JP2009290739A - Moving average arithmetic processing method in clock deviation change - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make the response speed of mapping amount adjustment during clock deviation change fast while keeping detection precision of a sequentially changing clock deviation high. <P>SOLUTION: A clock deviation detection unit 13 detects the clock deviation between an input clock and an output clock a plurality of times in a fixed period, and a moving average arithmetic processing function 13a uses detected clock deviation detected values thereof to compare the latest clock deviation detected value with the clock deviation detected value right before the latest one or the total value of a plurality of clock deviation detected values, performs weighting operation using the latest clock deviation detected value according to the comparison result, and replaces the operation value with the latest clock deviation detected value, and further similarly repeats comparison and operation, and divides the latest clock deviation detected value which is obtained finally by the total number of obtained clock deviation detected values to find a moving average value. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  In the processing of a high-speed transmission signal in an optical transmission system or the like, for example, when mapping client data to a frame format, the clock deviation is maintained while maintaining high detection accuracy of the deviation between the input clock and the output clock necessary for mapping. The present invention relates to a moving average calculation processing method at the time of clock deviation change, which can improve the response speed of the output clock for adjusting the mapping amount at the time of change.

  At present, in an optical transmission system, for example, as described in Patent Document 1, a wavelength division multiplexing system is employed, and a plurality of low-speed digital signals are transmitted in order to economically transmit a digital signal to a destination. Time-division multiplexing is used to form one high-speed digital signal (also referred to as a high-speed transmission signal), and this high-speed transmission signal is transmitted to an optical fiber.

  In order to time-division multiplex a plurality of low-speed digital signals, it is necessary that the frequencies of the respective low-speed digital signals are exactly the same. Therefore, the frequencies of the respective low-speed digital signals are synchronized by a stuff synchronization method or the like. . In the stuff synchronization method, when mapping a low-speed digital signal, which is client data, to a frame format on the receiving side, a stuff process is performed in which a stuff pulse having no information component is inserted in order to synchronize the frequency of each low-speed digital signal. . When the receiver side restores the client data by demapping the frame format signal on the receiving side, destuffing processing is performed to remove stuffing pulses.

When the above mapping is performed on the transmission side, generally, a write clock is generated based on a clock component extracted from a low-speed digital signal, and the digital signal is mapped to a frame format using this write clock. The digital signal written in the buffer memory is read out by a read clock oscillated from an oscillator.
Japanese Patent No. 3529713

  As described above, when mapping a low-speed digital signal to the frame format, the write clock (input clock) extracted from the low-speed digital signal for mapping and the read clock (output clock) output from the oscillator are asynchronous. Therefore, it is necessary to adjust the mapping amount by detecting a deviation (clock deviation) between the input clock and the output clock. In order to perform stable mapping, it is necessary to improve the detection accuracy of the clock deviation, but this can be done by calculating a moving average of a plurality of past clock deviation detection values (moving average calculation processing).

  However, since the total deviation detection period increases as the average number of clock deviations in the past (multiple times) increases, there is a problem that the response speed of the output clock to the input clock becomes slow.

  For example, when the phase of the input clock CK1 changes as shown in the correlation diagram between the JC processing amount that is the data adjustment amount corresponding to the clock deviation and the time shown in FIG. 1, the clock deviation is obtained by moving average calculation processing, When the phase of the output clock CK2 is adjusted so as to follow the phase of the input clock CK1 according to this result, the time (when the output clock CK2 follows the input clock CK1 when the average number of operations in the moving average calculation process is large ( Response time) is delayed.

  In order to solve the above-described problems, an object of the present invention is to increase the response speed of the output clock to the input clock when the clock deviation changes, while maintaining high detection accuracy of the clock deviation that changes sequentially.

