JP5062643B2 - Error correction coding apparatus, error correction coding method, and program - Google Patents

Description

  The present invention relates to an error correction encoding apparatus, an error correction encoding method, and a program, and more particularly, to an error correction encoding apparatus, an error correction encoding method, and a program based on redundancy of data representation.

  Today, in all digital communication fields, an error correction code technique for correcting an error generated in a communication path is used. Here, “communication” generally refers to information transmission between separation points performed via a communication medium such as wired or wireless, but is not limited to such a narrow meaning in this specification. For example, recording information on a recording medium such as a compact disk is also included in the concept of “communication”. Sending information is to carry information to a remote place, and recording is to carry information to a remote time (at a later date) assuming reading at a later date. Because it has the same meaning in terms of carrying.

  The code theory deals with digitized information. The data to be sent may be a digitalized audio signal or video signal, or digital data from a computer. The source of these information and the part that digitizes it are called information sources. The digitized information (data to be sent) is often expressed as a bit string using two kinds of symbols (numbers) of 0 and 1, as can be understood by considering a PCM signal, for example. For this reason, also in this specification, it is assumed that data to be sent is represented by a bit string or, more generally, a digit string.

FIG. 3 is a diagram illustrating a communication system model. In this figure, from the information source, 0 and 1 bit strings (information bit strings) are generated as data to be sent. The encoder (error correction encoding device) handles this information bit string by dividing it into blocks of every m bits. Referred to as "data to be sent" Each block represents a data to the feed _{i = (i 1, ····,} i m) at.

The encoder outputs an n-bit (n> m) code word w = (x ₁ ,..., X _n ) corresponding to the data i to be sent. That is, this encoder adds extra bits (redundant bits) of nm bits. n is the code length, m is the number of information bits, and this operation is called “encoding”. The code word sent to the communication path is called a transmission word w, and when the transmission word w is input, the communication path outputs an n-bit reception word v. If there is no influence of noise of the transmission / recording medium, the transmitted word is equal to the received word, but normally, due to noise or the like, 0 and 1 of each bit of the transmitted word are received differently with a certain probability. This is called an error, and this state is expressed as v = w + e, assuming that an error e is added to each bit in the communication path. However, an error of the j-th bit is set to e _j, and e represents a bit string e = (e ₁ ,..., E _n ) and is also called an error pattern. The decoder estimates which codeword is transmitted based on the received word v, and outputs an estimated value of the transmitted word w or an estimated value of the data i to be sent.

The actual transmission path or recording medium is sent out as a signal waveform corresponding to each bit x _j of the transmission word w, for example, “0” has no pulse and “1” has a pulse. However, for convenience of explanation, no specific waveform is mentioned in this specification.

Such encoding with a code length n and the number of information bits m is referred to as [n, m] encoding, and a set of 2 ^m codewords is referred to as [n, m] code (for example, Patent Document 1 and Patent Document 2).

More generally, the [n, m] linear error correction code on the Galois field GF (q) is an m-dimensional linear subspace of GF (q) ⁿ . GF (q) is a finite field whose original number is q. GF (q) ⁿ is a set of vectors (number vectors) that can be formed by arranging n elements of GF (q), and according to convention, a number vector on GF (q) is represented by a horizontal vector.

In conventional encoding using a linear code, data is encoded as follows. First, a base of an [n, m] linear error correction code (hereinafter referred to as a linear code C) is assumed to be (g ₁ ,..., G _m ). A string of length m to be encoded is represented by the following equation (13) as a number vector on GF (q). Specific examples of the linear code C include standard code theory textbooks (for example, “The Theory of Error-Correcting Codes”, FJ MacWilliams and NJA Slone, NORTH-HOLLAND, ISBN. : 0-444-85193-3 / 0-444-85009-0 / 0-444-85010-4), there are various types such as a Hamming code, a BCH code, and a Goppa code.

  Hereinafter, such data is referred to as a vector (of length m) or a vector of length m on GF (q). This is encoded into a vector f (x) belonging to the linear code C by the following equation (14). Of course, the calculation is performed on GF (q). For example, when q is a prime number, the calculation is performed using q numbers “0, 1,..., Q−1” and the modq calculation method.

