Method and apparatus for joint timing synchronization and frequency offset estimation

• US 6,363,131 B1
• Filed: 07/02/1998
• Issued: 03/26/2002
• Est. Priority Date: 04/29/1998
• Status: Expired due to Term

• Current Assignee: Hughes Network Systems LLC (Echostar Corporation)
• Original Assignee: Hughes Electronics Corporation (Rtx Corporation)
• Inventors: Antia, Yezdi F., Soleimani, Mohammad, Beidas, Bassel F.
• Primary Examiner(s): Pham, Chi
• Assistant Examiner(s): AL BESHRAWI, TONY
• Application Number:
  US09/109,798
• Time in Patent Office: 1,363 Days
• Field of Search: 375/368, 375/13, 375/90, 375/336, 375/340, 375/1, 375/207, 375/20, 375/210, 375/150, 375/206, 380/33, 178/88, 370/100.1, 370/206
• US Class Current: 375/368
• CPC Class Codes: H04L 2027/0065 Frequency error detectors H04L2027/0067 takes precedence
• CPC Class Codes: H04L 2027/0093 Intermittant signals
• CPC Class Codes: H04L 27/22 Demodulator circuits; Receiver circuits
• CPC Class Codes: H04L 7/042 Detectors therefor, e.g. correlators, state machines digital correlators in general G06F17/15

Abstract

A burst analyzer is useful in a digital communication system in which a signal burst has a plurality of reference segments distributed within the signal burst for transmission of a plurality of reference signals. The burst analyzer includes a filter that compares the received signal, for each reference segment thereof, with each reference signal offset by one of a plurality of time offsets to generate correlation data. The burst analyzer then determines, for each reference segment of the signal burst, a maximum correlation value from the correlation data for each time offset. Then the burst analyzer determines the time offset at which a sum of the maximum correlation values, over the plurality of reference segments, is a maximum. The burst analyzer jointly generates a frequency domain representation of the correlation data associated with the time offset at which the maximum correlation sum is a maximum, and then determines a frequency at which the frequency domain representation is a maximum.

Claims

1. A burst analyzer for use in a digital communication system transmitting a signal burst, the burst analyzer comprising:

a filter that generates correlation data based on the signal burst and a plurality of reference signals offset by a plurality of time offsets; and

an accumulator that combines quantities based on the correlation data generated by the filter for a plurality of reference segments distributed within the signal burst.

2. The burst analyzer of claim 1, further comprising means coupled to the filter for modifying the correlation data such that the accumulator non-coherently combines the correlation data.

3. The burst analyzer of claim 1, wherein the accumulator comprises a summer.

4. The burst analyzer of claim 3, further comprising:

first means for determining, for each reference segment of the signal burst, a maximum correlation value from the correlation data for each time offset; and

second means for determining the time offset at which a sum of the maximum correlation values calculated by the summer for the plurality of reference segments is a maximum.

5. The burst analyzer of claim 1, wherein the signal burst is modulated according to a memory-inducing modulation scheme.

6. The burst analyzer of claim 1, wherein the filter compares the signal burst and the reference signals over a correlation interval spanning a time period corresponding to the length of each reference segment.

7. The burst analyzer of claim 1, wherein the filter compares the signal burst and the reference signals over a correlation interval spanning a time period greater than the length of each reference segment.

8. The burst analyzer of claim 1, wherein:

each reference segment comprises a unique word; and

at least one of the reference signals is representative of an average of a plurality of possible signals representative of the unique word subsequent to modulation.

9. The burst analyzer of claim 1, wherein the signal burst comprises a GMSK-modulated signal.

10. The burst analyzer of claim 1, wherein the filter comprises:

a plurality of matched filters wherein each matched filter has an impulse response matched to a corresponding reference signal of the plurality of reference signals such that each matched filter generates a convolution signal representative of the convolution of the signal burst and the respective impulse response matched to the corresponding reference signal; and

a plurality of samplers coupled to the plurality of matched filters, respectively, that sample the convolution signals at a rate commensurate with a bit transmission rate for the signal burst.

11. The burst analyzer of claim 1, wherein each reference segment transmits an identical unique word.

12. The burst analyzer of claim 1, wherein:

a first reference segment of the plurality of reference segments transmits a first unique word and a second reference segment of the plurality of reference segments transmits a second unique word; and

the first and second unique words are different.

13. The burst analyzer of claim 4, further comprising:

a Fourier transformer coupled to the filter wherein the Fourier transformer generates, from the correlation data, a discrete-time frequency domain representation of the maximum correlation values based on the time offset at which the sum of the maximum correlation values is a maximum; and

third means responsive to the Fourier transformer for determining a frequency at which the discrete-time frequency domain representation is a maximum.

14. The burst analyzer of claim 13, wherein the Fourier transformer comprises means for performing a fast Fourier transform.

15. The burst analyzer of claim 13, wherein the filter compares the signal burst and the reference signals over a correlation interval spanning a time period greater than the length of each reference segment.

16. The burst analyzer of claim 1, wherein the reference segments are separated by respective information segments.

49. A burst analyzer as claimed in claim 1, wherein:

each of said reference segments includes at least one of said reference signals.

50. A burst analyzer as claimed in claim 1, wherein:

said accumulator combines said quantities by summing for a plurality of said reference segments the magnitudes of those of said correlation data having maximum correlation values.

