1 / 45

The Theory and Technology of Error Control Coding

The Theory and Technology of Error Control Coding. Chapter 2 Block code ----BCH. Summary. The arithmetic of Galois field Linear Block Code Cyclic Code Code Construction of BCH Codes Decoding of binary BCH codes ---BMA. The arithmetic of Galois field. Self-sealed arithmetic :

liuz
Download Presentation

The Theory and Technology of Error Control Coding

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The Theory and Technology of Error Control Coding Chapter 2 Block code ----BCH

  2. Summary • The arithmetic of Galois field • Linear Block Code • Cyclic Code • Code Construction of BCH Codes • Decoding of binary BCH codes ---BMA

  3. The arithmetic of Galois field • Self-sealed arithmetic: If the elements of a set are still in the set by some kind of operation (e.g. addition ,subtraction ,multiplication and division) ,then the set is called arithmetic seal up. • Field:the collection of arithmetic seal up elements is called field. • Finite field:a field with finite number of elements is called a finite field or Galois field ,written as GF(2).

  4. The arithmetic of Galois field • Binary Field A finite field which contains only the elements of 0 and 1 is called a binary field. • The arithmetic of addition and multiplication additionmultiplication 0+0=00*0=0 0+1=10*1=0 1+0=11*0=0 1+1=01*1=1

  5. The arithmetic of Galois field • Irreducible polynomial :It can not be decomposed as a product of lower degree polynomial • Example : x2 +x + 1 and x4 +x +1 are irreducible polynomials,while x2+1 is not. • The Reason :(x+1)2= x2 +x+x + 1= x2 +1

  6. The arithmetic of Galois field • GF(2m) There is a primitive element a ∈ GF(2^m), such that every nonzero element β in GF(2^m) can be expressed as , 0 ≤i ≤ 2^m − 2. • This element a is the root of an irreducible polynomial, called a primitive polynomial, p(x) over {0, 1}, that is, p(a) = 0. • A primitive element a of the field GF(2^m) satisfies the equation a^(2^m−1) = 1, and n = 2^m − 1 is the smallest positive integer such that a^n = 1.

  7. The arithmetic of Galois field • {0,1,a,a2,a3,·····,a -2 } make up the elements of GF(2m)。 • EXAMPLE: Let ‘a’ be a primitive element of GF(2^3) such that p(a) = a^3 + a + 1 = 0 and a^7 = 1. Show the ways to express the elements in GF(2^3).

  8. The arithmetic of Galois field • Minimal polynomial the minimal polynomial fi(x) of ai is the polynomial with the smallest degree which takes ai as its root. fi(x) has binary coefficients, and is irreducible. • ConjugateElements besides , 、 、····· are also the elements of fi(x),and they are called the conjugate elements of ai.

  9. Introductionof Linear Block Code • Definition The block code whose length is n and the number of the codewords is 2k , if and only if the 2k codewords constitute a k-dimensional vector subspace of all n-dimensional vector space  of the field GF (2) ,the code is called a (n, k) linear code. • Necessary and sufficient condition:adding any two codewords using mold 2 addition, the result is still the codeword of the block code.

  10. Introduction of linear block code • Encoding: • (n,k) linear block code C is a k-dimensional subspace of Vn,we can find k codewords which are linearly independent of each other,g0,g1,····,gk-1,and make any codeword v in C be a linear combination of these k codewords. • v=u0g0+u1g1+····+uk-1gk-1

  11. Introduction of linear block code • Taking the k-linearly independent code words as row vectors ,we can get a k*n matrix:

  12. Introduction of linear block code • IF u=(u0,u1,·····uk-1)is the message sequence which is to be encoded, then the codeword V is as follows described: v=u·G • The linear block word C is generated by the rows of the matrix G ,so the matrix G is called the generator matrix of C.

  13. Introduction of linear block code • For the linear block , the code word is divided into two parts , the message part and the redundant checking part. the message part is made up of the original message without changing, and the redundant checking part is made up of n-k parity codes. Linear block code with this construction is called systematic linear block code . Redundant checking part Message part n-k K

  14. Introduction of linear block code • Generator matrix G of systematic form can be written as follows G=[P,Ik] • The matrix G of systematic form can be got from G of non-systematic form via row/columm transformation.

