Error correction method for non-binary deletion error based on Debrueine sequence

By combining MDS code and de Bruin sequence, DB-MDS code is constructed, and the problem of correcting non-binary deletion errors in high-density memory is solved, efficient error correction ability and high encoding efficiency are achieved, and it is suitable for unknown codeword boundary situations.

CN119010922APending Publication Date: 2024-11-22AEROSPACE XINTONG TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411014226.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-26
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In high-density memory, non-binary symbol deletion errors are difficult to effectively correct, and existing methods are inefficient in encoding or cannot correct errors under unknown codeword boundaries.

Method used

Using the MDS code based on the debrün sequence, the original binary information is encoded, and the DB-MDS code is constructed, combining the error correction ability of the MDS code and the compactness of the debrün sequence, the correction of non-binary delete errors is achieved. It is wrong, and there is no need to know the codeword boundary during the error correction process.

Benefits of technology

Complete error correction for non-binary deletion errors not exceeding a given number is achieved, with high encoding efficiency, and when the number of deletion errors is certain, the encoding efficiency approaches the optimality as the code length increases.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119010922A_ABST
    Figure CN119010922A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of communication coding, in particular to a non-binary deletion error correction method based on a Debrueine sequence, which comprises the following steps of: coding an original binary information sequence u to obtain a DB-MDS code word sequence X; the code word sequence X passes through a non-binary deletion error channel to obtain a code word sequence Y with a deletion error; the code word sequence Y is subjected to error positioning and sequence synchronization, a code word # imgabs0 # containing erasure errors is obtained, a DB symbol in the code word # imgabs1 # is removed, an MDS code word # imgabs2 # containing erasure errors is obtained, and the MDS code word # imgabs3 # containing erasure errors is subjected to MDS code error correction to obtain a binary information sequence # imgabs4 #. Complete error correction can be performed on no more than a given number of non-binary deletion errors.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of communication coding, in particular to a non-binary deletion error correction method based on a De Bruijn sequence. Background Art

[0002] In some high-density memories such as racetrack memory technology, data reading is performed by multiple read ports or read heads, where the basic unit of information storage is called a "cell". During the process of reading information, each cell is moved to the nearest read head position. However, once a cell starts to move, all other cells will move synchronously. This means that during the cell movement, multiple read heads may read a block of bits, i.e. non-binary symbols, in parallel at the same time. However, if a cell moves more than a cell position, it is possible that one or more cells are not read by each read head. This phenomenon may lead to the deletion of non-binary symbols or even the deletion of consecutive non-binary symbols (i.e. bursts).

[0003] To address the non-binary erasure error problem, current work has proposed some correction methods, such as introducing additional read / write heads to reconstruct the data, but this will lead to reduced coding efficiency; or using a "guess and check" coding scheme to locate the error and correct it. However, this method assumes that the receiver knows the boundaries of the codeword and cannot be applied to practical scenarios. Summary of the invention

[0004] The present invention provides a non-binary erasure error correction method based on a de Bruijn sequence, which can completely correct non-binary erasure errors that do not exceed a given number; at the same time, the method can complete error positioning and error correction without knowing the codeword boundary during the error correction process; in addition, the method has a high coding efficiency. When the number of erasure errors is constant, the coding efficiency approaches the optimum as the code length increases.

[0005] This application provides the following technical solutions:

[0006] A non-binary deletion error correction method based on De Bruijn sequence, characterized by comprising:

[0007] S1. Encode the original binary information u to obtain the DB-MDS codeword x i The code word x formed by the code word x and the DB-MDS code word sequence X further formed by the code word x;

[0008] S2, the code word sequence X passes through a non-binary erasure error channel to obtain a code word sequence Y with an erasure error;

[0009] S3, error location and sequence synchronization are performed on the code word sequence Y to obtain a code word sequence containing erasure errors The codeword sequence A codeword containing erasure errors constitute;

[0010] S4, code word The DB symbol in is removed to obtain the MDS codeword containing erasure errors.

[0011] S5, the MDS codeword containing the erasure error The binary information sequence is obtained by MDS code error correction

[0012] Further, encoding the original binary information i includes:

[0013] S11, a non-binary message block u of length K, that is, u=[u1,u2,…,u K ] is encoded into a bit rate of The length of the MDS codeword c is N, that is, c = [c1, c2, ..., c N ],in and Multiple MDS codewords c are connected in series to form an MDS codeword layer Right now where q c The domain of the MDS codeword symbol;

[0014] S12, DB code π * =[π1,π2,…,π B ] is repeated and concatenated to obtain a layer of DB code sequence Π=[π * ,π * ,…]=[π1,π2,…,π i ,…],in And put the DB code layer below or above the MDS code layer, then each MDS code character c i Both with a DB symbol π i , forming a DB-MDS code symbol where q π The domain of the DB symbol;

[0015] S13, every N DB-MDS code symbols x i Construct a DB-MDS(N,K,π L ) code, each DB-MDS(N,K,π L ) code consists of an MDS(N,K) code and a period of π LThe length of DB sequence is N. Each DB-MDS (N, K, π L ) code can correct a segment of length not exceeding π L -1 deletion error, where π L -1≤NK.

