief propagation algorithm, it can approach the Shannon limit under low signal-to-noise ratio conditions and achieve efficient error correction. With the improvement of computing power, the research on LDPC codes has received widespread attention and has made significant progress in the field of communications. Compared with traditional error correction codes, LDPC codes have stronger resistance to noise interference and perform well in large-scale data transmission scenarios. LDPC codes have been widely used in modern communication technologies, such as wireless communication standards such as Wi-Fi, LTE, and 5G, as well as satellite communications, fiber-optic communications and other fields, and have become one of the key technologies in efficient and high-speed communication systems.

[0004] Although LDPC codes have excellent error correction performance, they face the problem of error flooring under certain channel conditions. Error flooring is the phenomenon that the decoding bit error rate tends to be stable and cannot be further reduced when the signal-to-noise ratio is very high. When the signal-to-noise ratio of the signal is high enough, the bit error rate of the LDPC code will drop significantly after a certain number of iterative decoding, but after a certain point, the bit error rate will no longer decrease significantly, but will remain at a lower level. This phenomenon is usually due to certain special types of error modes, especially error modes related to the codeword structure, such as trap sets, which are difficult to be effectively corrected by the iterative decoding process. Solving the error flooring problem is crucial to further improve the performance of LDPC codes, especially in application scenarios such as satellite communications that require extremely high data reliability. The error flooring phenomenon will significantly affect the actual performance of the system. Therefore, studying the LDPC code decoding method that reduces the error floor is of great significance to improving the application effect of LDPC codes in actual communication systems. Summary of the invention

[0005] Purpose of the invention: In order to solve the error floor problem of LDPC code decoding, the present invention proposes an LDPC code decoding method based on column layering to reduce the error floor. This decoding method can effectively make the decoding process jump out of the trap set by adjusting the weight of the information transmitted from the check node to the variable node in the case where the codeword enters the trap set during the decoding process, thereby improving the decoding success rate and reducing the error floor.

[0006] Technical solution: A LDPC code decoding method based on column stratification to reduce the error floor. This method groups all variable nodes by column, and processes all variable nodes in a group in parallel during the decoding process. Due to the influence of the trap set, when the decoding enters the trap set and causes the decoding failure, a post-processing process is added. The post-processing first obtains the verification results of all check nodes, and adjusts the weight of the information transmitted by the check nodes to the variable nodes in the post-processing process according to the verification results, that is, increases the weight of the information transmitted by the check nodes that failed the verification, and reduces the weight of the information transmitted by the check nodes that partially passed the verification, and decodes again to obtain the final decoding result. By modifying the weight of the information transmitted by the check nodes to the variable nodes, this method adds a certain disturbance to the decoding process, so that the codeword can jump out of the current trap set, improve the success rate of decoding, and effectively improve the error floor. The specific steps are as follows:

[0007] 1) Post-processing initialization: record the verification results of the check nodes after the decoding fails due to entering the trap set; group the variable nodes, that is, layer them by columns; initialize the number of iterations to one;

[0008] 2) Mark the check nodes that need to improve the weight of information transmission: According to the verification results of the check nodes, mark the check node set C1 that failed the verification; the check nodes in the set C1 are the check nodes that need to improve the weight of information transmission;

[0009] 3) Mark the check nodes that need to reduce the weight of information transmission: through the Internet network that transmits information from the check nodes to the variable nodes, mark all the variable node sets V1 connected to the check node set C1 that failed the verification; through the Internet network that transmits information from the variable nodes to the check nodes, mark all the check node sets C2 connected to the variable node set V1, where the set C1 is a subset of the set C2; define the set C3 as the check nodes in the set C2 except the set C1; the check nodes in the set C3 are the check nodes that need to reduce the weight of information transmission;