When g _i = (g _i1 ,..., g _in ) (i = 1,..., m), the following expression (15) is called a generator matrix of the linear code C.

  Such a known encoding is [n, m] encoding as described above, and in order to specify the encoding specifically, the above encoding is specified by the generator matrix G [ n, m] encoding. In this [n, m] encoding, information for m digits is sent using n digits (elements of GF (q)). The linear code C (that is, [n, m] coding specified by the generator matrix G) is most often used when q = 2. At this time, GF (q) = {0, 1}. Note that when q = 2, the “digit” is called a “bit”. Specific examples of the linear code C and its generation matrix G are also described in various standard code theory textbooks.

JP 2005-110251 A Japanese Patent No. 3281938

  However, although the known [n, m] code described above has sufficient error correction capability, it has a problem in that the countermeasure against eavesdropping during transmission is incomplete and security is lacking. In other words, when a third party is eavesdropped on a transmission word w input to the communication path or a reception word v output from the communication path, the original data i to be sent can be easily estimated. is there.

  SUMMARY OF THE INVENTION An object of the present invention is to provide an error correction encoding device, an error correction encoding method, and a program in which security measures are taken.

  According to the first aspect of the present invention, the data to be sent has a length N ′ by [[N ′, K ′]] encoding specified by the M ′ × N ′ generator matrix F on the Galois field GF (q). Vector of

For the first encoding means for encoding to i = 1,..., N,

A vector of length n _i by [[n _i , k _i ]] encoding specified by the generator matrix G _i

And a second coding means for ^{coding, w (1), ····,} w is obtained by concatenating the ^{(N) [w (1)} ···· w (N)] as a coding result An error correction coding apparatus comprising output means for outputting.
According to the second aspect of the present invention, the length N ′ is obtained by [[N ′, K ′]] encoding specified by the M ′ × N ′ generator matrix F on the Galois field GF (q). Vector of

For the first encoding step for encoding to i = 1,..., N,

A vector of length n _i by [[n _i , k _i ]] encoding specified by the generator matrix G _i

And a second encoding step of ^{encoding, w (1), ····,} w is obtained by concatenating the ^{(N) [w (1)} ···· w (N)] as a coding result An error correction coding method including an output step of outputting.
According to the third aspect of the present invention, the data to be sent to the computer is long by [[N ′, K ′]] encoding specified by the M ′ × N ′ generator matrix F on the Galois field GF (q). N 'vector

For the first encoding means for encoding to i = 1,..., N,

A vector of length n _i by [[n _i , k _i ]] encoding specified by the generator matrix G _i

And a second coding means for ^{coding, w (1), ····,} w is obtained by concatenating the ^{(N) [w (1)} ···· w (N)] as a coding result This is a program for realizing output means for outputting.
According to a fourth aspect of the invention, there is provided second output means for further encrypting and outputting the encoding result output from the output means with a function g representing an arbitrary encryption method. The error correction coding apparatus according to Item 1.
The invention according to claim 5 includes a second output step of further encrypting and outputting the encoding result output from the output step with a function g representing an arbitrary encryption method. 2. The error correction coding method according to 2.
According to the sixth aspect of the present invention, the data to be sent to the computer is long by [[N ′, K ′]] encoding specified by the M ′ × N ′ generator matrix F on the Galois field GF (q). N 'vector

For the first encoding means for encoding to i = 1,..., N,

A vector of length n _i by [[n _i , k _i ]] encoding specified by the generator matrix G _i

And a second coding means for ^{coding, w (1), ····,} w is obtained by concatenating the ^{(N) [w (1)} ···· w (N)] as a coding result This is a program for realizing output means for outputting and second output means for further encrypting and outputting the encoding result output from the output means with a function g representing an arbitrary encryption method.

For convenience of explanation, the [[n, k]] code is set to [[4, 1]], the data to be sent is set to {0, 1}, and for example, “0” of the data to be sent is set to {0000, 1111. }, “1” of the data to be sent is encoded as {0011, 1100}. In the present invention, when data is transmitted, any one pattern in {} is randomly selected and transmitted. For example, since the pattern of data “0” is “0000” and “1111”, one of them is transmitted at random. In this way, even if a third party in the middle of transmission can wiretap a part of the pattern, the original data cannot be estimated. This is because, for example, even if “00” can be wiretapped, “00” is also included in the pattern of data “1”.
Therefore, according to the present invention, it is possible to provide an error correction encoding device, an error correction encoding method, and a program in which security measures are taken.

