Cyclic Codes BCH Codes

Similar documents
Chapter 6. BCH Codes

Associative Memories

Lecture 5 Decoding Binary BCH Codes

we have E Y x t ( ( xl)) 1 ( xl), e a in I( Λ ) are as follows:

SUMMARY OF STOICHIOMETRIC RELATIONS AND MEASURE OF REACTIONS' PROGRESS AND COMPOSITION FOR MULTIPLE REACTIONS

Chapter Newton s Method

Chapter 6. Rotations and Tensors

LECTURE 21 Mohr s Method for Calculation of General Displacements. 1 The Reciprocal Theorem

Decoding of the Triple-Error-Correcting Binary Quadratic Residue Codes

Analysis of Non-binary Hybrid LDPC Codes

Note 2. Ling fong Li. 1 Klein Gordon Equation Probablity interpretation Solutions to Klein-Gordon Equation... 2

Solution of Linear System of Equations and Matrix Inversion Gauss Seidel Iteration Method

Neumann Asymptotic Eigenvalues of Sturm-liouville Problem with Three Turning Points

ISSN: ISO 9001:2008 Certified International Journal of Engineering and Innovative Technology (IJEIT) Volume 3, Issue 1, July 2013

THE METRIC DIMENSION OF AMALGAMATION OF CYCLES

Analysis of Bipartite Graph Codes on the Binary Erasure Channel

Neural network-based athletics performance prediction optimization model applied research

and decompose in cycles of length two

Foundations of Arithmetic

= z 20 z n. (k 20) + 4 z k = 4

between standard Gibbs free energies of formation for products and reactants, ΔG! R = ν i ΔG f,i, we

High-Speed Decoding of the Binary Golay Code

ON AUTOMATIC CONTINUITY OF DERIVATIONS FOR BANACH ALGEBRAS WITH INVOLUTION

The stress functions of the Cosserat continuum

Inthem-machine flow shop problem, a set of jobs, each

Nice plotting of proteins II

DIOPHANTINE EQUATIONS WITH BINOMIAL COEFFICIENTS AND PERTURBATIONS OF SYMMETRIC BOOLEAN FUNCTIONS

Quantum Runge-Lenz Vector and the Hydrogen Atom, the hidden SO(4) symmetry

Formulas for the Determinant

Monica Purcaru and Nicoleta Aldea. Abstract

Structure and Drive Paul A. Jensen Copyright July 20, 2003

( ) r! t. Equation (1.1) is the result of the following two definitions. First, the bracket is by definition a scalar product.

CSE4210 Architecture and Hardware for DSP

APPENDIX A Some Linear Algebra

Example: (13320, 22140) =? Solution #1: The divisors of are 1, 2, 3, 4, 5, 6, 9, 10, 12, 15, 18, 20, 27, 30, 36, 41,

QUARTERLY OF APPLIED MATHEMATICS

Lower bounds for the Crossing Number of the Cartesian Product of a Vertex-transitive Graph with a Cycle

G = G 1 + G 2 + G 3 G 2 +G 3 G1 G2 G3. Network (a) Network (b) Network (c) Network (d)

International Mathematical Olympiad. Preliminary Selection Contest 2012 Hong Kong. Outline of Solutions

Lower Bounding Procedures for the Single Allocation Hub Location Problem

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

COXREG. Estimation (1)

On the set of natural numbers

P.H.G. l-leseab.ci-l LABORATORIES - LABORATORY 1WTE. Di vision a r.e ed arose for decade pulse counters possessing certain

SOME CHARACTERS OF THE SYMMETRIC GROUP R. E. INGRAM, SJ.

