System and method of time and frequency synchronization in a radiocommunication system

• US 6,023,477 A
• Filed: 12/18/1996
• Issued: 02/08/2000
• Est. Priority Date: 12/18/1996
• Status: Expired due to Term

• Current Assignee: Unwired Planet LLC (JP Morgan Chase & Co)
• Original Assignee: Ericsson, Inc. (Telefonaktiebolaget LM Ericsson)
• Inventors: Dent, Paul W.
• Primary Examiner(s): VU, HUY DUY
• Application Number:
  US08/768,975
• Time in Patent Office: 1,147 Days
• Field of Search: 370/350, 370/318, 370/321, 370/337, 370/347, 370/442, 370/503, 370/507, 370/508, 370/509, 370/510, 370/511, 370/512, 370/513, 370/514, 370/515, 370/516, 375/356, 375/362, 375/365, 375/366, 375/367, 375/368
• US Class Current: 370/509
• CPC Class Codes: H04B 7/2675 Frequency synchronisation
• CPC Class Codes: H04B 7/2678 Time synchronisation
• CPC Class Codes: H04L 7/042 Detectors therefor, e.g. correlators, state machines digital correlators in general G06F17/15
• CPC Class Codes: H04L 7/06 the synchronisation signals differing from the information signals in amplitude, polarity or frequency or length
• CPC Class Codes: H04L 7/08 the synchronisation signals recurring cyclically
• CPC Class Codes: H04L 7/10 Arrangements for initial synchronisation
• CPC Class Codes: H04W 56/0035 detecting errors in frequency or phase
• CPC Class Codes: H04W 56/0085 detecting a given structure in the signal

Abstract

A communication station such as a hand held telephone quickly identifies a desired carrier signal or time slot of a carrier signal and synchronizes itself in time and frequency to that carrier or slot. Synchronization is achieved even when frequency and time reference signals in the communication station are offset, e.g., due to drift, with respect to the carrier signal or slot by averaging successions of signal measurements that occur at times when corresponding signal values are expected. In one method, a plurality of measurements of the received signal taken at a group of predetermined time points where corresponding values are expected are averaged together, in order to enhance the signal to noise ratio. Averaging is carried out several times using different groups of time points, the groups corresponding to respective, different expected clock frequency errors. The results of the averaging steps are used in determining a best hypothesis for the received signal. Dynamic programming is used to test synchronization hypotheses based on different hypotheses of timing drift rate and to discriminate between different hypotheses of timing and timing drift rate in order to maximize sums of correlations.

Claims

1. In a receiver, a method of obtaining time synchronization with a predetermined portion of a signal broadcast by a transmitter, the predetermined portion being broadcast at predetermined repetition intervals, wherein the receiver determines the repetition intervals with errors, comprising the steps of:

sampling the signal and storing samples at respective locations in a memory, wherein the memory locations are identified by respective addresses;

selecting a postulate of addresses of memory locations in which samples corresponding to a first occurrence of the predetermined portion are likely to have been stored;

determining a succession of measurement values related to the predetermined portion from the samples stored in the memory locations identified by the postulate of addresses, wherein a first measurement value is determined from samples that begin at a starting address and the successive measurement values are determined by incrementing the starting address until all samples stored at the postulate of addresses have been used;

storing the succession of measurement values in memory locations having sequential addresses;

additively combining a selected one of measurement values stored in memory locations having a first group of sequential addresses with measurement values stored in memory locations having a second group of addresses, wherein the first group of addresses corresponds to measurement values determined from a first likely occurrence of the predetermined portion and the second group of addresses corresponds to measurement values determined from a subsequent likely occurrence of the predetermined portion, and the addresses in the first group are separated from the addresses in the second group by amounts corresponding to the repetition interval and in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval; and

determining an optimal one of the combined values, thereby obtaining time synchronization with the predetermined portion.