It is a conceptual block diagram of 1st Embodiment ([[n, k]] encoding). It is a figure which shows the notional process flow of 2nd Embodiment (two-stage encoding). It is a figure which shows the model of a communication system.

Embodiments of the present invention will be described below.
First, as a definition of words, a vector x = (x ₁ ,..., X _a ) of length _a and a vector y = (y ₁ ,..., Y _b ) of length b can be connected. A vector of length a + b (x ₁ ,..., X _a , y ₁ ,..., Y _b ) is represented by [xy]. At this time, y is connected to x to obtain [xy]. Three or more vectors are also enclosed in “[” and “]”, and similarly represent vectors that are concatenated. For example, [xyz] represents a vector formed by concatenating y to x and further concatenating z.

<First Embodiment ([[n, k]] encoding)>
FIG. 1 is a conceptual block diagram of the first embodiment ([[n, k]] encoding). This block schematically shows what can be used as an encoder in the communication system model diagram (FIG. 3) described at the beginning. A specific configuration of this block is not particularly mentioned. For example, it may be configured by hard logic, or may be configured by software that can be executed by a computer, all or a part of its main functions. In any case, it is sufficient that the configuration described below can be realized.

  The name of the entire block shown in the figure may be referred to as an encoder as in the communication model at the beginning, for example, but here, in order to distinguish it from the conventional [n, m] encoding, The [[n, k]] encoding unit 1 is used to denote the encoding by the [[n, k]] encoding unit 1, and the [[n, k]] encoding is referred to as the [[n, k]] encoding. Will be referred to as the

  The [[n, k]] encoding unit 1 includes at least a u generation unit 2, an x ′ generation unit 3, and an [n, m] encoding unit 4. The u generator 2 is, for example, a physical random number generator or a pseudo-random number generator, and the [n, m] encoder 4 performs an operation corresponding to the conventional [n, m] encoding. It is.

  In the first embodiment, the following new encoding is performed to ensure information security. In short, this method is to send k digits of information using the well-known [n, m] encoding described at the beginning. However, k ≦ m.

  Assume that the data x to be sent is as shown in the following equation (28).

The x ′ generator 3 concatenates this data x with a vector u = (x _{k + 1} ,..., x _m ) consisting of m−k digits randomly generated by the u generator 2. Thus, x ′ in the following formula (29) is obtained.

  The [n, m] encoding unit 4 uses the [n, m] encoding specified by the generation matrix G to convert x ′ calculated by the x ′ generation unit 3 into a vector f having a length n. It encodes to (x ').

Thus, in the first embodiment, a vector u = (x _{k + 1} ,..., X _m ) consisting of m−k digits randomly generated in the data x to be sent is concatenated. X ′ is encoded, and the x ′ is encoded into a vector f (x ′) of length n using [n, m] encoding, which is output.

  As described above, the encoding in the first embodiment is [[n, k]] encoding, and different encoding can be obtained by changing the generator matrix G. Therefore, the encoding is specifically specified. In order to do this, this is called [[n, k]] encoding specified by the generator matrix G. An image of this [[n, k]] encoding can be expressed as the following equation (30).

  The entire expression (30) corresponds to the [[n, k]] encoding unit 1 in FIG. Further, the x → x ′ portion corresponds to the x ′ generation unit 3, and the x ′ → f (x ′) portion corresponds to the [n, m] encoding unit 4.

Here, it will be described that the example given in the column of “Effects of the Invention” can be realized by the [[n, k]] encoding of the first embodiment. The base of the linear code C is (g ₁ ,..., G _m ), and the subspace spanned by (g _{k + 1} ,..., G _m ) is B, and n = 4, If k = 1 and m = 2 , an example of g ₁ and g ₂ can be written as the following equations (31) and (32), and an example of the linear code C can be expressed by the following equation ( 33 ): Moreover, an example of B can be expressed as the following equation ( 34 ). These are examples when GF (q) = GF (2).