17. A method of synchronization in a digital communication system transmitting a signal burst, the method comprising the steps of:

(a) calculating correlation data based on the signal burst and a plurality of reference signals offset by a plurality of time offsets; and

(b) combining quantities based on the correlation data for a plurality of reference segments distributed throughout the signal burst.

18. The method of claim 17, further comprising the step of preparing the correlation data for non-coherent combination.

19. The method of claim 17, further comprising the step of determining, for each reference segment and each time offset, a maximum correlation value from the correlation data.

20. The method of claim 19, wherein step (b) comprises the step of combining, for each time offset, the maximum correlation values associated with each reference segment into a maximum correlation sum.

21. The method of claim 20, further comprising the step of finding the time offset of the plurality of time offsets at which the maximum correlation sum is a maximum.

22. The method of claim 21, further comprising the steps of:

generating a frequency domain representation of the maximum correlation values based on the time offset at which the maximum correlation sum is a maximum; and

determining a frequency at which the frequency domain representation is a maximum.

23. The method of claim 17, wherein step (a) comprises the step of convolving the signal burst with a set of filters matched to the plurality of reference signals, respectively.

24. The method of claim 17, wherein the signal burst is modulated according to a memory-inducing modulation scheme.

25. The method of claim 17, wherein step (a) is performed, for each reference segment, over an observation interval spanning a time period corresponding to the length of each reference segment.

26. The method of claim 17, wherein step (a) is performed, for each reference segment, over an observation interval spanning a time period greater than the length of each reference segment.

27. The method of claim 17, wherein:

each reference segment comprises a unique word; and

at least one of the reference signals is representative of an average of a plurality of possible signals representative of the unique word subsequent to modulation.

28. The method of claim 17, wherein the signal burst comprises a GMSK-modulated signal.

29. The method of claim 17, wherein step (a) comprises the step of convolving the signal burst with a plurality of filters having respective impulse responses matched to a corresponding reference signal to generate convolution data.

30. The method of claim 29, wherein step (a) further comprises the step of sampling the convolution data at a rate commensurate with a bit transmission rate for the signal burst.

31. The method of claim 17, wherein each reference signal transmits an identical unique word.

32. The method of claim 17, wherein:

a first reference segment of the plurality of reference segments transmits a first unique word and a second reference segment of the plurality of reference segments transmits a second unique word; and

the first and second unique words are different.

33. The method of claim 17, wherein the reference segments are separated by respective information segments.

51. A burst analyzer as claimed in claim 17, wherein:

each of said reference segments includes at least one of said reference signals.

52. A burst analyzer as claimed in claim 17, wherein:

said combining step combines said quantities by summing for a plurality of said reference segments the magnitudes of those of said correlation data having maximum correlation values.

34. A method of synchronization in a digital communication system transmitting a signal burst modulated by a modulation scheme that induces memory therein, the method comprising the steps of:

(a) comparing the signal burst with a plurality of possible reference waveforms to generate comparison data not reflecting the memory induced by the modulation scheme; and

(b) combining quantities based on the comparison data non-coherently to determine a timing offset for the signal burst.

35. The method of claim 34, wherein step (a) comprises the steps of:

convolving the signal burst with a plurality of filters having respective impulse responses associated with the plurality of possible reference waveforms, respectively, such that the comparison data comprises convolution data; and

sampling the convolution data.

36. The method of claim 34, wherein:

the signal burst comprises a plurality of reference segments distributed therein;

step (a) is performed over a plurality of observation intervals; and

each observation interval spans a respective reference segment.

37. The method of claim 36, wherein the observation interval extends beyond each reference segment.

38. The method of claim 36, wherein:

each reference segment comprises a unique word; and

at least one of the reference signals is representative of an average of a plurality of possible signals representative of the unique word subsequent to modulation.

39. The method of claim 34, wherein:

a first reference segment of the plurality of reference segments transmits a first unique word and a second reference segment of the plurality of reference segments transmits a second unique word; and

the first and second unique words are different.

40. The method of claim 34, further comprising the steps of:

generating a frequency domain representation from the comparison data based on the timing offset; and

evaluating the frequency domain representation to determine an offset frequency.

41. The method of claim 40, wherein the step of generating the frequency domain representation comprises calculating a fast Fourier transform of the comparison data based on the timing offset.

42. The method of claim 34, further comprising the step of recording the timing offset to provide information for subsequent burst synchronization.

43. A method for estimating a frequency of a signal burst transmitted in a digital communication system wherein the signal burst comprises a plurality of reference segments, the method comprising the steps of:

(a) calculating, for each reference segment, correlation data representative of an amount of correlation between the signal burst and one of a plurality of reference signals offset by a plurality of time offsets;

(b) generating a frequency domain representation of the correlation data associated with a particular time offset of the plurality of time offsets yielding a maximum amount of correlation over the plurality of reference segments; and

(c) determining the frequency at which the frequency domain representation of the correlation data is a maximum.

44. The method of claim 43, wherein the plurality of reference segments are distributed within the signal burst.

45. The method of claim 43, wherein the signal burst is modulated according to a memory-inducing modulation scheme.

46. The method of claim 43, wherein:

step (a) is performed over a plurality of observation intervals;

each observation interval spans a respective reference segment; and

the observation interval extends beyond each reference segment.

47. The method of claim 43, wherein step (a) comprises convolving the signal burst with a plurality of filters having respective impulse responses matched to the plurality of reference signals, respectively.

48. The method of claim 43, further comprising the step of recording the frequency to provide information for subsequent burst analysis.