Noncoherent receiver employing a dual-maxima metric generation process

• US 5,442,627 A
• Filed: 06/24/1993
• Issued: 08/15/1995
• Est. Priority Date: 06/24/1993
• Status: Expired due to Term

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Viterbi, Audrey, Viterbi, Andrew J.
• Primary Examiner(s): Marcelo, Melvin
• Assistant Examiner(s): VU, HUY DUY
• Application Number:
  US08/083,110
• Time in Patent Office: 782 Days
• Field of Search: 370/19, 370/21, 370/22, 370/18, 380/28, 380/33, 380/34, 375/1, 371/37.1, 371/37.8, 371/37.9, 371/43
• US Class Current: 370/209
• CPC Class Codes: H04B 1/707 using direct sequence modulation
• CPC Class Codes: H04B 1/7115 Constructive combining of multi-path signals, i.e. RAKE receivers

Abstract

A method and apparatus for decoding an orthogonally encoded data signal in a noncoherent receiver system. The method is referred to as dual-maxima metric generation. It comprises the steps of sequentially searching for a maximum energy level in each of two subsets of a given set of symbol indexes and associated energy levels and calculating a difference of the two values to form a soft decision output value. The two subsets are identified by the binary value (either "0" or "1") of a given digit of the binary equivalent of the symbol index. The soft decision output value reflects a measure of confidence of the value of the corresponding digit of the original signal. The dual-maxima generator sequences through the steps one time for each binary digit of the original signal. The method allows the correlated energy from multiple receivers to be combined before the decoding of the signals, thus further reducing the complexity of the circuitry and improving the performance of the decoder.

Claims

1. A method of decoding an orthogonally encoded data signal corresponding to an original data signal, said orthogonally encoded data signal having a fixed number of possible code values each having an index number wherein a binary equivalent of said index number corresponds to said original data signal, comprising the steps of:

receiving a set of energy values, each energy value corresponding to a respective one of said fixed number of code values and corresponding index number;

searching a first subset of said set of energy values for a maximum energy value wherein said first subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a preselected digit;

searching a second subset of said set of energy values for a maximum energy value wherein said second subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as said preselected digit; and

forming a difference between said maximum energy value of said first subset and said maximum energy of said second subset wherein said difference is a measure of confidence of the value of a particular digit of said original data signal.

2. The method of claim 1 further comprising the steps of:

searching a third subset of said set of energy values for a maximum energy value wherein said third subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a second preselected digit;

searching a fourth subset of said set of energy values for a maximum energy value wherein said fourth subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as said second preselected digit; and

forming a difference between said maximum energy value of said third subset and said maximum energy of said fourth subset wherein said difference is a measure of confidence of the value of a second particular digit of said original data signal.

3. The method of claim 1 wherein each energy value in said set of energy values is generated by an orthogonal code transformer.

4. The method of claim 1 wherein each energy value in said set of energy values is generated by summing a plurality of partial energy values corresponding to a like symbol index, each of said plurality of partial energy values corresponding to said like symbol index is generated by a different one of a plurality of orthogonal code transformers.

5. The method of claim 3 wherein said orthogonal code transformer comprises a fast Hadamard transformer.

6. The method of claim 4 wherein said plurality of orthogonal code transformers comprises a fast Hadamard transformer.

7. The method of claim 1 wherein said orthogonally encoded signal is a Walsh encoded signal.

8. The method of claim 1 wherein said measure of confidence represents a soft decision data value.

9. A method of decoding an orthogonally encoded data signal corresponding to an original data signal, said orthogonally encoded data signal having a fixed number of possible code values each having an index number wherein a binary equivalent of said index number corresponds to said original data signal, comprising the steps of:

(a) receiving a set of energy values, each energy value corresponding to a respective one of said fixed number of code values and corresponding index number;

(b) searching a first subset of said set of energy values for a maximum energy value wherein said first subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a first digit;

(c) searching a second subset of said set of energy values for a maximum energy value wherein said second subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as a first digit;

(d) forming a difference of said maximum energy value of said first subset and said maximum energy of said second subset wherein said difference is a measure of confidence of the value of a particular digit of said original data signal; and

(e) repeating steps (b), (c), and (d) for each subsequent digit of said index number.

10. A method of decoding an orthogonally encoded data signal corresponding to an original data signal, said orthogonally encoded data signal having a fixed number of possible code values each having an index number wherein a binary equivalent of said index number corresponds to said original data signal, comprising the steps of:

receiving a set of energy values, each energy value corresponding to a respective one of said fixed number of code values and corresponding index number;

searching a first subset of said set of energy values for a maximum energy value wherein said first subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a first digit;

searching a second subset of said set of energy values for a maximum energy value wherein said second subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as said first digit;

forming a difference of said maximum energy value of said first subset and said maximum energy of said second subset wherein said difference is a measure of confidence of the value of a first digital of said original data signal;

searching a third subset of said set of energy values for a maximum energy value wherein said third subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as an intermediate digit;

searching a fourth subset of said set of energy values for a maximum energy value wherein said fourth subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as said intermediate digit;

forming a difference of said maximum energy value of said third subset and said maximum energy of said fourth subset wherein said difference is a measure of confidence of the value of an intermediate digit of said original data signal;

searching a fifth subset of said set of energy values for a maximum energy value wherein said fifth subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a last digit;

searching a sixth subset of said set of energy values for a maximum energy value wherein said sixth subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as said last digit; and

forming a difference of said maximum energy value of said fifth subset and said maximum energy of said sixth subset wherein said difference is a measure of confidence of the value of a last digit of said original data signal.

