Error detection and correction in digital communication systems

• US 4,597,083 A
• Filed: 04/06/1984
• Issued: 06/24/1986
• Est. Priority Date: 04/06/1984
• Status: Expired due to Term

• Current Assignee: Ampex Corporation (Delta Information Systems, Inc.)
• Original Assignee: Ampex Corporation (Delta Information Systems, Inc.)
• Inventors: Stenerson, Roger M.
• Primary Examiner(s): Fleming, Michael R.
• Application Number:
  US06/597,538
• Time in Patent Office: 809 Days
• Field of Search: 371/37, 371/38, 371/39
• US Class Current: 714/753
• CPC Class Codes: H03M 13/15 Cyclic codes, i.e. cyclic shifts of codewords produce other codewords, e.g. codes defined by a generator polynomial, Bose-Chaudhuri-Hocquenghem [BCH] codes H03M13/17 takes precedence

Abstract

An error detection and correction system locates and corrects double errors in a received data block word digitally encoded in a Reed-Solomon code (n,K) as coefficient terms of an n-order codeword polynomial in the form ##EQU1## where the coefficients are powers of the primitive a of a Galois field, which codeword is divisible by a code generator polynomial in the form of a product of a plurality of different factors in the form (x+aⁱ). Four syndrome signals are derived, each corresponding to a respective first order syndrome equal to the remainder upon dividing a received data block word by a respective factor. An error locator responsive to the four syndrome signals derives a reference signal corresponding to the position h of a reference term in the received data block word equally spaced relative to the two terms containing errors, and a relative position signal k corresponding to the number of terms by which the terms containing the errors are spaced from the reference term. An error solution circuit responsive to the four syndrome signals produces first and second error solution signals corresponding to the differences between the respective received terms containing the errors and the corresponding terms as encoded. An error correction circuit adds the error solution signals to the received data block at the proper places to reproduce the encoded codeword. A fifth syndrome signal may be used to determine if there are more than two errors. Two syndromes may be used to correct single errors.

Claims

1. An error detection and correction system for locating and correcting double errors in a received data block word digitally encoded in a Reed-Solomon code (n,K) as coefficient terms of an n-order codeword polynomial in the form ##EQU20## where i is an integer and the coefficients are powers of the primitive a of a Galois field, which codeword is divisible by a code generator polynomial in the form of a product of a plurality of different factors in the form (x+aⁱ), where aⁱ is a power of said primitive a, said system comprising:

syndrome generation means responsive to signals representing a received data block word for deriving four syndrome signals each corresponding to a respective first order syndrome equal to the remainder upon dividing in said Galois field said received data block word by a respective said factor;

error locator means for locating the terms in said received data block word containing respective said errors, said error locator means comprising means operating in said Galois field and responsive to said four syndrome signals for deriving a reference signal corresponding to the position of a reference term in said received data block word equally spaced relative to said terms containing said errors, means operating in said Galois field and responsive to said four syndrome signals for deriving a relative position signal corresponding to the number of terms by which the terms containing said errors are spaced from said reference term, and means for combining said reference signal and said relative position signal to form combined error location signals corresponding to the positions of the terms containing the respective errors;

error solution means responsive to said four syndrome signals for producing first and second error solution signals corresponding to the differences between the respective received terms containing said errors and the corresponding terms as encoded; and

error correction means responsive to said error location signals and said error solution signals for correcting said signals representing a received data block word.

2. An error detection and correction system according to claim 1 wherein said syndrome generator means includes means for deriving a fifth syndrome signal corresponding to a respective first order syndrome equal to the remainder upon dividing in said Galois field said received data block word by a fifth different respective said factor, and wherein said system includes means responsive to said fifth syndrome and to the outputs of said error locator means and said error solution means for producing an inhibiting signal when a received data block signal contains more than two errors, and means responsive to said inhibiting signal for inhibiting operation of said error correction means when there are more than two errors in a received data block signal.

3. An error detection and correction system according to claim 2 further comprising means responsive to said syndrome signals for providing a single error indicating signal when said received data block signal contains but a single error, said error locator means including means responsive to said single error indicating signal and two of said syndrome signals for providing an error locator signal corresponding to the position of the term containing said single error, and said error solution means including means responsive to said single error indicating signal and two said syndrome signals for providing an error solution signal corresponding to the difference between the received term containing said single error and the corresponding term as encoded.

4. An error detection and correction system according to claim 1 further comprising means responsive to said syndrome signals for providing a single error indicating signal when said received data block signal contains but a single error, said error locator means including means responsive to said single error indicating signal and two of said syndrome signals for providing an error locator signal corresponding to the position of the term containing said single error, and said error solution means including means responsive to said single error indicating signal and two said syndrome signals for providing an error solution signal corresponding to the difference between the received term containing said single error and the corresponding term as encoded.

5. An error detection and correction system according to any one of claims 1 to 4 wherein said means for deriving a reference signal operates in said Galois field to derive a signal corresponding to the term a^h, where h corresponds to said reference term, and includes means for evaluating h to form said reference signal, and wherein said means for deriving a relative position signal derives a signal corresponding to the term (a^k +a^-k), where k corresponds to the number of terms by which the terms containing said errors are spaced from said reference term, and includes means for evaluating k to form said relative position signal.

