Peterson-Gorenstein-Zierler Decoder

(1) e(X) = e0 + e1X + … + en-1Xn-1 where at most t coefficients are nonzero

(2) Suppose v errors actually occur, 0 ≤ v ≤ t, in unknown locations
i1, i2, … ,iv
∴ e(X) = ei1Xi1 + ei2Xi2 + … + eivXiv ij, eij, v are unknown !
(3) S1 = v(α) = c(α) + e(α) = e(α)
= ei1αi1 + ei2αi2 + … + eivαiv
Let Yl = eil : error magnitude 1≤l≤v
Xl = αil : error-location number
Note : all Xl are different (∵αn = 1)
S1 = Y1X1 + Y2X2 + … + YvXv
Q

Q: If j0 is equal to 2, not 1, how do the peterson’s method work?

2006/12/13 Yuh-Ming Huang BCH_Code_1 1
(4) For α α2 α3 … α2t
S1 = Y1X1 + Y2X2 + … + YvXv Yl  GF(q)
(＊)
nonlinear S2 = Y1X1
2 + Y X 2 + … + Y X 2
2 2 v v Xl  GF(qm)
equations S3 = Y1X13 + Y2X23 + … + YvXv3
.
.
,
S2t = Y1X12t + Y2X22t + … + YvXv2t
1.(＊)至少有一解 ∵Syndrome Si is defined in such way
2.證明此解unique

2006/12/13 Yuh-Ming Huang BCH_Code_1 2

Note: j0 =1
S1 = v(β) = c(β) + e(β) = e(β)
= ei1βi1 + ei2βi2 + … + eivβiv

Let Yl = eil : error magnitude 1≤l≤v

Xl = βil : error-location number
Note : all Xl are different (∵αn = 1)
S1 = Y1X1 + Y2X2 + … + YvXv v=1, j0 =2 ?
S1 = Y1X12
= Y 1 β12
Note: If β= 3 and v=1 & i1=6 = Y1 36
Then S1 = Y1X1j0 = Y1X1 = Y1 β6 = Y1 18
2006/12/13 Yuh-Ming Huang BCH_Code_1 3
Consider the polynomial Λ(x) ≡ (1-X1x)(1-X2x)…(1-Xvx)
= 1 + Λ1x + Λ2x2 + … + Λvxv – (1)
error-locator polynomial
If we knew the coefficients Λl of Λ(x)
we could find the zeros of Λ(x) to obtain the error locations Xl
S i  Λl
1≤i≤2t 1≤l≤v
Multiply both sides of (1) by YlXlj+v and set x = Xl-1

2006/12/13 Yuh-Ming Huang BCH_Code_1 4

 0  Yl X l j  v 1   1 X l1   2 X l2  ...   v X l v 
 Yl X l j v   1 X l j v 1  ...   v X l j   0

  Yl X l j  v   1 X l j  v 1  ...   v X l j   0
v

l 1
v v v

 Yl X l
l 1
j v
  1  Yl X l
l 1
j  v 1
 ...   v  Yl X l j  0
l 1

S j  v   1 S j  v 1   2 S j  v 2  ...   v S j  0
 1 S j v 1   2 S j v 2  ...   v S j   S j  v 1 j  v

 S1 S2 S3 . . . S v 1 S v    v    S v 1  ∵v ≤ t
S2 S3 S4 . . . Sv S v 1    v 1   S v 2  ∴S1 ~ S2v
S S4 S5 . . . S v 1 S v  2   v 2    S v 3  all known
 3   
.  .   . 
.  .   . 
.  .   . 
 S v S v 1 S v 2 . . . S 2 v 2 S 2 v 1    1    S 2 v 
2006/12/13 Yuh-Ming Huang BCH_Code_1 5
 Thm : The Vandermonde matrix has a nonzero determinant
iff all of Xi 1 ≤ i ≤ u are distinct

 1 1 . . . 1 
 X1 X2 . . . Xu 
 X2 X 22 . . . X u2 
 .1 . 
A .
 . . . 
 . . . 
 X 1u 1 X 2u 1 . . . X uu 1 

Thm : The matrix of syndromes M is nonsingular if u is equal to v, the number

of errors that actually occurred. The matrix is singular if u > v.