2. The method of claim 1, wherein the measurement values are signal strength values.

3. The method of claim 1, wherein the measurement values are determined by correlating samples with a known signal pattern.

4. The method of claim 3, wherein the known signal pattern is a digital code.

5. The method of claim 4, wherein the digital code is one of a set of orthogonal digital codes.

6. The method of claim 5, wherein the measurement values are determined by correlating samples with every digital code in the set of orthogonal digital codes.

7. The method of claim 6, wherein a Fast Walsh Transform is used for correlating samples with every digital code in the set of orthogonal digital codes.

8. The method of claim 1, wherein the step of additively combining comprises combining moduli of the measurement values.

9. The method of claim 1, wherein the selected one of the measurement values stored in memory locations having the first group of sequential addresses is the largest measurement value within the range from the repetition interval less the uncertainty to the repetition interval plus the uncertainty and centered on an address corresponding to the address of a measurement value stored in memory locations having the second group of addresses with which the largest measurement value is combined.

10. In a receiver, a method of obtaining time synchronization with characteristic portions of a signal broadcast repetitively by a transmitter at predetermined repetition intervals, comprising the steps of:

receiving the broadcast signal and storing samples of the received broadcast signal at respective locations in a memory, each memory location being identified by a respective address;

postulating a region of addresses identifying memory locations in which samples corresponding to a first occurrence of the characteristic portion are likely to be found;

determining a first measurement value corresponding to the first occurrence of the characteristic portion using samples stored in memory locations having addresses in the postulated address region and beginning with a start address;

determining succeeding measurement values by incrementing the start address until substantially all sample values have been used and storing the first and succeeding measurement values in respective memory locations having successive addresses;

replacing each measurement value or already replaced measurement value in a first group of measurement values or already replaced measurement values stored in memory locations having sequential addresses by a combination of the measurement value or already replaced measurement value and a selected one of a second group of measurement values stored in memory locations having earlier addresses, wherein the addresses of the second group are separated from the addresses of the first group by amounts that correspond substantially to the repetition interval and are in a range from the repetition interval minus an uncertainty in the repetition interval to the repetition interval plus the uncertainty in the interval; and

the replacing starts at a first address for which such a second group exists; and

determining an optimal one of the replaced values, thereby obtaining time synchronization with the repetitive characteristic portion.

11. The method of claim 10, wherein the first and succeeding measurement values are signal strength values.

12. The method of claim 10, wherein the first and succeeding measurement values are determined by correlating samples with a known signal pattern.

13. The method of claim 12, wherein the known signal pattern is a digital code.

14. The method of claim 13, wherein the digital code is one of a set of orthogonal digital codes.

15. The method of claim 14, wherein the first and succeeding measurement values are determined by correlating samples with every digital code in the set of orthogonal digital codes.

16. The method of claim 15, wherein a Fast Walsh Transform is used for correlating samples with every digital code in the set of orthogonal digital codes.

17. The method of claim 10, wherein the combinations are formed by adding moduli of the measurement values.

18. A method of estimating a frequency error between a transmitter and a receiver based on a characteristic portion of a signal broadcast repetitively by the transmitter at predetermined repetition intervals, comprising the steps of:

generating samples of the signal and storing the samples in a memory;

postulating a region of addresses identifying locations in the memory where samples corresponding to an occurrence of the characteristic portion are likely to start to be found;

determining a first measurement value related to the characteristic portion using samples stored in memory locations having addresses in the postulated region;

determining succeeding measurement values related to the characteristic portion by incrementing a start address in the postulated region until all samples stored at memory locations having addresses in the postulated region have been used and storing the first and succeeding measurement values in memory locations having successive memory addresses;

initializing a plurality of counters that are associated with respective ones of the memory values;

additively combining measurement values selected from memory locations having sequential addresses in the memory with a selected one of a group of measurement values stored in memory locations having addresses corresponding to an earlier occurrence of the characteristic portion, wherein the addresses of the measurement values in the group are separated from the addresses of the selected measurement values by amounts corresponding to the repetition interval and in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval;

transferring a counter value associated with the selected one of the group of measurement values to the counter associated with the selected one after incrementing or decrementing the counter value according to which of the group was selected;

determining an optimal one of the additively combined values; and

obtaining a frequency error estimate based on the counter value associated with the optimal one.