[0010] 4) Calculate the information that the check node transmits to the adjacent variable node: when the number of iterations is equal to one, adjust the weight of the information transmitted by the check node to the variable node according to the marking result, set the absolute value of the information transmitted by the check node in set C1 to the adjacent variable node to S1, and set the absolute value of the information transmitted by the check node in set C3 to the adjacent variable node to S2, where S1 and S2 are respectively the absolute value of the enhanced information and the absolute value of the weakened information, which can be simulated and determined according to the main trap set structure of the specific LDPC codeword, and the absolute value of the information transmitted by other unmarked check nodes is the minimum value or the second minimum value of the information stored in the check node, and the absolute value of the information is combined with the sign bit and transmitted to the variable node; when the number of iterations is greater than one, according to the result of sorting the information stored in the check node, if the transmitted variable node is the source of the minimum value, then select the second minimum value and the calculated sign bit to be combined and transmitted to the variable node, otherwise select the minimum value and the calculated sign bit to be combined and transmitted;

[0011] 5) Calculate the information passed by the variable node to the adjacent check node: calculate the information passed by the variable node in a group to the check node adjacent to the current variable node each time, the information passed is the sum of the initialization information of the variable node and the information passed to the variable node by all the check nodes connected to the variable node minus the information passed to the variable node by the passed check node;

[0012] 6) Update the information of the check node: sort the information received by each check node from the adjacent variable nodes according to the absolute value, update the minimum and second minimum absolute values ​​of the information received by each check node from the variable nodes and the variable nodes from which they come, and record the signs of all the information;

[0013] 7) Repeat steps 4) to 6) until all variable nodes are processed, which is recorded as one iteration; make a decoding decision. When the information completely passes the check matrix, the decoding is terminated. Otherwise, when the number of iterations is less than the maximum number of post-processing iterations, the number of iterations is increased by one, and steps 4) to 6) are repeated. When the number of iterations reaches the maximum number of post-processing iterations, the decoding is terminated.

[0014] The beneficial effect of the method is as follows: when the initial decoding fails due to entering the trap set, according to the check result of the check node, the check node that currently fails the check is the check node in the trap set structure. Through the structure of the check matrix and the interconnection network that transmits information between the check node and the variable node, the variable node set V1 with decoding failure, the check node set C1 with higher confidence in the check result, and the check node set C3 with lower confidence in the check result can be marked. In the first iteration of the post-processing decoding, by enhancing the information transmitted to the connected variable nodes by the check node set C1 and weakening the information transmitted to the connected variable nodes by the check node set C3, the variable node set V1 can be effectively The variable nodes of some decoding failures in V1 are flipped, so that the decoding process jumps out of the trap set and finally the decoding is successful; the method is widely applicable to different LDPC codewords, and the values ​​of enhanced or weakened information strength S1 and S2 can be selected according to the structure of the check matrix of the LDPC codeword and the structure of the corresponding main trap set. At the same time, in some scenarios with extremely high requirements for low error floor, the post-processing stage can be repeated, and different S1 and S2 parameters can be used in different post-processing stages, which further enhances the effect of reducing the error floor; the method only needs to add modules related to post-processing marks on the basis of the decoder architecture based on column layering to realize the post-processing function, and the overhead is low during the hardware implementation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The present invention is a flowchart of a method for decoding LDPC codes based on column stratification to reduce error levels.

[0016] Figure 2 The figure is a schematic diagram of the architecture of the LDPC code decoder for reducing error levels based on column stratification according to the present invention. DETAILED DESCRIPTION

[0017] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.