Image Classification Using EM And JE algorithms

On the Interval Zoro Symmetric Single-step Procedure for Simultaneous Finding of Polynomial Zeros

The Geometry of Logit and Probit

Boundary Value Problems. Lecture Objectives. Ch. 27

LECTURE 9 CANONICAL CORRELATION ANALYSIS

Independent Device Currents

Bernoulli Numbers and Polynomials

n-step cycle inequalities: facets for continuous n-mixing set and strong cuts for multi-module capacitated lot-sizing problem

Chapter 8. Potential Energy and Conservation of Energy

Entropy Coding. A complete entropy codec, which is an encoder/decoder. pair, consists of the process of encoding or

Module 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur

Note On Some Identities of New Combinatorial Integers

arxiv: v1 [cs.cr] 18 Feb 2016

A Comparison between Weight Spectrum of Different Convolutional Code Types

LOW-DENSITY Parity-Check (LDPC) codes have received

Principles of Food and Bioprocess Engineering (FS 231) Solutions to Example Problems on Heat Transfer

First day August 1, Problems and Solutions

Section 3.6 Complex Zeros

Coincidences of Hypercubic Lattices in 4 dimensions

Quantum Mechanics I - Session 4

VQ widely used in coding speech, image, and video

Strain Energy in Linear Elastic Solids

The line method combined with spectral chebyshev for space-time fractional diffusion equation

L-Edge Chromatic Number Of A Graph

Chapter 2. Pythagorean Theorem. Right Hand Rule. Position. Distance Formula

Development of whole CORe Thermal Hydraulic analysis code CORTH Pan JunJie, Tang QiFen, Chai XiaoMing, Lu Wei, Liu Dong

Report on Image warping

Motion in One Dimension

The optimal delay of the second test is therefore approximately 210 hours earlier than =2.

Circular units of an abelian field ramified at three primes

Min Cut, Fast Cut, Polynomial Identities

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

Transfer Functions. Convenient representation of a linear, dynamic model. A transfer function (TF) relates one input and one output: ( ) system

MARKOV CHAIN AND HIDDEN MARKOV MODEL

Analysis and construction of a family of refinable functions based on generalized Bernstein polynomials

COMPLEX NUMBERS AND QUADRATIC EQUATIONS

Module 2. Random Processes. Version 2 ECE IIT, Kharagpur

Physics 5153 Classical Mechanics. Principle of Virtual Work-1

Selected Student Solutions for Chapter 2

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP

Introduction to Vapor/Liquid Equilibrium, part 2. Raoult s Law:

CHALMERS, GÖTEBORGS UNIVERSITET. SOLUTIONS to RE-EXAM for ARTIFICIAL NEURAL NETWORKS. COURSE CODES: FFR 135, FIM 720 GU, PhD

EEL 6266 Power System Operation and Control. Chapter 3 Economic Dispatch Using Dynamic Programming

PHASED TILINGS AND GENERALIZED FIBONACCI IDENTITIES

The BCH Bound. Background. Parity Check Matrix for BCH Code. Minimum Distance of Cyclic Codes

A summation on Bernoulli numbers

Some Properties of a Recursive Procedure for High Dimensional Parameter Estimation in Linear Model with Regularization

Remark: Positive work is done on an object when the point of application of the force moves in the direction of the force.

Mathematical Preparations

1 Derivation of Point-to-Plane Minimization

From the Euclidean Algorithm for Solving a Key Equation for Dual Reed Solomon Codes to the Berlekamp-Massey Algorithm

3. Stress-strain relationships of a composite layer

Randić Energy and Randić Estrada Index of a Graph

THEOREMS OF QUANTUM MECHANICS

Approximate merging of a pair of BeÂzier curves

Transcription:

Cycc Codes BCH Codes Gaos Feds GF m A Gaos fed of m eements can be obtaned usng the symbos 0,, á, and the eements beng 0,, á, á, á 3 m,... so that fed F* s cosed under mutpcaton wth m eements. The operator + s defned by ddng by p where p s a prmte rreducbe poynoma n GF m. = q. p + a where a s a poynoma of degree m- or ess oer GF. The outcome s a set of m non zero poynomas of á oer GF m wth degree m- or ess. Exampe.. tartng wth m = 4, p = 4 + 3 +, whch s a prmte poynoma oer GF and a factor of 5 +, et pá = á 4 + á 3 + = 0. Hence á 4 = + á 3 and the GF 4 can be constructed and s gen by Tabe 4. Eements of GF 4 usng p = 4 + 3 + oer GF 4 Power Eements Poynoma 4-Tupe 0 0 0 0 0 0 0 0 0 á á 0 0 0 á á 0 0 0 á 3 á 3 0 0 0 á 4 + á 3 0 0 á 5 + á + á 3 0 á 6 + á + á + á 3 á 7 + á + á 0 á 8 á + á + á 3 0 á 9 + á 0 0 á 0 á + á 3 0 0 á + á + á 3 0 á + á 0 0 á 3 á + á 0 0 á 4 á + á 3 0 0 Tabe 4. 4 + + s rreducbe oer GF and does not hae roots oer GF, but t has 4 roots oer GF 4. These are gen by á 7, á, á 3 and á 4. It can be shown, usng Tabe 4., that + á 7 + á + á 3 + á 4 = + + 4.

Further Feds A new fed eement â s ntroduced n an extenson fed of GF wth â a root of a poynoma f so that fâ = 0. L L L For any 0, s aso a root of f so that f = 0. The eement s the conugate of â. Ths aso mpes that f â, an eement n GF m s a root of f oer GF, then a the dstnct conugates of â, aso eements of GF m are roots of f. Exampe Usng f = 4 + + + + = 6 + 5 + 4 + 3 +, á 7 s a root. The conugates of á 7 are á 7, 7 3, 7 4. Note that 7 s á = á / á 05 = á 7 and hence coses the group. The conugates of á 7 are á 4, á 8 /á 5 = á 3, and á 56 /á 45 = á. The other prmte roots are á 5 and á 0. Further snce â s an eement of GF m m m, n genera,, and 0, and the m nonzero eements of GF m m form a the prmte roots of 0. Aso snce the zero eement 0 of GF m s the root of, then the m eements form a the roots of m Mnma poynomas. The fed eement â can aso be a root of a poynoma of degree ess than m. The poynoma of smaest degree oer GF for whch f = fâ = 0 s known as the mnma poynoma of â, and denoted by Ö. Ths poynoma s aso rreducbe. Further f f, a poynoma oer GF has â as a root, then f n genera s dsbe by Ö, the mnma poynoma. If f tsef s an rreducbe poynoma then f = Ö. It foows that the conugates of â, â, roots of Ö = Öâ. It can be shown that f L 0, 3,.. are aso Exampe: For GF 4 and the fed eements of Tabe 4., startng from â = á 3, we obtan â = á 6, â 4 = á, â 8 = á 4. These resut n the poynoma + á 3 + + á + á + á 3 + + á + + á resutng n ++ á+ á + + á + á+ á + + á + á + á 3 fnay resutng n the mnma poynoma 4 + 3 + + +. Another way to obtan the mnma poynoma s the foowng. Let ã = á n GF 4 be used as the prmte eement. Hence ã = á, ã 4 = á 4, ã 8 = á 8, and ã 6 = á 6 = ã coses the group. Hence Ö of degree 4 must hae the foowng form. Ö = a 0 + a + a + a 3 3 + a 4 4. Usng the poynoma representaton and substtutng for ã = á, Ö = a 0 + a á+ a á + a 3 á 3 + a 4 á 4.