19. The method of claim 18, wherein the measurement values are determined by correlating samples with a known signal pattern.

20. The method of claim 19, wherein the known signal pattern is a digital code.

21. The method of claim 20, wherein the digital code is one of a set of orthogonal digital codes.

22. The method of claim 18, wherein the counter value associated with the selected one is incremented by a number of addresses that the selected one lies from a nominal location if the counter value is associated with an earlier timing, is decremented by the number of addresses if the counter value is associated with a later timing, and is unchanged if the optimal one lies at the nominal location.

23. In a receiver, a method of estimating a frequency error between a transmitter and the receiver based on a signal characteristic broadcast repetitively by the transmitter at predetermined repetition intervals, comprising the steps of:

sampling a received signal characteristic and storing the samples in respective locations in a memory;

postulating a region of addresses identifying memory locations where samples of an occurrence of signal characteristic are likely to start;

determining a measurement value related to the signal characteristic using samples stored in locations having addresses in the postulated region of addresses;

determining succeeding measurement values by incrementing a start address until all sample values stored in memory locations having addresses in the postulated region have been used and storing the measurement values in memory locations having successive addresses;

initializing a plurality of counters associated with the measurement values;

replacing each measurement value or already replaced measurement value in a first group of measurement values or already replaced measurement values stored in memory locations having sequential addresses by a combination of the measurement value or already replaced measurement value and a selected one of a second group of measurement values stored in memory locations having earlier addresses, wherein the addresses of the second group are separated from the addresses of the first group by amounts that correspond substantially to the repetition interval and are in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval, starting at a first address for which such a second group exists;

transferring a counter value associated with the selected one of the second group of measurement values to the counter associated with the selected one after incrementing or decrementing the counter value according to which of the group was selected;

determining an optimal one of the replaced values; and

determining a frequency error estimate based on the counter value associated with the optimal one.

24. The method of claim 23, wherein the measurement values are determined by correlating samples with a known signal pattern.

25. The method of claim 24, wherein the known signal pattern is a digital code.

26. The method of claim 25, wherein the digital code is one of a set of orthogonal digital codes.

27. The method of claim 23, wherein the counter value associated with the selected one is incremented by a number of addresses that the selected one lies from a nominal location if the counter value is associated with an earlier timing, is decremented by the number of addresses if the counter value is associated with a later timing, and is unchanged if the optimal one lies at the nominal location.

28. A method of time synchronizing a receiver with a signal characteristic broadcast repetitively by a transmitter at predetermined repetition intervals, comprising the steps of:

storing samples of a broadcast signal characteristic in locations in a memory;

postulating a region of addresses identifying memory locations where samples of the signal characteristic are likely to start;

determining a measurement value related to the signal characteristic based on samples from the postulated region of addresses;

determining succeeding measurement values by incrementing a start address until all sample values having addresses in the postulated region have been used and storing the measurement values in memory locations having successive addresses;

determining cumulative measurement values associated with the measurement values by combining each measurement value with a selected one of a group of measurement values stored in memory locations having addresses separated by amounts corresponding to the repetition interval and in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval, starting at a first address for which such a group exists; and

determining an optimal one of the cumulative measurement values, thereby obtaining synchronization with the signal characteristic.