 S1 S2 . . . Su 
S2 S3 . . . S u 1  If v > t
. . .  i.e. u < v時
M   det(M) = ?
. . . 
. . . 
 S u S u ! . . . S 2 u 1 
2006/12/13 Yuh-Ming Huang BCH_Code_1 6
Pf:Let Xu  0 for u  v
 1 1  1 
 X X2  Xu 
Let A   1  with A  Xi-1
     ij j

 u 1 u 1 
 X 1 X 2u 1  X u 
Y1 X 1 0  0 
 0 0  1 i j
 with Bij  Yi X i ij ,  ij  
Y2 X 2 
B  
    0 i  j
 
 0 0  Yu X u 
u u u u

 ABA T
ij
  
l 1
X li-1
k 1
Yl X l lk X kj 1  
l 1
X li-1Yl X l X lj 1  
l 1
Yl X li  j 1  M ij
T
 M  ABA
(1) u  v det(B)  0  det(M)  det(A)det(B)det(A T )  0
(2) u  v det(B)  0 & det(A)  0  det(M)  0

2006/12/13 Yuh-Ming Huang BCH_Code_1 7

 Enter υ(x)
special case
Compute syndromes
j0 = 1
Sj = υ(αj+j0-1) j = 1 … 2t α α2 … α2t

v=t

 S1  Sv 
M     Find error location
S  S 2v 1  Xl (l = 1…..v)
 v
by finding zeros of Λ(x)

Yes
det(M) = 0 vv-1 1
 Y1   X 1 X v   S1 
    
 Yv   X v v
X v  S v 
   1
No

 Λv    S v 1 
 Λ v 1  1   S v  2  Halt
    M  
 Λ   
 1    S 2v 

Figure 7.1 The Peterson-Gorenstein-Zierler decoder.

2006/12/13 Yuh-Ming Huang BCH_Code_1 8
Eg:
BCH(15,5) triple-error-correcting code.
g(X) = X10 + X8 + X5 + X4 + X2 + X + 1
received polynomial
 V(X) = X7 + X2

In GF(24) S1 = α7 + α2 = α12
S2 = α14 + α4 = α9
S3 = α21 + α6 = 0
S4 = α28 + α8 = α3
S5 = α35 + α10 = α0
S6 = α42 + α12 = 0

2006/12/13 Yuh-Ming Huang BCH_Code_1 9

 S3   9 0
12
 S1 S2
 
set v  3 M   S 2 S3 S4     9 0  3  det  M   0
 
 S3 S4 S5   0 3

1 

 S1 S2   12  9 
v2 M    det  M   0
 S2 S3    9 0 

-1
 0  6

M  
  
6 9

 2  -1  0    9

   M   3    
 1     
12

  X   1   12 x   9 x 2
 
 1 2x 1 7x 
 
  9 x   8 x   13 
1
 7
8
1
2
 13
e  X   X 7  X 2
2006/12/13 Yuh-Ming Huang BCH_Code_1 10
Reed-Solomon Code (maximum-distance code)
(1)The symbol field GF(q) and the error-locator field GF(qm) are the same
(Yl) (Xl)
(i.e. m = 1)
(2)Take α as primitive n = qm – 1 = q – 1
(3)The minimal polynomial over GF(q) of an element β in the same GF(q)
is fβ(X) = X – β
(4)Take j0 = 1 g(X) = (X – α)(X – α2)… (X – α2t) ∴ n – k = 2t

eg1. (15,11) t = 2 GF(16) j0 = 1

g(X) = (X – α)(X – α2) (X – α3)(X – α4)
= X4 + (Z3 + Z2 + 1)X3 + (Z3 + Z2)X2 + Z3X + (Z2 + Z + 1)
= X4 + α13X3 + α6X2 + α3X + α10
n – k = 4  k = 11 (11 16-ary symbols  44 bits)

2006/12/13 Yuh-Ming Huang BCH_Code_1 11

eg2. (7,3) t = 2 GF(8) j0 = 4
g(X) = (X – α4)(X – α5) (X – α6)(X – α0)
= X4 + (Z2 + 1)X3 + (Z2 + 1)X2 + (Z + 1)X + Z
n – k = 4  k = 3 (3 8-ary symbols  9 bits)
If i(X) = (Z2 + Z)X2 + X + (Z + 1)
c(X) = i(X)g(X) = (α4X2 + X + α3)(X4 + α6X3 + α6X2 + α3X + α)
= α4X6 + αX5 + α6X4 + 0X3 + 0X2 + α5X + α4

※ 1. n – k + 1 = 2t + 1 ≤ d* ≤ 1 + n – k
∴ d* = 1 + n – k
2. R-S codes always have relatively short block-length as compared
to other cyclic codes over the same alphabet !

2006/12/13 Yuh-Ming Huang BCH_Code_1 12

2006/12/13 Yuh-Ming Huang BCH_Code_1 13
2006/12/13 Yuh-Ming Huang BCH_Code_1 14
2006/12/13 Yuh-Ming Huang BCH_Code_1 15
2006/12/13 Yuh-Ming Huang BCH_Code_1 16
2006/12/13 Yuh-Ming Huang BCH_Code_1 17
2006/12/13 Yuh-Ming Huang BCH_Code_1 18