  15. Introduction of linear block code • Parity-check matrix • For any k*n matrix G which was made up of k linearly independent row vectors, there exist a (n-k)*n matrix H which was made up of n-k linearly dependent row vector ,and the row vectors of H are orthogonal with any row vector of G. H is called the parity-check matrix.Thus, G·HT=0 ,that is v·HT=0

  16. Introduction of linear block code • If G is the systematic form ,then: • H=[In-k,PT]

  17. Introduction of linear block code • Decoding • Let v=(v0,v1,····,vn-1)is a codeword that was transmitted over a noisy channel, r=(r0,r1,····,rn-1)is the received code V and R may be different, then e=v+r=(e0,e1,····,en-1) e is called error pattern.

  18. Introduction of linear block code • Syndrome • The value of syndrome is defined as follows:

  19. Introduction of linear block code • Since • While v·HT=0 so we can get s=e·HT • That is • The n - k linear equations above do not have an unique solution but have 2^k solutions . To minimize the probability of a decoding error, the most probable error pattern that satisfies the equations is chosen as the true error vector. • If the channel is a BSC, the most probable error pattern is the one that has the smallest number of nonzero digits.

  20. Introduction of Cyclic Code • Definition • A linear code that satisfies that any cyclic shift of a codeword is still a codeword of the linear code is called Cyclic code • Advantages: a . simpler encoded mechanism because the process can be implemented by Shift registers b . simpler decoded mechanism because of fixed algebra structure

  21. Introduction of Cyclic Code • Generator polynomial. • Parity-check polynomial From the equality :g(x)·h(x)=xn+1 , we can get the parity-check polynomial h(x)

  22. Introduction of Cyclic Code • Encoding of cyclic code Non-systematic encoding: v(x)=u(x)g(x) Systematic encoding: v(x)=xn-ku(x)+[xn-ku(x)modg(x)]

  23. Introduction of Cyclic Code • General decoding of cyclic codes Let r(x) = v(x) + e(x) e(x) the error polynomial v(x) the code polynomial r(x) the received code polynomial • Then define the syndrome polynomial as

  24. BCH code • BCH code which was named after the inventors (Bose- Chaudhurl-Hocquenghem), is one of cyclic code that can correct more than one errors since 1959. As the code generator polynomial is related to the minimum distance, the code construction can be easily determined by the error correcting ability. Therefore, BCH is one of the widely used error control codes .

  25. BCH code • Definition: A BCH code of dmin ≥ 2td + 1 is a cyclic code whose generator polynomial g(x) has 2td consecutive roots: So, a binary BCH (n, k, dmin) code has a generator polynomial: g(x) = LCM{φb(x), φb+1(x), . . . , φb+2td−1(x)}, Length n = LCM{nb, nb+1, . . . , nb+2td−1}, and dimension k = n − deg[g(x)]

  26. BCH code Code Construction of BCH Codes : t-error correcting binary BCH code of length 2m 1 • Find a degree m primitive polynomial p(x) over GF(2). • Use p(x) to construct GF(2^m). • Let  be a primitive element of GF(2^m), i.e.  is a root of p(x) in GF(2^m). • Find the Φ(x) ,the minimal polynomial of ^i, for i = 1, 3, .., 2t  1. • Let g(x) be the minimal degree polynomial with the Φ(x) as factors,,

  27. BCH code Simplification of generator matrix: • According to the arithmetic of finite field , • If • then , while i’ is odd. • i and i’ are conjugate elements ,so they have the same minimal polynomial, then the generator matrix can be simplified as follows:

  28. BCH code • Example consider GF(24), p(x) = x4 + x + 1,with td = 2 and b =1. then g(x) = LCM{φ1(x), φ3(x)} = (x4 + x + 1)(x4 + x3 + x2 + x + 1) = x8 + x7 + x6 + x4 + 1 generates a double-error-correcting binary BCH (15,7,5) code.

  29. BCH code • Encoding of BCH: So we have got the generator polynomial g(x), the code word v(x) can be consturcted by v(x) = b(x) + X^(n-k)*m(x). Just the same as encoding the cyclic code mentioned before.