[0016] Further, the S3 includes:

[0017] S31, observe each received codeword containing deletion error The remaining DB sequence pattern in determines y j Whether there is a deletion error in j-1 represents the sequence synchronization compensation coefficient, that is, the jth codeword y j The total length of all previous deletion errors;

[0018] S32, if y is determined j A deletion error occurred, through y j The remaining DB sequence patterns in the position where the deletion error occurs;

[0019] S33, for y j Remove random caret at the location of the error Make y j becomes

[0020] Further, the MDS codeword c used is an RS code.

[0021] Principles and advantages of the present invention:

[0022] MDS code (Maximum Distance Separable) is a linear coding scheme with the maximum possible minimum Hamming distance. As one of the most common MDS codes, Reed-Solomon code is widely used in digital communication systems to combat non-binary burst errors.

[0023] De Bruijn sequence is a special sequence that has applications in many fields such as mathematics, computer science and cryptography. A De Bruijn sequence DB(k,n) is a sequence of k different symbols with a length of kn, and contains all possible kn different n-tuples exactly once in all consecutive subsequences of length n. It is often used to design efficient encoding and decoding schemes in the field of communications because of its compactness, periodicity and uniqueness.

[0024] The present invention combines the error correction capability of the MDS code with the compactness and integrity of the De Bruijn sequence to construct a non-binary error deletion correction code. First, the non-binary message block is encoded to obtain the MDS codeword, and then the MDS codeword symbol and the DB codeword symbol are used to form a DM-MDS codeword symbol. N DB-MDS codewords are combined to form a DB-MDS (N, K, π L ) code, that is, codeword x. This DB-MDS code can meet the needs of non-binary deletion error correction in data storage (caused by mechanical vibration or interference), ensuring that data can be accurately recovered. The De Bruijn sequence ensures complete data coverage, while the MDS code provides powerful error detection and correction capabilities.

[0025] Define RS(n,k) as the Galois field GF(2 m ) with a code rate of k / n, where 2 m -1=n+σ, σ≠0 represents the shortened RS code, and σ=0 represents the standard RS code. The generator matrix of the system RS code is represented by G k×n , which can be written as:

[0026]

[0027] where e k×k Represents the identity matrix of dimension k.

[0028] Let M j×k is the message matrix, where j is the number of codewords and k is the message length of the codeword. j×n is the codeword (group) matrix. The encoding process of RS code can be simply expressed as:

[0029] C j×n =M j×k G k×n (2)

[0030] Let G (k,:) , G (:,n) and G (k,n) Represents the matrix G k×n The k-th row, n-th column, and the element at the k-th row and n-th column. is an information symbol block to be encoded, and m i After RS ​​encoding, we can get:

[0031]

[0032] Let 2t = nk. Since the RS code is a cyclic code, the generating polynomial of the RS (n, k) code can be written as:

[0033] g(x) = (x-α)(x-α 2 )…(x-α 2t )=g0+g1x+…+g 2t-1 x 2t-1 +g 2t x 2t (4)

[0034] where α,α 2 ,…,α 2t It is the Galois Field GF(2 m ) on the primitive element, and i = 0, ..., 2t. It should be noted that the generator polynomial is also the codeword polynomial.

[0035] Since RS codes are Maximum Distance Separable (MDS) codes, RS codes can reach the Singleton bound. If c is a linear code with code length n, dimension k, and the minimum distance on the Galois field containing q elements is d, then the maximum number of codewords is q k .

[0036] The Singleton bound can be expressed as the following inequality:

[0037] q k ≤q n-d+1 , (5)

[0038] That is

[0039] k≤n-d+1 (6)

[0040] The Singleton bound can also be equivalently expressed as:

[0041] d≤n-k+1, (7)

[0042] That is, the maximum distance d is n-k+1, that is, the minimum distance of the MDS code is:

[0043] d=n-k+1 (8)

[0044] This means that RS(n,k) can correct errors, or nk erase errors, where an erase error is an error with a known error location.