As described above , g ₁ and g ₂ can be expressed as in the previous formula ( 31 ) and the previous formula ( 32 ). Then, as shown in the following equation ( 35 ), x ₂ = 0, 1 is randomly generated in the u generation unit 2, so whether the data to be sent is x ₁ = 0 or x ₁ = 1. In response, the output of [[n, k]] encoder 1 is {0000, 1111} or {1100, 0011}. Here, the vector (1, 1, 1, 1) is written as “1111”.

Thus, according to the first embodiment, a vector u = (x _{k + 1} ,..., X _m ) consisting of m−k digits randomly generated in the data x to be sent is By concatenating to generate x ′, the x ′ is encoded into a vector f (x ′) of length n using [n, m] encoding, thereby generating and outputting information for k digits. Therefore, the security of information can be ensured for the following reasons.

For simplification, the [[n, k]] code is a [[4, 1]] code and the data to be sent is {0, 1}. In this case, in the first embodiment, for example, when “0” of data to be transmitted is encoded as {0000, 1111} and “1” of data to be transmitted is encoded as {0011, 1100}, {
}, One of the patterns is randomly selected and transmitted. For example, since the pattern of data “0” is “0000” and “1111”, one of them is transmitted at random. Even if a third party in the middle of transmission can wiretap a part of the pattern, the original data cannot be estimated. This is because, for example, even if “00” can be wiretapped, “00” is also included in the pattern of data “1”.

  Therefore, in the encoding method ([[n, k]] encoding) of the first embodiment, it is possible to effectively prevent eavesdropping by a third party during transmission and to take security measures. The problems at the beginning of the present invention can be achieved.

<Second embodiment (two-stage encoding)>
The encoding method ([[n, k]] encoding) of the first embodiment described above is advantageous in that it can take security measures. On the other hand, it is known from theoretical analysis that a relatively long code length is advantageous in terms of decoding error probability and security. However, as the code length increases, the encoding and decoding procedures are increased. There is no denying that the complexity increases, and the complexity of this procedure may become an obstacle to practical use. As described above, [[n, k]] coding has an exceptionally beneficial merit that can prevent eavesdropping on the communication path, so it is a must to overcome such an obstacle factor (complexity of procedures). Both are technical issues that must be achieved. The second embodiment described below is for that purpose. To put it briefly, this second embodiment is based on the idea of division and governance (a method of solving a given problem by dividing it into several small problems and performing governance based on it and obtaining a solution to the problem) Divide and Conquer Law: Divide
Also called and Conquer. ) To overcome the above-mentioned obstacle factor by dividing the encoding process equivalent to [[n, k]] encoding having a large code length into several processes in stages. Thus, an error correction coding technique that can sufficiently withstand industrial use is provided.

  A second embodiment will be described. In the second embodiment, as described in detail below, as an example, a method of encoding using two linear codes is disclosed. Hereinafter, this encoding is referred to as “two-stage encoding”.

In this two-stage encoding, N> 0, n _i ≧ m _i ≧ k _i ≧ 0 (i = 1,..., N) is an integer (practically, n _i , m _i , k _i is i And [[n _i , k _i ]] encoding specified by the m _i × n _i generator matrix G _i and M ′ × N ′ generation on GF (q). Encoding is performed in combination with [[N ′, K ′]] encoding specified by the matrix F. This two-stage encoding has the same transmission efficiency as [[N ″, K ′]] encoding in the sense that K ′ digit information is sent using N ″ digits as a total. However, N ′ and N ″ are given by the following equation (36).

A specific two-stage encoding method is as follows. Here, the length of the transmission data is K ′.

  FIG. 2 is a diagram showing a conceptual processing flow of the second embodiment (two-stage encoding). This processing flow schematically represents what can be used as an encoder in the communication system model diagram (FIG. 3) described at the beginning. As in the first embodiment, this processing flow may be configured by hard logic, or all or some of its main functions are configured by software that can be executed by a computer. May be. In any case, it is sufficient that the configuration described below can be realized.

  In this processing flow, first, [[N ′, K ′]] specified by the M ′ × N ′ generator matrix F on GF (q) in [[N ′, K ′]] encoding in step S1. By encoding, arbitrary data to be sent is encoded into a vector of length N ′. This vector can be divided into N blocks and written as the following equation (37).