29. A method of time synchronizing a receiver with a signal characteristic broadcast repetitively by a transmitter at predetermined repetition intervals, comprising the steps of:

receiving the signal characteristic and storing samples of the received signal characteristic in locations in a memory;

postulating a region of addresses of memory locations where samples of a first occurrence of the signal characteristic are likely to start;

determining a measurement value related to the signal characteristic from samples stored in memory locations having addresses in the postulated address region;

determining succeeding measurement values by incrementing a start address until all stored samples have been used and storing the measurement values in memory locations having successive addresses;

determining cumulative measurement values associated with the stored measurement values by combining each stored measurement value with a selected one of a group of measurement values and cumulative measurement values, the group being stored in memory locations having addresses separated by amounts corresponding to the repetition interval and in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval, starting at a first address for which such a group exists; and

determining an optimal one of the cumulative measurement values, thereby obtaining synchronization with the signal characteristic.

30. A method of time synchronizing a receiver and obtaining an estimate of a synchronization error based on a signal characteristic broadcast repetitively by a transmitter at predetermined repetition intervals, comprising the steps of:

receiving the signal characteristic and storing samples of the signal characteristic in locations in a memory;

postulating an address region in the memory, wherein the address region includes addresses that identify memory locations where samples corresponding to a first occurrence of the signal characteristic are likely to start;

determining a measurement value related to the signal characteristic from samples retrieved from memory locations having addresses in the postulated address region;

determining succeeding measurement values by incrementing a start address until all sample values have been used and storing the measurement values in memory locations having successive addresses;

initializing a plurality of counters that are associated with respective ones of the measurement values;

determining cumulative measurement values associated with the stored measurement values by combining each stored measurement value with a selected one of a group comprising at least either measurement values or cumulative measurement values, the group being stored at earlier addresses separated by amounts corresponding to the repetition interval and in a range from the repetition interval less an uncertainty in the interval to the repetition interval plus the uncertainty in the interval, starting at a first address for which such a group exists;

transferring a counter value associated with the selected one of the group to the counter associated with a measurement value after incrementing or decrementing the counter value according to which of the group was selected; and

determining an optimal one of the cumulative values, thereby obtaining time synchronization with the signal characteristic, and obtaining an error estimate based on the counter value associated with the optimal one.

31. The method of claim 30, wherein the counter value associated with the selected one is incremented by a number of addresses that the selected one lies from a nominal location if the counter value is associated with an earlier timing, is decremented by the number of addresses if the counter value is associated with a later timing, and is unchanged if the optimal one lies at the nominal location.

32. A method of providing coarse time synchronization to time division multiple access (TDMA) transmissions from a communication satellite to a receiving terminal, comprising:

transmitting from the communication satellite a sequence of signals occupying sequential time slots in a TDMA frame;

designating at least one of the time slots in the TDMA frame for transmitting a synchronization signal;

transmitting a succession of TDMA frames in which the designated time slots in certain frames are transmitted at a higher power level according to a pattern that is known to the communication satellite and the receiving terminal;

receiving the signal sequence and averaging a received signal power over a sliding time window one time slot in duration to obtain measurements of slot energy at a succession of time intervals;

storing the slot energy measurements for a plurality of TDMA frames in a memory having addressable storage locations, a memory address for a given slot energy measurement corresponding to the sliding time window'"'"'s position for the measurement;

determining a cumulative energy result corresponding to a memory address postulate by the steps of;

postulating a memory address at which a slot energy measurement for one of the designated time slots is stored;

determining, based on the postulated memory address and the pattern, memory addresses of other slot energy measurements for other designated time slots;

combining the slot energy measurements stored at the postulated memory address and the determined memory addresses to obtain a cumulative energy result for the postulated address over an energy accumulation period equal to a duration of the plurality of TDMA frames; and

storing the cumulative energy result in a memory location corresponding to the postulated address;

determining cumulative energy results corresponding to respective different memory address postulates; and

finding a largest of the cumulative energy results, thereby determining a most likely location of a designated slot in the received signal sequence.

33. The method of claim 32, wherein the pattern is a regular pattern in which designated time slots are separated by a fixed number of TDMA frames.

34. The method of claim 32, wherein the pattern is a pattern in which designated time slots are separated by a variable number of TDMA frames to reduce ambiguity in determining the most likely location of a designated slot.