  In order to achieve the above object, the inventors have written data extracted from a digital signal used for mapping to a frame format by writing and reading a digital signal in a buffer memory or performing demapping separated from the frame format. The deviation between the input clock and the output clock input from the outside is detected multiple times in a certain period, and the moving average calculation processing of the clock deviation detection value sequentially obtained by this detection is performed, and the clock deviation detection value detected most recently The latest clock deviation detection value is compared with the total value that is the sum of one or more clock deviation detection values detected before the latest detection, and when this result is small, equal, or large In each case, the weighting calculation using the latest clock deviation detection value is performed, and the value obtained by this calculation is updated to the latest. After replacing the lock deviation detection value, repeating the comparison and calculation described above, the latest clock deviation detection value obtained at the end, it was decided to carry out is divided by the detection number of all clock deviation detection value.

  Specifically, an input clock is generated based on a clock component extracted from a digital signal whose frequency changes, and the digital signal is written to a buffer memory in order to map the digital signal to a frame format using this input clock. When reading out a digital signal with an externally input output clock, the deviation between the input clock and the output clock is detected a plurality of times in a certain period, and the moving average calculation processing of the clock deviation detection values sequentially obtained by this detection is performed. In the moving average calculation processing method at the time of clock deviation change to be performed, the moving average calculation processing includes the latest clock deviation detection value that is the latest clock deviation detection value among the clock deviation detection values sequentially obtained by the detection; Sum of one or more clock deviation detection values detected before the latest detection A comparison is made with a certain sum value, and when the latest clock deviation detection value is smaller than, equal to, or larger than the sum total value, a weighting operation is performed using the latest clock deviation detection value. Then, after replacing the value obtained in this calculation with the latest clock deviation detection value, repeatedly performing the comparison according to the comparison using the latest clock deviation detection value after the replacement and the result of the comparison, The latest clock deviation detection value obtained after the calculation using all the clock deviation detection values detected a plurality of times in the predetermined period is divided by the number of detections of all the clock deviation detection values. It is a moving average calculation processing method at the time of change.

  According to this method, when mapping a digital signal, a moving average value is sequentially obtained by sequentially using a plurality of clock deviation detection values. The detected value is compared with the total value of one or more previous clock deviation detected values, and a weighting operation is performed using the latest clock deviation detected value according to the result, and the calculated value is used as the latest clock deviation value. After replacing with the detected value, the comparison and calculation were repeated in the same manner, and the latest clock deviation detected value obtained at the end was divided by the number of detected clock deviation detected values to obtain the moving average value. For this reason, a moving average value closer to the finally obtained average value (the average value of the latest clock deviation detection value obtained at the end) can be obtained in a time process earlier than the conventional method. The moving average value obtained in the middle of this process can be processed to bring the clock deviation closer to zero in a shorter time than in the past. In other words, the response of the output clock to the input clock when the clock deviation changes for a certain period can be accelerated.

  More specifically, an input clock is generated based on the clock component extracted from the digital signal in the frame format, and the frame format digital signal is written to the buffer memory for demapping with the input clock, When reading out the written digital signal in frame format with the output clock input from the outside, the deviation between the input clock and the output clock is detected multiple times in a certain period, and the clock deviation detection obtained sequentially by this detection In the moving average calculation processing method at the time of clock deviation change for performing a moving average calculation process of values, the moving average calculation process is a clock deviation detection value detected most recently among clock deviation detection values sequentially obtained by the detection. The latest clock deviation detection value and one or more detected before the latest detection The latest clock deviation detection value is compared with the sum total value that is the sum of the lock deviation detection values, and when the latest clock deviation detection value is smaller than, equal to, or larger than the total value, After performing the weighting calculation used and replacing the value obtained by this calculation with the latest clock deviation detection value, the comparison using the newest clock deviation detection value after the replacement and the result according to the result of the comparison Repeat the calculation, and divide the latest clock deviation detection value obtained after the calculation using all the clock deviation detection values detected a plurality of times in the certain period by the number of detections of all the clock deviation detection values. This is a moving average calculation processing method when the clock deviation changes.