Next, with [[n _i , k _i ]] encoding in step S2, for i = 1,.

A vector of length n _i by [[n _i , k _i ]] encoding specified by the generator matrix G _i

Is encoded.

Then, in the last step ^{S3, w (1), ····} , obtained by connecting a ^{w (N) [w (1} ) ···· w (N)] and outputs as coding results.

Here, regarding the codes used in step S1, it is particularly useful that the [[N ′, K ′]] encoding specified by the M ′ × N ′ generator matrix F on GF (q) is GF (q). This is the case derived from the code on the extension field GF (q ^k ). In this case, for example, a Reed-Solomon code on GF (q ^k ) can be used. In practice, the [N, K] Reed-Solomon code on GF (q ^k ) is used as the [kN, kK] code on GF (q). For example, a Reed-Solomon code used for a compact disk (q = 2).

That is, in the [[N ′, K ′]] encoding in step S1, [N ′, M ′] encoding corresponding to the conventional [n, m] encoding can be used. [N, M] codes on GF (q ^k ) can be used as [kN, kM] codes on GF (q). However, N ′ = kN and M ′ = kM.

This is obtained by expressing an element of GF (q ^k ) as a vector of length k on GF (q). Specifically, it is as follows. Now, let b = (β _j ) ^k _{j = 1} be the basis of GF (q ^k ), which is a linear space on GF (q). Using this, any element ζ∈GF (q ^k ) can be written as in the following equation (40), so the number of row vectors (x ₁ ,... , X _k ) is represented by the symbol φ (ζ).

Corresponding “ξ = (ξ ₁ ,..., Ξ _N ) ← → [φ (ξ ₁ ) using this transformation“ ζ → φ (ζ) = (x ₁ ,..., X _k ) ” ... Φ (ξ _N )] ”, a vector ξ of length N on GF (q ^k ) and a vector ^k of length kN on GF (q) [φ (ξ ₁ ). (Ξ _N )] can be identified.

A specific example of two-stage encoding will be described. When n _i , m _i , and k _i do not depend on i, that is, n ≧ m ≧ k ≧ 0 is an integer, n _i = n, m _i = m, k _i = k (i = 1,...・ Specific examples in the case of N) are given below. In the above step S1, [kN, kM] encoding as described above is used. For example, if q = 2 and k = 4, [4 · 15,4M] encoding by Reed-Solomon code can be used. However, 0 <M ≦ 15. In this case, in the above step S1, the following equation (41) is obtained.

In step S2, [[n, k]] encoding specified by the same generator matrix G may be applied to y ₁ ⁽ⁱ⁾ ,..., Y _k ⁽ⁱ⁾ (i = 1). ,..., N). For example, when n = 6, [[6,4]] encoding specified by the generator matrix G in which the i-th row is g _i is performed as the following equations (42) to (46). Thus, the output of [[6, 4]] coding specified by the generator matrix G is a vector whose hamming weight is an even number.

In the above, the reason why the codes on the extension field GF (q ^k ) are handled is that efficient decoding can be performed using such codes. However, this merely illustrates the best (best mode) implementation in terms of performance. Therefore, the present invention is not limited to the case of using a code on this extension field GF (q ^k ).

As described above, the second embodiment adopts the concept of divide and conquer and divides the encoding process equivalent to [[n, k]] encoding having a large code length into several processes. Therefore, the encoding and decoding procedures can be efficiently performed, and the procedure can be simplified. Therefore, the [[n, k]] code is utilized while taking advantage of the coding method ([[n, k]] coding) of the first embodiment described above (security measures can be taken). Therefore, it is possible to provide a technique for error correction coding that can solve the drawbacks (complexity of coding and decoding procedures) associated with a long code length and can withstand practical use.

  Although the second embodiment relates to two-stage encoding, the second embodiment is not limited to two stages and includes three or more stages. This is because if there are more than two stages, two-stage encoding is included as part of the procedure. Therefore, it can be said that the essential matter for the invention is in the part relating to the two-stage encoding (step S1, step S2 in FIG. 2).

  The first embodiment relates to [[n, k]] encoding, and the second embodiment relates to two-stage encoding. These two embodiments can be constructed very simply, for example, by black-boxing conventional [n, m] coding, as shown in equation (30). Supplementing this, the following equation (47) is obtained.

