Frame synchronization with unique-word dependent filter coefficients

• US 6,480,559 B1
• Filed: 11/12/1998
• Issued: 11/12/2002
• Est. Priority Date: 11/12/1997
• Status: Expired due to Term

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Dabak, Anand G.
• Primary Examiner(s): Chin, Stephen
• Assistant Examiner(s): Kim, Kevin
• Application Number:
  US09/191,096
• Time in Patent Office: 1,461 Days
• Field of Search: 375/341, 375/368, 375/354, 370/350, 370/324
• US Class Current: 375/368
• CPC Class Codes: H04L 7/042 Detectors therefor, e.g. correlators, state machines digital correlators in general G06F17/15
• CPC Class Codes: H04L 7/046 using a dotting sequence

Abstract

A TDMA system in which the mobile receiver performs unique word detection (and hence frame synchronization) by using real correlation coefficients which are not equal to the binary unique word (nor to any shift or scaling of it), but which are dependent on the unique word. In some embodiments, the correlation coefficients are dependent both on the unique word and also on the bit_sync pattern.

Claims

1. A method for detection of a unique word, which is normally preceded by a synchronization pattern, in a data stream, comprising the actions of:

(a.) feeding the data stream into a finite impulse response filter structure which applies a coefficient vector to derive a receiver output signal from said data stream, said coefficient vector

having more than binary resolution, and

being dependent on the specific value of the unique word, and lying within the space spanned by a collection of vectors Vⁱ, and being near said vectors Vⁱ;

wherein each said vector Vⁱ represents the difference between said unique word and a respective candidate vector Sⁱ which is a subsequence taken from the concatenation of said unique word with said synchronization pattern; and

(b.) detecting said unique word in said bit stream whenever said step (a.) produces a receiver output signal which is above a threshold.

2. The method of claim 1, wherein said unique word is also optimized, in combination with said coefficient vector, for optimal performance of said detecting step (b.).

3. The method of claim 1, wherein elements of said coefficient vector each have at least 4 bits of resolution.

4. The method of claim 1, wherein said coefficient vector is derived from less than two-thirds of the set of all possible vectors Vⁱ.

5. A method for detection of a unique word in a bit stream, comprising the actions of:

(a.) repeatedly testing to detect a near match between said incoming bit stream and a set of stored coefficients;

wherein said stored coefficients vary in dependence on the specific value of said unique word, and are not equal to said unique word nor to any shift or multiple thereof; and

(b.) detecting said unique word in said bit stream whenever said step (a.) detects a near match.

6. The method of claim 5, wherein said stored coefficients are also dependent on a pattern of symbol synchronization bits which predictably precede said unique word with an unknown phase offset interposed there between.

7. The method of claim 5, wherein said testing step is a correlation operation.

8. The method of claim 5, wherein said testing step uses a filter structure with a non-binary coefficient vector X_opt which is optimized according to the formula ${\underset{\_}{X}}_{\mathrm{opt}}=\underset{\underset{\_}{X}}{\max }$

 



mini

 



〈

U_-S_i,X〉

where U is the unique word, and Sⁱ are subsequences which are extracted from the concatenation of said unique word with at least part of a bit synchronization pattern which normally precedes said unique word.

9. The method of claim 8, wherein said subsequences Sⁱ each have the same length as said unique word.

10. A method for receiving synchronous communication, in a format where a unique word, which is normally preceded by a synchronization pattern, marks frame boundaries in a data stream, comprising the actions of:

(a.) feeding the data stream into a finite impulse response filter structure which applies a coefficient vector to derive a receiver output signal from said data stream, said coefficient vector

having more than binary resolution, and

being dependent on the specific value;

of the unique word, and lying within the space spanned by a collection of vectors Vⁱ, and being near said vectors Vⁱ;

wherein each said vector Vⁱ represents the difference between said unique word and a respective candidate vector Sⁱ which is a subsequence taken from the concatenation of said unique word with said synchronization pattern; and

(b.) detecting a frame boundary in said;

bit stream whenever said step (a.) produces a receiver output signal which is above a threshold; and

(c.) extracting data from said bit stream at only some offsets from the frame boundary detected in said step (b.).

11. The method of claim 10, wherein said unique word is also optimized, in combination with said coefficient vector, for optimal performance of said detecting step (b.).