[0018] In this example, the specific steps of the complete LDPC code initial decoding process and post-processing process based on column layering are as follows:

[0019] 1) Decoding initialization: Divide the variable nodes into L groups, i.e., layer by column; initialize each variable node, and obtain the initialization log-likelihood ratio of each variable node according to the information of the transmission channel; set the maximum number of initial decoding iterations M and the maximum number of post-processing iterations N, and the current number of iterations is one;

[0020] 2) Calculate the information passed by the check node to the adjacent variable nodes: In the initial decoding stage, when the number of iterations is equal to one, since the check node has not obtained the information of all adjacent variable nodes, the information passed by the check node to the adjacent variable nodes is zero. When the number of iterations is greater than one, according to the sorted results of the information received by the check node, if the transmitted variable node is the source of the minimum value, the next smallest value is selected and combined with the calculated sign bit and passed to the variable node, otherwise the minimum value is selected and combined with the calculated sign bit for transmission;

[0021] 3) Calculate the information passed by the variable node to the adjacent check node: calculate each time the information passed by the variable node in a group to the check node adjacent to the current variable node. When the number of iterations is one, the information passed is the initialization information of the variable node. When the number of iterations is greater than one, the information passed is the sum of the initialization information of the variable node and the information passed to the variable node by all the check nodes connected to the variable node minus the information passed to the variable node by the passed check node.

[0022] 4) Update the information of the check node: sort the information received by each check node from the adjacent variable nodes according to the absolute value, update the minimum and second minimum absolute values ​​of the information received by each check node from the variable nodes and the variable nodes from which they come, and record the signs of all the information;

[0023] 5) Repeat steps 2) to 4) L times, which is recorded as one iteration; make a decoding decision. When the information completely passes the check matrix, the decoding is terminated. Otherwise, when the number of iterations is less than M, the number of iterations is increased by one, and steps 2) to 4) are repeated. When the number of iterations is equal to M, enter the post-processing stage and initialize the number of iterations to one again;

[0024] 6) Mark the check nodes that need to modify the weight: according to the check result of the check node, mark the check node set C1 that failed the check; through the Internet network that transmits information from the check node to the variable node, mark all the variable node sets V1 connected to the check node set C1 that failed the check; through the Internet network that transmits information from the variable node to the check node, mark all the check node sets C2 connected to the variable node set V1, where set C1 is a subset of set C2, and set C3 is defined as the check nodes in set C2 except set C1; the check nodes in set C1 are marked as the check nodes that need to increase the weight of the information transmitted, and the check nodes in set C3 are marked as the check nodes that need to reduce the weight of the information transmitted;

[0025] 7) Calculate the information passed by the check node to the adjacent variable node: In the post-processing decoding stage, when the number of iterations is one, the absolute value of the information passed by the check node in set C1 to the adjacent variable node is set to S1, the absolute value of the information passed by the check node in set C3 to the adjacent variable node is set to S2, and the absolute value of the information passed by other unmarked check nodes is the minimum value or the second minimum value of the information stored in the check node, and the absolute value of the information is combined with the sign bit and passed to the variable node; when the number of iterations is greater than one, according to the result of sorting the information stored in the check node, if the variable node to be passed is the source of the minimum value, the second minimum value is selected and combined with the calculated sign bit and passed to the variable node, otherwise the minimum value is selected and combined with the calculated sign bit for transmission;

[0026] 8) Calculate the information transmitted by the variable node to the adjacent check node: calculate the information transmitted by the variable node in a group to the check node adjacent to the current variable node each time, the transmitted information is the sum of the initialization information of the variable node and the information transmitted to the variable node by all the check nodes connected to the variable node minus the information transmitted to the variable node by the transmitted check node;

[0027] 9) Update the information of the check node: sort the information received by each check node from the adjacent variable nodes according to the absolute value, update the minimum and second minimum absolute values ​​of the information received by each check node from the variable nodes and the variable nodes from which they come, and record the signs of all the information;

[0028] 10) According to the marking result, repeat steps 7) to 9) L times, which is recorded as one iteration; make a decoding decision. When the information completely passes the check matrix, the decoding is terminated. Otherwise, when the number of iterations is less than N, the number of iterations is increased by one, and steps 7) to 9) are repeated. When the number of iterations is equal to N, the decoding is terminated.