6. A method for locating and correcting double errors in a received data block word digitally encoded in a Reed-Solomon code (n,K) as coefficient terms of an n-order codeword polynomial in the form ##EQU21## where i is an integer and the coefficients are powers of the primitive a of a Galois field, which codeword is divisible by a code generator polynomial in the form of a product of a plurality of different factors in the form (x+aⁱ), where aⁱ is a power of said primitive a, said method comprising:

responsive to signals representing a received data block word, deriving four syndrome signals each corresponding to a respective first order syndrome equal to the remainder upon dividing in said Galois field said received data block word by a respective said factor;

locating the terms in said received data block word containing respective said errors by operating in said Galois field in response to said four syndrome signals to derive a reference signal corresponding to the position of a reference term in said received data block word equally spaced relative to said terms containing said errors, operating in said Galois field in response to said four syndrome signals to derive a relative position signal corresponding to the number of terms by which the terms containing said errors are spaced from said reference term, and combining said reference signal and said relative position signal to form combined error location signals corresponding to the positions of the terms containing the respective errors;

responsive to said four syndrome signals, producing first and second error solution signals corresponding to the differences between the respective received terms containing said errors and the corresponding terms as encoded; and

responsive to said error location signals and said error solution signals, correcting said signals representing a received data block word.

7. A method according to claim 6 further comprising deriving a fifth syndrome signal corresponding to a respective first order syndrome equal to the remainder upon dividing in said Galois field said received data block word by a fifth different respective said factor, and responsive to said fifth syndrome and to the outputs of said error locator means and said error solution means, producing an inhibiting signal when a received data block signal contains more than two errors, and responsive to said inhibiting signal inhibiting said error correction when there are more than two errors in a received data block signal.

8. A method according to claim 7 further comprising in response to said syndrome signals providing a single error indicating signal when said received data block signal contains but a single error, in response to said single error indicating signal and two of said syndrome signals providing an error locator signal corresponding to the position of the term containing said single error, and responsive to said single error indicating signal and two said syndrome signals providing an error solution signal corresponding to the difference between the received term containing said single error and the corresponding term as encoded.

9. A method according to claim 6 further comprising in response to said syndrome signals providing a single error indicating signal when said received data block signal contains but a single error, in response to said single error indicating signal and two of said syndrome signals providing an error locator signal corresponding to the position of the term containing said single error, and responsive to said single error indicating signal and two said syndrome signals providing an error solution signal corresponding to the difference between the received term containing said single error and the corresponding term as encoded.

10. A method according to any one of claims 6 to 9 wherein deriving said reference signal includes operating in said Galois field to derive a signal corresponding to the term a^h, where h corresponds to said reference term, and evaluating h to form said reference signal, and wherein deriving said relative position signal includes deriving a signal corresponding to the term (a^k +a^-k), where k corresponds to the number of terms by which the terms containing said errors are spaced from said reference term, and evaluating k to form said relative position signal.

11. A decoder for error correcting a received digital data block R(x) encoded as a codeword C(x) having a plurality of data bytes and a plurality of check bytes, said codeword being encoded in a Reed-Solomon code such that division of the received data block by the generator polynomial will produce a plurality of syndromes S_i which contain information about the errors contained in the received code words, said decoder comprising:

means for generating a plurality of syndromes from a received code word,

input register means for storing the generated syndromes when they become available from said syndrome generator,

an arithmetic/logic unit (ALU) for performing arithmetic or logical operations on one or two operands input to a first channel or to a second channel of said unit in accordance with a control word and for outputting the results of the operation to an accumulator,

output register means for storing the locations of the errors and the values of the errors in said received data block R(x),

internal register means for the temporary storage of the output of said accumulator,

a first data path connecting the output of said input register means and the output of said internal register means to one of said first and second channels of said ALU,

a second data path connecting the output of said accumulator to the input of said output register means, the input of said internal register means, and the other of said first and second channels of said ALU, and

processor means for controlling the transfer of data over said first and second data paths and for generating a control word which selects an operation for the (ALU).

12. A decoder as set forth in claim 11 wherein:

said first data path has disposed between its input and output a plurality of translational means adapted to transform data transmitted along said select path into a corresponding value of a predetermined function of the data.

13. A decoder as set forth in claim 12 wherein one of said translational means comprises:

a buffer which transmits received data unmodified.

14. A decoder as set forth in claim 12 wherein one of said translational means comprises:

means for translating a received data value into the logarithm of that value.

15. A decoder as set forth in claim 12 wherein said processor means includes:

means for enabling one of said translational means while disabling the rest during a data transfer over said first path, said enablement providing for the selection of the type of translations for the data value desired at the destination of the transfer.

16. A decoder as set forth in claim 11 wherein:

said second data path has disposed between its input and output a plurality of translational means adapted to transform data transmitted along said second path into a corresponding value of a predetermined function of the data.

17. A decoder as set forth in claim 16 wherein one of said translational means comprises:

first means which transmits received data unmodified.

18. A decoder as set forth in claim 16 wherein one of said translational means comprises:

second means for translating a received data value into an antilogarithm of that value.

19. A decoder as set forth in claim 16 wherein one of said translational means comprises:

third means for translating a received data value R into a value for k according to the function;

space="preserve" listing-type="equation">a.sup.k +a.sup.-k =R

where a is the primitive of a Galois field.

20. A decoder as set forth in claim 16 wherein said processor means includes:

means for enabling one of said translational means while disabling the rest during a data transfer over said second path, said enablement providing for the selection of the type of translation for the data value desired at the destination of the transfer.