35. The method of claim 32, wherein the step of determining memory addresses of other slot energy measurements includes a step of compensating for timing drift over the duration of the plurality of TDMA frames.

36. The method of claim 35, wherein the compensating step includes a step of adding a slot energy measurement to a previous cumulative energy measurement to obtain a new cumulative energy measurement, the previous cumulative energy measurement being the largest of a plurality of adjacent previous cumulative energy measurements.

37. A method of providing fine time synchronization and frequency synchronization to time division multiple access (TDMA) transmissions from a communications satellite to a receiving terminal, comprising:

transmitting from the satellite a succession of superframes, each superframe comprising a succession of TDMA frames and each TDMA frame comprising a succession of time slots, wherein a sequence of signals occupies sequential time slots in each TDMA frame;

designating at least one of the time slots in each superframe for transmitting a synchronization signal comprising a known symbol pattern;

identifying at the receiving terminal a first occurrence of the designated time slot with a coarse accuracy of better than half a time slot, sampling a received transmission during the first occurrence, and storing the samples in a memory;

postulating a frequency error and compensating the stored samples for the postulated frequency error;

determining correlations of the known symbol pattern with respective different time shifts of the compensated stored samples, the time shifts comprising stored samples selected according to a range from an earliest possible occurrence of the known symbol pattern to a latest possible occurrence of the known symbol pattern;

determining correlations of the known symbol pattern with the different time shifts for different frequency error postulates in steps from a lowest frequency or most negative error to a highest frequency or most positive error;

storing the correlations in a two-dimensional table, one table dimension representing the different time shifts and the other table dimension representing the different frequency error postulates;

accumulating correlations from successive occurrences of the designated time slot by adding correlations corresponding to the same frequency error postulate and time shift, and producing a cumulative correlation table; and

identifying a largest cumulative correlation in the cumulative correlation table, thereby providing frequency and time synchronization.

38. The method of claim 37, wherein the accumulating step includes compensating for possible timing drift over an accumulation period by adding a latest correlation from a latest occurrence of the designated time slot to a largest of a plurality of previous cumulative correlations corresponding to adjacent time shifts and corresponding to the same frequency error postulate as the latest correlation in order to obtain a new cumulative correlation.

39. The method of claim 37, wherein the accumulating step includes compensating for possible frequency drift over an accumulation period by adding a latest correlation from a latest occurrence of the designated time slot to a largest of a plurality of previous cumulative correlations corresponding to adjacent frequency error postulates and corresponding to the same time shift as the latest correlation in order to obtain a new cumulative correlation.

40. The method of claim 37, wherein the accumulating step includes compensating for possible frequency and timing drift over an accumulation period by adding a latest correlation from a latest occurrence of the designated time slot to a largest of a plurality of previous cumulative correlations corresponding to adjacent frequency error postulates and time shifts in order to obtain a new compensated cumulative correlation.

41. A method of time synchronizing a receiver with a signal characteristic broadcast repetitively by a transmitter at known repetition intervals, comprising the steps of:

storing samples of the signal characteristic in an addressable memory in the receiver;

making postulates of an address region in the addressable memory identifying samples of a first occurrence of the signal characteristic;

determining a succession of measurement values related to the signal characteristic using respective groups of samples from the postulated address region by incrementing a start address until all samples have been used;

storing the succession of measurement values at successive memory addresses in the addressable memory;

computing first candidate new values respectively associated with each of the stored measurement values by selecting each stored measurement value in turn from sequential addresses and combining each selected stored measurement value with a measurement value or an available associated new value computed from an earlier address corresponding to a nominal estimate of the repetition interval;

computing second candidate new values respectively associated with each of the stored measurement values by selecting each stored measurement value in turn from sequential addresses and combining each selected stored measurement value with a measurement value or an available associated new value computed from an earlier address corresponding to the nominal estimate plus an uncertainty in the repetition interval and with a new value computed from an earlier address corresponding to twice the nominal estimate plus the uncertainty in the repetition interval;