12. The method of claim 10, wherein elements of said coefficient vector each have at least 4 bits of resolution.

13. The method of claim 10, wherein said coefficient vector is derived from less than two-thirds of the set of all possible vectors Vⁱ.

14. A receiving method for synchronous communication, comprising the actions of:

(a.) repeatedly testing to detect a near match between an incoming bit stream and a set of stored coefficients;

wherein said stored coefficients vary in dependence on the specific value of a unique word which marks the timing of a frame boundary, and are not equal to said unique word nor to any bit rotation thereof;

(b.) detecting a frame boundary in said bit stream whenever said step (a.) detects a near match; and

(c.) extracting data from said bit stream at only some offsets from the frame boundary detected in said step (b.).

15. The method of claim 14, wherein said stored coefficients are also dependent on a pattern of symbol synchronization bits which predictably precede said unique word.with an unknown phase offset interposed there between.

16. The method of claim 14, wherein said testing step is a correlation operation.

17. The method of claim 14, wherein said testing step uses a filter structure with a non-binary coefficient vector X_opt which is optimized according to the formula ${\underset{\_}{X}}_{\mathrm{opt}}=\underset{\underset{\_}{X}}{\max }$

 



mini

 



〈

U_-S_i,X〉

where U is the unique word, and Sⁱ are subsequences which are extracted from the concatenation of said unique word with at least part of a bit synchronization pattern which normally precedes said unique word.

18. The method of claim 17, wherein said subsequences Sⁱ each have the same length as said unique word.

19. A time-domain-segmented communication system, comprising:

a transmitter which transmits a bit stream which is formatted in frames of data; and

a receiver which performs the actions of;

(a.) repeatedly testing to detect a near match between an incoming bit stream and a set of stored coefficients;

wherein said stored coefficients vary in dependence on the specific value of a unique word which marks the timing of a frame boundary, and are not equal to said unique word nor to any bit rotation thereof;

(b.) detecting a frame boundary in said bit stream whenever said step (a.) detects a near match; and

(c.) extracting data from said bit stream at only some offsets from the frame boundary detected in said step (b.).

20. The system of claim 19, wherein said stored coefficients are also dependent on the pattern of symbol synchronization bits which predictably precede said unique word with an unknown phase offset interposed there between.

21. The system of claim 19, wherein said testing step is a correlation operation.

22. The system of claim 19, wherein said testing step uses a filter structure with a non-binary coefficient vector X_opt which is optimized according to the formula ${\underset{\_}{X}}_{\mathrm{opt}}=\underset{\underset{\_}{X}}{\max }$

 



mini

 



〈

U_-S_i,X〉

where U is the unique word, and Sⁱ are subsequences which are extracted from the concatenation of said unique word with at least part of a bit synchronization pattern which normally precedes said unique word.

23. The system of claim 22, wherein said subsequences Sⁱ each have the same length as said unique word.

24. A portable telephone system, comprising:

a wireless receiver, and digitizing stages which provide a digital bit stream as output in dependence on signals received by said receiver; and

at least one signal processing stage receiver which is connected to receive said bit stream, and which performs the actions of;

(a.) repeatedly testing to detect a near match between said bit stream and a set of stored coefficients;

wherein said stored coefficients vary in dependence on the specific value of a unique word which marks the timing of a frame boundary, and are not equal to said unique word nor to any bit rotation thereof;

(b.) detecting a frame boundary in said bit stream whenever said step (a.) detects a near match; and

(c.) extracting data from said bit stream at only some offsets from the frame boundary detected in said step (b.).

25. The system of claim 24, further comprising a microphone and a speaker.

26. The system of claim 24, wherein said stored coefficients are also dependent on a pattern of symbol synchronization bits which predictably precede said unique word with an unknown phase offset interposed there between.

27. The system of claim 24, wherein said testing step is a correlation operation.

28. The system of claim 24, wherein said testing step uses a filter structure with a non-binary coefficient vector X_opt which is optimized according to the formula ${\underset{\_}{X}}_{\mathrm{opt}}=\underset{\underset{\_}{X}}{\max }$

 



mini

 



〈

U_-S_i,X〉

where U is the unique word, and Sⁱ are subsequences which are extracted from the concatenation of said unique word with at least part of a bit synchronization pattern which normally precedes said unique word.

29. The system of claim 28, wherein said subsequences Sⁱ each have the same length as said unique word.