Ths resuts n a 0 + a á+ a á + a 3 á 3 + a 4 á 3 + = 0. Ths s rearranged to get a 0 + a 4 + a á+ a á + a 3 + a 4 á 3 = 0 Hence a 0 + a 4 = 0 a = 0 a = 0 a 3 + a 4 = 0 Ths resuts n a 3 = a 4 = a 0 = ; a = a = 0; and therefore the poynoma + 3 + 4 = Ö. The Tabe 4. shows the mnma poynomas wth the prmte eements as the prmte roots of the mnma poynomas usng p = + 3 + 4 Conugate roots Mnma poynoma 0 + Ö 0 á, á, á 4, á 8, + 3 + 4 Ö á 3, á 6, á 9, á, 4 + 3 + + + Ö 3 á 5, á 0, ++ Ö 5 á 7, á, á 3, á 4, + + 4 Ö 7 Tabe 4. Tabe 4. shows that the degree of each mnma poynoma n GF 4 usng + 3 + 4 as the prmte poynoma g. Note that budng up other generator poynomas g from g, st uses g so that g w aways ncude the prmte root á. BCH Code It s charactersed by the foowng: Bock ength n = m ; Party checks n-k mt; Mnmum dstance d mn t+; The generator poynoma g s specfed n terms of ts roots n GF m. Eery prmte eement á s a root of a mnma poynoma Ö. It can be shown that a een powers of á, beong to a mnma poynoma wth a precedng odd power as one of ts roots. Ths s ustrated by Tabe 4. aboe. BCH Bound: The mnmum dstance of the code generated by g s greater than the argest number of consecute prmte roots of g. Usng a generator poynoma g = Ö 0. Ö. Ö 7 yeds the set of prmte roots whose ndex s 0,., 4,,, 7, 8,,,,, 3, 4. Note that there are 5 consecute prmte roots n the sequence so that g has a mnmum dstance of at east 6. Lookng at Tabe 4., t can be seen that eery odd root s n the same poynoma as. Hence t consecute odd roots guarantee t consecute roots. Aso t can be shown 3

m that the degree of eery dsor of, cannot exceed m. nce at most t mnmum poynomas are requred to guarantee that g has t consecute odd roots, the order of g s m.t and, at most, m.t party checks are requred. Encodng a BCH codeword. The encodng process s dentca to the standard cycc code. For a k-bt data d the resutant party bts are found from rem { n-k..d}/ g whch are appended to the front of the d to obtan the codeword. Eery codeword n a BCH code s a codeword f t s dsbe by the GF m roots, á, á,...á t. Decodng a BCH codeword. Assume a codeword sent, and r s receed. Then the error pattern can be dered from r = + e. The syndrome of a t-correctng BCH code s gen by =,,.. t, and = rá Dde r, n turn, by each of the mnma poynomas comprsng g. In each case a remander term b s obtaned. Ths remander s n GF. Ths s substtuted by the correspondng prmte root beongng to the mnma poynoma. Exampe: Usng g = + 3 + 4 n GF 4 the 5,7 code uses as prmte poynomas, Ö = + 3 + 4, and Ö 3 = 4 + 3 + + +. Ths ges g = + + + 4 + 8 for a 5,7 code. Usng a data pattern [0000] that ges d = + 3 + 6, a code word s but gen by = + 5 + 8 + + 4. Let r be + 8 + + 4. Ths resuts n an e = + + 5. To determne the syndrome =,, 3, 4 the r s dded by each of the mnma poynomas. Usng Ö = + 3 + 4, the remander s b = + + 3. Usng the roots of the mnma poynoma, á, á, á 4, Hence = + á + á 3 = á = + á 4 + á 6 = + á + á = á 7 4 = + á 8 + á = á + á 3 = á 4 3 s obtaned from Ö 3 = 4 + 3 + + +. The remander s b 3 = + +. Usng the frst root of ths mnma poynoma, á 3, 3 = + á 3 + á 6 = á + á = á 3. Hence = á, á 7, á 4, á 3 4