computing third candidate new values respectively associated with each of the stored measurement values by selecting each stored measurement value in turn from sequential addresses and combining each selected stored measurement value with a measurement value or an available associated new value computed from an earlier address corresponding to the nominal estimate minus the uncertainty and with a new value computed from an earlier address corresponding to twice the nominal estimate minus the uncertainty;

selecting best ones of the first, second, and third candidate new values as new values associated with each respective measurement value, whereby new values are associated with all of the measurement values;

determining the best of the new values, thereby time synchronizing the receiver with the signal characteristic.

42. A method of time synchronizing a receiver with a signal characteristic broadcast repetitively by a transmitter at known repetition intervals, comprising the steps of:

storing samples of the broadcast signal characteristic in an addressable memory;

making postulates of a region of addresses in the addressable memory where samples corresponding to a first occurrence of the signal characteristic are likely to start;

determining a measurement value related to the signal characteristic using samples stored in the memory at addresses in the postulated region, beginning with a start address, and determining successive measurement values by incrementing the start address;

storing the measurement values in an addressable memory at successive memory addresses;

initializing a plurality of counters respectively associated with the stored measurement values;

computing a first candidate new value associated with each of the measurement values selected in turn from sequential addresses by combining the selected measurement value with a measurement value or an available associated candidate new value stored at an earlier address that corresponds to a nominal estimate of the repetition interval and associating the counter associated with the earlier address with the first candidate new value;

computing a second candidate new value associated with each of the measurement values selected in turn from sequential addresses by combining the selected measurement value with a measurement value or an available associated candidate new value stored at an earlier address that corresponds to the nominal estimate plus an uncertainty in the repetition interval and with a new value computed from an earlier address corresponding to twice the nominal estimate plus the uncertainty and associating the counter associated with the latter new value with the second candidate new value;

computing a third candidate new value associated with each of the measurement values selected in turn from sequential addresses by combining the selected measurement value with a measurement value or an available associated candidate new value stored at an earlier address corresponding to the nominal estimate minus the uncertainty and with a new value computed from an earlier address corresponding to twice the nominal estimate minus the uncertainty and associating the counter associated with the latter new value with the third candidate new value;

selecting the best of the first, second, and third candidate values as a new value associated with the selected measurement value;

transferring the counter value associated with the selected best candidate value to the counter value associated with the new value after incrementing the counter value by one when the third candidate value was selected as the best value and decrementing the counter value by one when the second candidate value was selected as the best value;

repeating the computing, selecting, and transferring steps until all stored measurement values have been selected and new values and counter values have been associated with each;

determining the best of the new values, thereby time synchronizing with the signal characteristic, and using the associated counter value to obtain a frequency error estimate.

43. The method of claim 42, wherein the measurement values are signal strength values.

44. The method of claim 42, wherein the measurement values are determined by correlating samples with a known signal pattern.

45. The method of claim 44, wherein the known signal pattern is a digital code.

46. The method of claim 45, wherein the digital code is one of a set of orthogonal digital codes.

47. The method of claim 46, wherein the measurement values are determined by correlating samples with every digital code in the set.

48. The method of claim 47, wherein samples are correlated with digital codes by determining a Fast Walsh Transform.

49. The method of claim 42, wherein the combining steps comprise the step of adding moduli of combined values.

50. A method of facilitating coarse timing synchronization of a receiving station with a signal transmitted by a transmitting station, comprising the steps of:

transmitting a time division multiplex (TDMA) signal comprising a succession of TDMA frames, wherein each TDMA frame includes a first number of TDMA timeslots, at least one TDMA timeslot in each frame includes broadcast information, and the TDMA timeslots in in successive frames form a repetitive pattern, the repetitive pattern repeating with a period of a second number of frames;

varying a power level at which certain ones of the TDMA timeslots including broadcast information are transmitted in the repetitive pattern such that corresponding TDMA timeslots in successive repeats of the repetitive pattern are transmitted at the same power level;

measuring in the receiving station a received strength of the TDMA signal over successive timeslots;

accumulating a current TDMA signal strength measurement with TDMA signal strength measurements made at corresponding time positions in previous repeats of the repetitive pattern, thereby forming accumulated values; and

determining a highest accumulated value, the highest accumulated value indicating coarse timing synchronization.