11. An apparatus for decoding an orthogonally encoded signal to produce a corresponding original signal comprising:

at least one demodulator for receiving and demodulating said encoded signal and providing groups of samples of said signal as an output;

at least one signal transformer connected to receive said groups of samples and to generate a plurality of soft decision data which each correspond to a measure of confidence that a data sample of said encoded signal is substantially similar to one orthogonal code from a set of mutually orthogonal codes, and said soft decision data each having a corresponding index symbol with a binary equivalent value which corresponds to one of said mutually orthogonal codes;

metric generation means coupled to said transformer for finding a set of paired data values, each paired data value of said set of paired data values corresponding to a digit of said binary equivalent of each of said index symbols, a first value of said paired data values corresponding to a maximum value of said soft decision data having a binary equivalent of said corresponding index symbol with a "0" as a corresponding digit and a second value of said paired data values corresponding to a maximum value of said soft decision data having a binary equivalent of said corresponding index symbol with a "1" as a corresponding digit, and for subtracting said second value from said first value of said paired data values so as to form a soft decision output value for each digit of said binary equivalent of said index symbols.

12. The apparatus of claim 11 further comprising conversion means for converting said resultant soft decision data to an estimate of said original signal.

13. The apparatus of claim 12 wherein said conversion means comprises a convolutional decoder.

14. The apparatus of claim 12 wherein said conversion means comprises a Viterbi decoder.

15. The apparatus of claim 11 wherein said, paired data values correspond to dual-maxima aggregate data values.

16. The apparatus of claim 11 further comprising:

a plurality of demodulators for receiving and demodulating said encoded signal and providing groups of samples of said signal as outputs;

a plurality of signal transformers each connected to receive groups of samples from one of said demodulators and to generate a plurality of soft decision data which each correspond to a measure of confidence that a data sample of said encoded signal is substantially similar to one orthogonal code from a set of mutually orthogonal codes, and said soft decision data each having a corresponding index symbol which corresponds to one of said mutually orthogonal codes; and

summation means coupled to said plurality of signal transformers for summing said plurality of soft decision data according to like corresponding index symbols to provide an aggregate plurality of soft decision data and corresponding index symbols, and having an output coupled to said metric generator.

17. A method of decoding an orthogonally encoded data signal corresponding to an original data signal, said orthogonally encoded data signal having a fixed number of possible code values each having an index number wherein a binary equivalent of said index number corresponds to said original data signal, comprising the steps of:

(a) receiving a set of energy values, each energy value corresponding to a respective one of said fixed number of code values and corresponding index number;

(b) determining a first resultant value from a first subset of said set of energy values according to a predetermined function wherein said first subset contains each of said energy values corresponding to an index number having a binary equivalent with a "0" as a first digit;

(c) determining a second resultant value from a second subset of said set of energy values according to said predetermined function wherein said second subset contains each of said energy values corresponding to an index number having a binary equivalent with a "1" as a first digit;

(d) forming a difference between said first resultant value and said second resultant value wherein said difference is a measure of confidence of the value of a particular digit of said original data signal; and

(e) repeating steps (b), (c), and (d) for each subsequent digit of said binary equivalent for each index number.

18. The method of claim 17 wherein each energy value in said set of energy values is generated by an orthogonal code transformer.

19. The method of claim 17 wherein each energy value in said set of energy values is generated by summing a plurality of partial energy values corresponding to a like symbol index, each of said plurality of partial energy values corresponding to said like symbol index is generated by a different one of a plurality of orthogonal code transformers.

20. The method of claim 17 wherein said orthogonal code transformer comprises a fast Hadamard transformer.

21. The method of claim 19 wherein said plurality of orthogonal code transformers comprises a fast Hadamard transformer.

22. The method of claim 17 wherein said orthogonally encoded signal is a Walsh encoded signal.

23. The method of claim 17 wherein said measure of confidence represents a soft decision data value.

24. The method of claim 17 wherein said first and second resultant values are maxima and said predetermined function is a searching function for finding a maximum value.

25. The method of claim 17 wherein said predetermined function is a summation function for finding a sum of said set of values.

26. The method of claim 17 wherein said predetermined function is an exponential function of said set of values.

27. A method of decoding an orthogonally encoded data signal corresponding to an original data signal in a noncoherent receiver system, where the encoded data signal has a substantially constrained number of possible code values each in turn having an index number with an N-bit binary equivalent value which corresponds to said original data signal, comprising the steps of:

receiving a set of energy values, each corresponding to a respective one of said code values;

searching a first subset of said set of energy values which comprises those values having a binary equivalent for the index number with a "0" as a k^th digit, where in k is an integer and 1≦

k≦

N, for a desired energy value determined by a preselected energy function;

searching a second subset of said set of energy values which comprises those values having a binary equivalent for the index number with a "1" as said k^th digit, for a second desired energy value determined by said preselected energy function; and

generating a sort decision value from a difference between said desired values for said first and second subsets.

28. The method of claim 27 further comprising the steps of repeating said searching and generating steps for first and second subsets for each other value of k.

29. The method of claim 27 wherein said soft decision value comprises a measure of confidence of the value of a particular digit of said original signal.