  From this equation (47), it can be said that the two embodiments are encoding methods only for the part before being input to the “communication path or storage device”.

  Incidentally, although it is not the essence of the present invention, the content of the decoder corresponding to the above two embodiments can be generally expressed as the following equation (48) when “decoder” is also referred to.

  Here, the estimated values of x and u are respectively given by the following equation (49).

From the above, if the two-stage encoding is also viewed as a total, it can be regarded as equivalent to the method of the following equation (50), so the decoding principle of the above equation (4 8 ) can be applied.

  However, in terms of efficiency, it is desirable to apply two-stage decoding corresponding to two-stage encoding. This is similar to the known two-stage encoding = decoding of concatenated codes.

  In the first embodiment, the [n, m] encoding of the [n, m] encoding unit 4 is “specified by the generator matrix G”, which is given as a specific example. The generation matrix G is not limited to a specific code. Any [n, m] linear code generator matrix can be used, not limited to a specific generator matrix G. The same applies to the two-stage encoding in the second embodiment.

  In addition, as described above, the previous equation (47) means that the above-mentioned two embodiments are, in the end, an encoding method only for a portion before being input to the “communication path or storage device”. However, this does not restrict the usage of the present invention. That is, according to the previous equation (47), only the simplest usage mode in which the [[n, k]] encoded data f (x ′) is directly input to the communication path or the storage device is directly used. Although it may be imaged, it is merely an example of the mode of use of the present invention.

For example, in order to further increase the encryption strength, the use of this method in combination with another encryption method is one of the possible forms of use. This combined form will be described. Now, let g be a function representing an encryption method used together. Then, after f (x ′) is further encrypted to g (f (x ′)) and then g (f (x ′)) is input to the communication path or storage device, f (x Since the encryption strength of ′) is further reinforced by g, a stronger encryption method that is more practically preferable can be obtained. However, g currently depends on the encryption key K in the currently popular cryptosystem, and is often in the form of g = g _K as a notation, but in the above description, it is used together for the sake of simplification. The function representing the encryption method is simply g. Therefore, this g includes the form g = g _K.

1: [[n, k]] encoder (error correction encoding device)
2: u generation part (generation means, generation process)
3: x ′ generator (generator, generator)
4: [n, m] encoding unit (output means, output process)
S1: Step (first encoding means, first encoding step)
S2: Step (second encoding means, second encoding step)
S3: Step (output means, output process)

Claims (6)

generating means for randomly generating a vector u = (x _{k + 1} ,..., x _m ) composed of m−k digits ;
A vector u = (x _{k + 1} ,..., X _m consisting of m−k digits randomly generated by the generating means in the data x = (x ₁ ,..., X _k ) to be sent. ) To generate x ′ = [xu] = (x ₁ ,..., X _m ),
The present invention relates to an error correction coding apparatus for performing [[n, k]] coding, comprising: [n, m] coding of x ′ generated by the generating means and output means for outputting a vector of length n. Yes,
Divide and conquer processing with transmission efficiency equivalent to [[N ", K ']] encoding that uses a predetermined code length N" and transmits the data to be sent with a length K' that is a vector of K 'digits A means for performing division into a plurality of stages by law,
The data to be sent of length K ′ is designated by the M ′ × N ′ generator matrix F on the Galois field GF (q) , and the variables [n, k] in the [[n, k]] encoding are respectively N ′, A vector of length N ′ is obtained by [[N ′, K ′]] encoding obtained by substituting K ′.

First encoding means for encoding into:
obtained by the first encoding means for i = 1,..., N

The specified by m _i × n _i generator matrix G _i, the [[n, k]] variables in the encoding n, respectively k n _i, obtained by substituting the _{k i [[n i, k} i] ] A vector of length n _i by encoding

Second encoding means for encoding
[W ⁽¹⁾ ... W ^(N) ] obtained by concatenating w ⁽¹⁾ ,..., W ^(N) obtained by the second encoding means is output as an encoding result. and output means for,
In here, the variable N, n _i, m _i and k _{i is, N> 0, n i ≧} m i ≧ k i ≧ 0 (i = 1, ····, N) is an integer that satisfies, N ′ and N ″ are expressed by the following equation (36).