  30. Decoding algorithm of BCH • BMA(Berlekamp-Massey Agorithm) The BMA was invented by Berlekamp (1984), Massey (1969). This is a computationally efficient method to solve the key equation, in terms of the number of operations in GF(2^m). • EA(Euclidean Agorithm) • Direct solution

  31. Decoding algorithm of BCH • Let v(X)=v0+v1X+····+vn-1Xn-1 is the code polynomial of a BCH code whose length n is 2m-1 and the error correcting ability is t , FOR 1<=i<=2t, i is the roots of v(X),so we can get : v(i)= v0+v1 i +····+vn-1 i(n-1)=0 From the above property ,we can construct H as follows: H= 1 23··································n-1 1 2 (2)2 (2)3····················(2)n-1 1 2t (2t)2 (2t)3··············(2t)n-1

  32. Decoding algorithm of BCH • According to the concept of conjugate element, H can be simplified as follows: • H= 1 23··································n-1 1 3 (3)2 (3)3····················(3)n-1 1 2t-1 (2t-1)2 (2t-1)3·····(2t-1)n-1

  33. Decoding algorithm of BCH • We assume that the transmission errors occurred,v(X)=v0+v1X+····+vn-1Xn-1 changed into r(X)=r0+r1X+····+rn-1Xn-1 In the receiver side, r(X)=v(X)+e(x) s=(s1,s2,·····,s2t)=r·HT so we can get the ith component of S: si= r(i)=r0+r1 i +····+rn-1 i(n-1) while r(X)=ai(x)φi(X)+bi(X) φi(i)=0 si= r(i)= bi(i)=v(i)+e(i)= e(i) let

  34. Decoding algorithm of BCH • From the above analysis ,we can get a series of equations  about the syndrome: Any method for solving these equations is a kind of decoding algorithm for BCH define the error position polynomial as follows :

  35. Decoding algorithm of BCH • The error correcting process of BCH code: 1、calculate syndrome S from the receiver polynomial r(X): s=(s1,s2,·····,s2t) 2、get the error polynomial by s1,s2,·····,s2t . 3、resolve the roots of ,find the position of errors ,and correct the errors of r(X).

  36. BMA • The goal of the BMA is to find a (connection) polynomial σ^ (i+1)(x) of minimal degree that satisfies the following equations, derived from • (a) • The discrepancyat iteration i defined as • If di = 0, then the equations (a) are satisfied for If di! = 0: Let σ ^(m)(x) be the solution at iteration m, such that −1 ≤ m < i, dm = 0,and (m − lm) is maximal. Then

  37. BMA With an initial value of i = 0, the computation of σ (i+1)(x) continues until the conditions i ≥ i+1 + td − 1 or i = 2td − 1 or both are satisfied. The initial conditions of the algorithm are:

  38. BMA • Example 3.5.1 Let C be the triple-error-correcting BCH (15,5,7) code of Example 3.3.3. As a reference, to check the numerical computations, the power and vector representations of GF(2^4), with primitive polynomial p(x) = 1 + x + x4, are listed. A generator polynomial for C is

  39. BMA Suppose that the information polynomial is Then the corresponding code polynomial is given by Let be the polynomial associated with a vector r = v + e received after transmission of code word v over a BSC channel.

  40. BMA The vector e corresponds to the error polynomial e(x) = 1 + x6 + x12. Obviously, the decoder does not know this. However, in computing the syndromes below, this knowledge is used to simplify expressions. The syndromes are:

  41. BMA Berlekamp–Massey algorithm:

  42. BMA

  43. BMA

  44. BMA Chien search and error correcting • To find the roots of σ(x), a simple trial-and-error procedure – called Chien search – is performed. All nonzero elements β of GF(2m) are generated in sequence 1, a, a^2, . . . And the condition = 0 tested. • For binary BCH codes, once the error locations j1, . . . , jv are known, the corresponding bits in the received word are complemented and the estimated code word v(x) generated.

  45. Reference • The Art of Error Correction Coding (SECOND EDITION)Written byRobert H.morelos-Zaragoza • Error Control Coding(SECOND EDITION) Written by Shu Lin Daniel J.Costello,Jr.

More Related