[0045] From the above analysis, it can be seen that through the implementation of the present invention, non-binary erasure errors not exceeding a given number can be fully corrected in the case of unknown codeword boundaries, and the encoding and decoding structure is simple. That is, each DB-MDS (N, K, π L ) code can correct a segment of length not exceeding π L -1 deletion error, where πL -1≤NK.

[0046] This method also has a high coding efficiency, DB-MDS (N, K, π L )The overall bit rate is When the deletion error is certain, that is, given q π In this case, the coding efficiency (overall bit rate) approaches the optimum as the code length N increases, that is, where lim N→∞ log2q c =lim N→∞ log2q π q c ,because BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flowchart of a non-binary deletion error correction method based on De Bruijn sequence of the present invention;

[0048] Figure 2 It is a schematic diagram of the DB-MDS encoding method of the present invention. DETAILED DESCRIPTION

[0049] The following is further described in detail through specific implementation methods:

[0050] In some high-density memories such as racetrack storage technology, data reading is performed by multiple read ports or read heads, where the basic unit of information storage is called a "cell". In the process of reading information, each cell is moved to the nearest read head position. Once a cell starts to move, all other cells will move synchronously. This means that during the movement of the cell, multiple read heads may read a block of bits, i.e., non-binary symbols, in parallel at the same time. However, if the movement distance of a cell exceeds the position of a cell, it is possible that one or more cells are not read by each read head. This phenomenon may lead to the deletion of non-binary symbols, and may even lead to the deletion of continuous non-binary symbols (i.e., bursts), and corresponding error correction methods need to be developed to recover from the deletion of non-binary symbols.

[0051] The present invention combines the MDS code with the De Bruijn sequence to construct a non-binary erasure error correction code, which can meet the needs of non-binary erasure error correction (caused by mechanical vibration or interference) in data storage. Among them, the MDS code (Maximum Distance Separable) code is a linear coding scheme with the maximum possible minimum Hamming distance. As one of the most common MDS codes, the Reed-Solomon code is widely used in digital communication systems to combat non-binary burst errors; and the De Bruijn sequence is a special sequence, which is a sequence composed of k different symbols, with a length of kn, and in all continuous subsequences of length n of the sequence, all possible kn different n-tuples are included exactly once. Because of its compactness, periodicity and uniqueness, it is often used in the design of efficient encoding and decoding schemes in the field of communication.

[0052] Embodiment 1

[0053] like Figure 1 As shown, a non-binary deletion error correction method based on De Bruijn sequence comprises the following steps:

[0054] S1. Encode the original binary information u to obtain the DB-MDS codeword x i The code word x formed by the code word x and the DB-MDS code word sequence X further formed by the code word x;

[0055] like Figure 2 As shown, the process of encoding the original binary information i includes:

[0056] S11, a non-binary message block u of length K, that is, u=[u1,u2,…,u K ] is encoded into a bit rate of The length of the MDS codeword c is N, that is, c = [c1, c2, ..., c N ],in and Multiple MDS codewords c are connected in series to form an MDS codeword layer Right now Wherein, qc is the field of the MDS codeword symbol; the MDS codeword c in this embodiment uses the coding scheme of the RS code.

[0057] S12, DB code π * =[π1,π2,…,π B ] is repeated and concatenated to obtain a layer of DB code sequence Π=[π * ,π * ,…]=[π1,π2,…,π i ,…],in And put the DB code layer below or above the MDS code layer, then each MDS code character c i Both with a DB symbol π i , forming a DB-MDS code symbol where q π The domain of the DB symbol;

[0058] S13, every N DB-MDS code symbols x i Construct a DB-MDS(N,K,π L ) code, each DB-MDS(N,K,π L ) code consists of an MDS(N,K) code and a period of π L The length of DB sequence is N. Each DB-MDS (N, K, π L ) code can correct a segment of length not exceeding π L -1 deletion error, where π L -1≤NK.

[0059] S2, the code word sequence X passes through a non-binary erasure error channel to obtain a code word sequence Y with an erasure error;

[0060] S3, error location and sequence synchronization are performed on the code word sequence Y to obtain a code word sequence containing erasure errors The codeword sequence A codeword containing erasure errors The composition specifically includes:

[0061] S31, by observing each received codeword containing deletion errors The remaining DB sequence pattern in determines y j Whether there is a deletion error in j-1 represents the sequence synchronization compensation coefficient, that is, the jth codeword y j The total length of all previous deletion errors;

[0062] S32, if y is determined j A deletion error occurred, through y j The remaining DB sequence patterns in the position where the deletion error occurs;

[0063] For example, define the period as π L =8 DB sequence is If a deletion error with an error length of 3 occurs at the 5th to 7th position, the remainder of the DB sequence π becomes in Since only deletion errors occur, and the length of deletion errors does not exceed π L-1=7, the receiving end can easily determine that the deleted symbols "4,5,6" occur at Between the 4th symbol "3" and the 5th symbol "7".