Claims

1. A LDPC code decoding method based on column layering to reduce error level, characterized in that The following steps are involved: 1) Post-processing initialization: record the verification results of the check nodes after the decoding fails due to the decoding entering the trap set; Group variable nodes, i.e., stratify by columns; Initialize the number of iterations to one; 2) Mark the check nodes that need to improve the weight of information transmission: According to the verification results of the check nodes, mark the check node set C1 that failed the verification; the check nodes in the set C1 are the check nodes that need to improve the weight of information transmission; 3) Mark the check nodes that need to reduce the weight of information transmission: mark all variable node sets V1 connected to the check node set C1 that failed the verification; then mark all check node sets C2 connected to the variable node set V1, where set C1 is a subset of set C2; define set C3 as the check nodes in set C2 except set C1; the check nodes in set C3 are the check nodes that need to reduce the weight of information transmission; 4) Calculate the information that the check node transmits to the adjacent variable node: when the number of iterations is one, adjust the weight of the information transmitted by the check node to the variable node according to the marking result, set the absolute value of the information transmitted by the check node in set C1 to the adjacent variable node to S1, set the absolute value of the information transmitted by the check node in set C3 to the adjacent variable node to S2, and the absolute value of the information transmitted by other unmarked check nodes is the minimum value or the second minimum value of the information stored in the check node, and the absolute value of the information is combined with the sign bit and transmitted to the variable node; when the number of iterations is greater than one, according to the result of sorting the information stored in the check node, if the transmitted variable node is the source of the minimum value, select the second minimum value and the calculated sign bit to be combined and transmitted to the variable node, otherwise select the minimum value and the calculated sign bit to be combined and transmitted; 5) Calculate the information passed by the variable node to the adjacent check node: calculate the information passed by the variable node in a group to the check node adjacent to the current variable node each time, the information passed is the sum of the initialization information of the variable node and the information passed to the variable node by all the check nodes connected to the variable node minus the information passed to the variable node by the passed check node; 6) Update the information of the check node: sort the information received by each check node from the adjacent variable nodes according to the absolute value, update the minimum and second minimum absolute values ​​of the information received by each check node from the variable nodes and the variable nodes from which they come, and record the signs of all the information; 7) Repeat steps 4) to 6) until all variable nodes are processed, which is recorded as one iteration; make a decoding decision. When the information completely passes the check matrix, the decoding is terminated. Otherwise, when the number of iterations is less than the maximum number of post-processing iterations, the number of iterations is increased by one, and steps 4) to 6) are repeated. When the number of iterations reaches the maximum number of post-processing iterations, the decoding is terminated.

2. The process of marking check nodes that need to increase the weight of information transmission according to claim 1, characterized in that: The verification results of the check nodes after the initial decoding fails are stored, and the check node set C1 that fails the verification is marked as the check nodes that need to increase the weight of the transmitted information.

3. The process of marking check nodes that need to reduce the weight of information transmission according to claim 1, characterized in that: The Internet network that transmits information from the check nodes to the variable nodes marks all the variable node sets V1 connected to the check node set C1 that failed the verification; the Internet network that transmits information from the variable nodes to the check nodes marks all the check node sets C2 connected to the variable node set V1, where set C1 is a subset of set C2, and set C3 is defined as the check nodes in set C2 except set C1; the check nodes in set C3 are the check nodes that need to reduce the weight of the information transmitted.

4. The process of calculating information transmitted from a check node to an adjacent variable node according to claim 1, characterized in that: In the first iteration, it is necessary to adjust the weight of information transmitted from the check node to the variable node according to the marking result. The absolute value of the information transmitted by the check node marked as needing to increase the weight of information transmitted is S1, and the absolute value of the information transmitted by the check node marked as needing to reduce the weight of information transmitted is S2. The absolute value of information transmitted by other check nodes is the minimum value or the second smallest value, and the absolute value of the information is combined with the sign bit and transmitted to the variable node, where S1 and S2 are respectively the absolute value of the enhanced information and the absolute value of the weakened information, which can be simulated and determined according to the main trap set structure of the specific LDPC codeword.

Images