51. The method of claim 50, wherein the accumulating step comprises the step of determining a corresponding time position in a previous repeat of the repetitive pattern based on a largest of three previous accumulated values.

52. The method of claim 51, wherein the three previous accumulated values correspond to neighboring time positions, whereby time drift over an accumulation period is accounted for.

53. The method of claim 50, wherein the step of determining a highest accumulated value further comprises the step of combining selected accumulated values having predetermined time spacings that are equal to known time spacings at which the TDMA signal is transmitted at a relatively higher power level.

54. A method of facilitating coarse timing synchronization of a receiving station with a signal transmitted by a transmitting station, comprising the steps of:

transmitting from said transmitting station a time division multiplex signal (TDMA signal) comprising repetitive TDMA frame periods each containing a first number of TDMA timeslots, at least one of which slots in every frame contains broadcast information (BCCH slots) and others may or may not contain traffic information, and wherein said BCCH slots in successive frames furthermore form a repetitive pattern over a period of a second number of frames;

varying the transmitted power level of certain of said BCCH slots within said repetitive pattern such that corresponding slots in successive repeats of said pattern use the same power;

forming in said receiving system measurements of received signal strength over successive timeslots;

accumulating said measurements with measurements made at corresponding time positions in previous repeats of said pattern over a second number of pattern repeats to form a block accumulation for each time position;

further accumulating values from a current block accumulation with corresponding values from previous block accumulations to produce extended accumulation values for each time position; and

determining the highest of said extended accumulation values as indicative of coarse timing synchronization.

55. The method of claim 54, wherein the further accumulating step comprises determining that a corresponding value from a previous block accumulation to be accumulated with a value from a current block accumulation is indicated by the largest of three of the extended accumulation values.

56. The method of claim 55, wherein the three previous extended accumulation values correspond to neighboring time positions, whereby time drift over the extended accumulation period is accounted for.

57. The method of claim 54, wherein the step of determining a highest extended accumulation value further comprises combining selected extended accumulated values with predetermined time spacings equal to the known time spacings at which the TDMA signal is transmitted at a high power level.

58. In a receiving system, a method of providing fine synchronization with a signal transmitted by a transmitting system, comprising the steps of:

transmitting from said transmitting station a time division multiplex signal (TDMA signal) comprising repetitive TDMA frame periods each containing a first number of TDMA timeslots, at least one of which slots in every frame contains broadcast information (BCCH slots) and others may or may not contain traffic information, and wherein said BCCH slots in successive frames furthermore form a repetitive pattern over a period of a second number of frames;

varying the format of bits or symbols in certain of said BCCH slots within said repetitive pattern such that corresponding slots in successive repeats of said pattern use the same format, and at least one of said various formats contains a fixed word being a number of bits or symbols always having the same values;

postulating a number of fine symbol timing position and correlating received samples with said fixed word starting at said postulated position to obtain a correlation value for each of said postulated timing positions;

accumulating the magnitudes of said correlation values for corresponding timing positions within different slots that correspond to the same format to produce a cumulative correlation for each postulated fine timing position; and

determining the highest cumulative correlation as indicative of fine timing synchronization.

59. The method of claim 58, wherein the accumulating step comprises determining that a corresponding timing position is indicated by the largest of three of the cumulative correlation values.

60. The method of claim 59, wherein the three cumulative correlation values correspond to neighboring symbol-timing positions, whereby time drift over the accumulation period is accounted for.