An error correction coding apparatus characterized by the above. a generating step for randomly generating a vector u = (x _{k + 1} ,..., x _m ) consisting of m−k digits ;
A vector u = (x _{k + 1} ,..., X _m) consisting of m−k digits randomly generated in the generating step in the data x = (x ₁ ,..., X _k ) to be sent. ) To generate x ′ = [xu] = (x ₁ ,..., X _m ),
The present invention relates to an error correction coding method for performing [[n, k]] coding, comprising [n, m] coding of x ′ generated in the generating step and outputting a vector of length n. Yes,
Divide and conquer processing with transmission efficiency equivalent to [[N ", K ']] encoding that uses a predetermined code length N" and transmits the data to be sent with a length K' that is a vector of K 'digits a step of performing divided into a plurality of stages by law,
The data to be sent of length K ′ is designated by the M ′ × N ′ generator matrix F on the Galois field GF (q) , and the variables [n, k] in the [[n, k]] encoding are respectively N ′, A vector of length N ′ is obtained by [[N ′, K ′]] encoding obtained by substituting K ′.

A first encoding step of encoding into
For i = 1,..., N, obtained in the first encoding step.

The specified by m _i × n _i generator matrix G _i, the [[n, k]] variables in the encoding n, respectively k n _i, obtained by substituting the _{k i [[n i, k} i] ] A vector of length n _i by encoding

A second encoding step of encoding into
[W ⁽¹⁾ ... W ^(N) ] obtained by concatenating w ⁽¹⁾ ,..., W ^(N) obtained in the second encoding step is output as an encoding result. and an output step of,
In here, the variable N, n _i, m _i and k _{i is, N> 0, n i ≧} m i ≧ k i ≧ 0 (i = 1, ····, N) is an integer that satisfies, N ′ and N ″ are expressed by the following equation (36).

An error correction coding method characterized by the above. generating means for randomly generating a vector u = (x _{k + 1} ,..., x _m ) composed of m−k digits ;
A vector u = (x _{k + 1} ,..., X _m consisting of m−k digits randomly generated by the generating means in the data x = (x ₁ ,..., X _k ) to be sent. ) To generate x ′ = [xu] = (x ₁ ,..., X _m ),
The present invention relates to an error correction coding program for [[n, k]] coding that realizes [n, m] coding of x ′ generated by the generating means and outputs a vector of length n. Is,
Divide and conquer processing with transmission efficiency equivalent to [[N ", K ']] encoding that uses a predetermined code length N" and transmits the data to be sent with a length K' that is a vector of K 'digits A means for performing division into a plurality of stages by law,
On the computer,
The data to be sent of length K ′ is designated by the M ′ × N ′ generator matrix F on the Galois field GF (q) , and the variables [n, k] in the [[n, k]] encoding are respectively N ′, A vector of length N ′ by [[N ′, K ′]] encoding obtained by substituting K ′

First encoding means for encoding into:
obtained by the first encoding means for i = 1,..., N

The specified by m _i × n _i generator matrix G _i, the [[n, k]] variables in the encoding n, respectively k n _i, obtained by substituting the _{k i [[n i, k} i] ] A vector of length n _i by encoding

Second encoding means for encoding
[W ⁽¹⁾ ... W ^(N) ] obtained by concatenating w ⁽¹⁾ ,..., W ^(N) obtained by the second encoding means is output as an encoding result. to realize and output means for,
In here, the variable N, n _i, m _i and k _{i is, N> 0, n i ≧} m i ≧ k i ≧ 0 (i = 1, ····, N) is an integer that satisfies, N ′ and N ″ are expressed by the following equation (36).

An error correction coding program characterized by the above.   2. The error correction coding apparatus according to claim 1, further comprising second output means for further encrypting and outputting the encoding result output from the output means with a function g representing an arbitrary encryption method. .   3. The error correction coding method according to claim 2, further comprising a second output step of further encrypting and outputting the encoding result output from the output step with a function g representing an arbitrary encryption method. . 4. The computer according to claim 3, further comprising: a second output unit configured to further encrypt and output an encoding result output from the output unit with a function g representing an arbitrary encryption method. Error correction coding program.

Images