5 The second step, after determnng the syndrome n terms of the prmte eements s to determne the error ocaton poynoma ó from the syndrome components. There are arous methods aaabe. They are based on a genera souton nong the foowng. Gen the errors, t, the error postons are denoted by..,, nce the syndromes = eá, eery syndrome s reated drecty to the error parameters. Ths ges rse to a set of equatons t t t t Defne the error ocator poynoma as The prmte eement roots of ths poynoma are the nerse error ocaton postons. It s easy to show from the aboe the set of Newton Identtes gen by + ó = 0 + ó + ó = 0 3 + ó + ó + 3ó 3 = 0.................... + ó - +....+ ó - + ó = 0 Note that snce n GF + = = 0, ó = ó for odd, and 0 for een. The Berekamp-Massey Agorthm w be used for the souton of the Newton Identtes. The goa of the agorthm s to fnd at teraton + connecton poynoma ó + n terms of the error poynoma prmte eements, and gen by usng as the error dscrepancy that becomes a correcton factor the aue, d, usng d where the upper ndces assocated wth ó ndcate the coeffcent aue assocated wth an approprate n the equaton óat the th teraton. If d = 0, then there s no dscrepancy at that stage, and the present aue of ó, ó, s carred to the next teraton ó +. If d 0, fnd a preous teraton row, ñ, for whch d 0, and the aue of ñ- ñ where ñ denotes the order of ó ñ. Then work out the aue of the next teraton ó + usng, max d d 4.3 The teratons are contnued unt the quantty, +t becomes ad

Exampe: The BCH 5,5 code, whch has t=3, s generated usng Ö = + 3 + 4 ; Ö 3 = 4 + 3 + + +; Ö 5 = ++. Ths resuts n a g = + + 5 + 6 + 8 + 9 + 0. A code poynoma s but usng the data pattern [00] whch s d = + + 4. The codeword s but by usng 0 d/ g to obtan the remander. In ths case the remander s gen by + + 6 + 8 so that the codeword = + + 6 + 8 + + + 4. The receed word s r = + 8 + + 4. Ths mpes an error poynoma e = + 6 +. Of course the decoder does not know ths. The procedure for decodng starts wth the syndrome cacuaton, obtaned by ddng the receed word r by each mnma poynoma n turn to work out the correspondng prmte eement assocated wth the syndrome eement. In ths case = [,, 3, 4, 5, 6 ]. nce á, á, á 4, are obtaned from the same poynoma Ö = + 3 + 4, r s dded by Ö, to obtan b = + + 3 Hence = + á + á 3, and usng the GF 4 arthmetc, based on + 3 = 4, and Tabe 4. = á. Usng á, = + á 4 + á 6 = + á + á = á 7. Usng á 4, 4 = + á 8 + á = á + á 3 = á 4. á 3, á 6, are obtaned from Ö 3 = 4 + 3 + + +, to obtan b 3 = ++, and usng the prmte eements, á 3 3 = + á 3 + á 6 = á + á = á 3 and usng á 6 6 = + á 6 + á = + á + á 3 = á. Fnay á 5 s obtaned from Ö 5 = ++, to obtan b 5 =. Therefore 5 =. The Berkeamp-Massey Agorthm s now used. Intasaton Iteraton 0: ó - = ; d - =0; - =0; - =0-0=0; snce d - =0; ó 0 = ; Iteraton: =0; d 0 = = á and usng 4.3 ó = ó 0 + á. Therefore at end of teraton the entry s + á. 0 0 Check on d : d = + ó = á 7 + á.á = á 7 + á = á 7 + á 7 = 0 Iteraton : =; d = 0; Hence ó = ó ; =; - = ; d = 3 + ó = á 3 + á 9.á = á 3 + á 8 = á + á + á 3 = á 8. Entry + á. á 8 Iteraton 3: =; d = á 8 ; Hence update ó, usng row teraton 0, to obtan 6