[0064] S33, for y j Remove random caret at the location of the error Make y j becomes

[0065] S4, code word The DB symbol in is removed to obtain the MDS codeword containing erasure errors.

[0066] Remove DB-MDS codewords by decoding The DB code layer in the code is used to obtain the MDS codeword containing erasure errors.

[0067] S5, the MDS codeword containing the erasure error The binary information sequence is obtained by MDS code error correction

[0068] The MDS code error correction process includes error detection, error location, error value calculation and data recovery.

[0069] Error detection is accomplished by calculating the adjoint matrix of the received data. If the adjoint matrix is ​​non-zero, an error is present.

[0070] If an error is detected, the specific location where the error occurs is determined by using a polynomial of the error location;

[0071] Then calculate the error value of each error symbol;

[0072] Finally, the error value is added to the corresponding error symbol to restore the original encoded data.

[0073] The error correction method provided by the present invention has many advantages, including:

[0074] 1) It can fully correct non-binary errors that do not exceed a given number in the case of unknown codeword boundaries, and the coding structure is simple. That is, each DB-MDS (N, K, π L ) code can correct a segment of length not exceeding π L -1 deletion error, where π L -1≤NK.

[0075] 2) This method has a high coding efficiency, DB-MDS(N,K,π L )The overall bit rate is When the deletion error is certain, that is, given q π In this case, the coding efficiency (overall bit rate) approaches the optimum as the code length N increases, that is, where lim N→∞ log2q c =lim N→∞ log2q π q c ,because

[0076] Embodiment 2

[0077] The difference between this embodiment and the first embodiment is that other commonly used check code types other than RS codes are used, including CRS code (Cauchy Reed-Solomon code), Evenodd code (Even-odd, parity check) and RDP code (ReducedRedundancy Parity, reduced redundant parity check), etc. The encoding and decoding processing steps are the same as those in the first embodiment, and are not repeated here.

[0078] The above are only embodiments of the present invention. The invention is not limited to the field involved in this implementation case. The common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A non-binary deletion error correction method based on De Bruijn sequence, characterized in that: include: S1. Encode the original binary information u to obtain the DB-MDS codeword x i The code word x formed by the code word x and the DB-MDS code word sequence X further formed by the code word x; S2, the code word sequence X passes through a non-binary erasure error channel to obtain a code word sequence Y with an erasure error; S3, error location and sequence synchronization are performed on the code word sequence Y to obtain a code word sequence containing erasure errors The codeword sequence A codeword containing erasure errors constitute; S4, code word The DB symbol in is removed to obtain the MDS codeword containing erasure errors. S5, the MDS codeword containing the erasure error The binary information sequence is obtained by MDS code error correction 2. The method for correcting non-binary deletion errors based on De Bruijn sequence according to claim 1, characterized in that: The encoding of the original binary information u comprises: S11, a non-binary message block u of length K, that is, u=[u1,u2,…,u K ] is encoded into a bit rate of The length of the MDS codeword c is N, that is, c = [c1, c2, ..., c N ],in and Multiple MDS codewords c are connected in series to form an MDS codeword layer Right now where q c The domain of the MDS codeword symbol; S12, DB code π * =[π1,π2,…,π B ] is repeated and concatenated to obtain a layer of DB code sequence Π=[π * ,π * ,…]=[π1,π2,…,π i ,…],in And put the DB code layer below or above the MDS code layer, then each MDS code character c i Both with a DB symbol π i , forming a DB-MDS code symbol where q π The domain of the DB symbol; S13, every N DB-MDS code symbols x i Construct a DB-MDS(N,K,π L ) code, each DB-MDS(N,K,π L ) code consists of an MDS(N,K) code and a period of π L The length of DB sequence is N. Each DB-MDS (N, K, π L ) code can correct a segment of length not exceeding π L -1 deletion error, where π L -1≤NK.

3. The method for correcting non-binary deletion errors based on De Bruijn sequence according to claim 2, characterized in that: The S3 includes: S31, observe each received codeword containing deletion error The remaining DB sequence pattern in determines y j Whether there is a deletion error in j-1 represents the sequence synchronization compensation coefficient, that is, the jth codeword y j The total length of all previous deletion errors; S32, if y is determined j A deletion error occurred, through y j The remaining DB sequence patterns in the position where the deletion error occurs; S33, for y j Remove random caret at the location of the error Make y j becomes 4. The method for correcting non-binary deletion errors based on de Bruijn sequences according to claim 2, characterized in that: The MDS code word c is an RS code.