ó 3 = ó + d.d 0 -. -0. ó 0 = ó + á 8. / á.. = ó + á.. Hence ó 3 = + á + á.. Entry on Iteraton 3 s 3 + á + á.. 0 Check on d 3 : d 3 = 4 + 3 ó 3 + ó 3 = á 4 + á 3.á + á 7.á = á 4 + á 9 + á 4 = 0 Iteraton 4: : =3; d 3 = 0; Hence ó 4 = ó 3 ; 4 = 3 =; - = ; Check : 4 + 3 = 4. therefore 3 s not ad. Contnue. Current entry 4 + á + á.. d 4 d 4 = 5 + 4 ó 4 + 3 ó 4 = + á 4.á + á 3.á = + á 5 + á 5 = Iteraton 5: =4; d 4 = ; Hence update ó 4 to ó 5, usng row teraton, to obtan ó 5 = ó 4 + d 4.d -. 4-. ó =. / á 8. + á = ó 4 + á 7 + á 8 3 = + á + á + á 7 + á 8 3 = + á + á + á 3 3. Entry for teraton 5 5 + á + á + á 3 3 0 3 Check on d 5 : d 5 = 6 + 5 ó 5 + 4 ó 5 + 3 ó 3 5 = á +.á + á 4.á + á 3.á 3 = 0 Iteraton 6: =5; d 5 = 0; Hence ó 6 = ó 5 ; 6 = 5 =3; - = 3; Check : 6 + 3 = 5. therefore 5 s true. Iteraton stopped. The outcome of the agorthm s ó 6 = + á + á + á 3 3. The roots of ths cubc poynoma are found to be n ths case by a process of tra and error on the ffteen prmte eements = ; = á 3 ; = á 9 ; eg for =; + á + á + á 3 = + á + á 3 + á + á 3 = 0 The error ocatons n e are the nerse of these roots. o error ocatons are at poston, 6,. Ths s the expected resut. The oera teratons are gen n the Tabe 4.3 beow 7

I ó d - - 0 0-0 á 0 0 + á 0 0 + á á 8 3 + á + á.. 0 4 + á + á.. 5 + á + á + á 3 3 0 3 6 + á + á + á 3 3 - - - Other BCH Codes Tabe 4.3 Bnary BCH codes wth ength n m can be constructed as for those wth n = m. Let â be an eement of order n n GF m. Consder a poynoma that has as roots â, â, â 3,..., â t. n tsef s a factor of some m. A the eements are roots of n +. Therefore ths s a cycc code. In partcuar, for a sequence of t roots, the g that s the LCM of the mnma poynomas of a the roots, generates a t-error correctng BCH code. nce â s not a prmte eement of GF m and n m, the BCH code generated s caed a nonprmte BCH code. Non-bnary BCH codes Reed oomon Codes Bnary BCH codes can be generazed to any GFq where p s a prme number and q any power of p to obtan a q-ary code. An n,k q-ary cycc code s generated by a poynoma of degree n-k wth coeffcents from GFq whch s a factor of n +. Let á be a prmte eement, n GFq s, where n = q s. For a t error correctng code the generator poynoma g has t roots from GFq gen by á, á,.., á t. The degree of each mnma poynoma s s or ess, and hence the number of party check dgts generated by by g s no more than st. The speca subcass for whch s= s the most mportant subcass of q-ary BCH codes. These codes are usuay caed Reed-oomon codes. A t-error correctng R code wth symbos from GFq has the foowng parameters Bock ength n = q Number of party-check dgts n-k = t Mnmum dstance d mn = t+. Usng GFq = GF m, and usng á as a prmte eement n GF m, a Reed- oomon code, t-error correctng, can be generated usng a g = +á+á +á 3. +á t so that g = g 0 + g + g + + g t- t- + t, so that the g s are now not from GF but from GF m. Generatng a codeword s st the process of ddng t d by g and usng the remander to bud up the systematc codeword. 8

Decodng foows the nes of a BCH code nong:. yndrome cacuaton. Error ocaton usng an error ocaton poynoma, and an agorthm for the souton such as the Berekamp-Massey agorthm for the souton of ó 3. Obtan from the error ocaton poynoma, the error aues, Z, n terms of á s usng Newton s denttes 4. Fnay obtan the error aues at the ocatons obtaned from the error ocaton poynoma usng an equaton reatng the error ocatons and Z. 9