  According to this method, when digital signal demapping is performed, a moving average value is sequentially obtained by sequentially using a plurality of clock deviation detection values. The deviation detection value is compared with the sum of the previous one or more clock deviation detection values, and a weighting operation is performed using the latest clock deviation detection value according to the result, and the calculated value is used as the latest clock. After the replacement with the deviation detection value, the comparison and calculation were repeated in the same manner, and the latest clock deviation detection value obtained at the end was divided by the number of detections of all the clock deviation detection values to obtain the moving average value. For this reason, a moving average value closer to the finally obtained average value (the average value of the latest clock deviation detection value obtained at the end) can be obtained in a time process earlier than the conventional method. The moving average value obtained in the middle of this process can be processed to bring the clock deviation closer to zero in a shorter time than in the past. In other words, the response of the output clock to the input clock when the clock deviation changes for a certain period can be accelerated.

The moving average calculation processing method at the time of clock deviation change according to the present invention uses the latest clock deviation detection value, which is the latest clock deviation detection value among the clock deviation detection values sequentially obtained by the detection described above, as T _n (n is 1 or an integer), 1 or more of the plurality detected before the latest detection is 2n, and the total value is (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ) The latest clock deviation detection value T _n is compared with the total value (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), and the latest clock deviation detection value T _n is compared with the total value (S ₂ ^{( n-1)} ₊₁ +... + S ₂ ⁿ ) is smaller than T _{n + 1} = T _n × 2 + 1, the latest clock deviation detection value T _n is the sum (S ₂ ^(n-1) ₊₁ + ... + S ₂ ⁿ ), T _{n + 1} = T _{When n} × 2 is calculated and the latest clock deviation detection value T _n is larger than the total value (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), T _{n + 1} = T _n × 2-1 performs calculation, according to the value T _{n + 1} obtained by this calculation to the after replacing the latest clock deviation detection value T _n, the comparison and the comparison with the most recent clock deviation detection value T _n after the replacement results The above-mentioned calculation is repeated, and the latest clock deviation detection value T _{n + 1} obtained after the calculation using all the clock deviation detection values detected a plurality of times in the predetermined period is detected as the detection of all the clock deviation detection values. It is desirable to divide by the number ²ⁿ .

According to this method, in the process of sequentially obtaining a moving average value using a plurality of clock deviation detection values sequentially, the latest clock deviation detection value _Tn and the previous clock deviation detection value Tn are not sequentially performed as in the prior art. One or more sum values (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ) of a plurality of clock deviation detection values are compared, and a weighting operation using the latest clock deviation detection value is performed according to this result. After this calculated value is replaced with the latest clock deviation detection value, the comparison and calculation are repeated in the same manner, and the latest clock deviation detection value T _{n + 1} obtained at the end is divided by the detected number 2 ⁿ of all the clock deviation detection values. The moving average value was calculated. For this reason, the moving average value (5/8) closer to the finally obtained average value (for example, 8/8) is obtained for the fifth time in the conventional method, but more than the second time in the method of the present invention. It can be obtained in the early time process. The moving average value (5/8) obtained in the middle of this process can be processed to bring the clock deviation closer to zero in a shorter time than in the past. In other words, the response of the output clock to the input clock when the clock deviation changes for a certain period can be accelerated.

  More specifically, the stuff control method inserts stuff bits according to the moving average value obtained by the moving average calculation processing method at the time of clock deviation change according to claim 1 or 3.

  According to this method, stuff bits are inserted using a moving average value obtained in a faster calculation process than the conventional method, so that the response speed of stuff processing can be increased.

  Furthermore, specifically, when performing the demapping in the moving average calculation processing method at the time of clock deviation change according to claim 2 or 3, destuffing processing for removing stuff bits from the digital signal was performed. In this case, the destuffing control method performs moving average calculation processing by the moving average calculation processing method using a clock deviation detection value with respect to the output clock corresponding to the fluctuation of the input clock caused by the destuffing processing.

  According to this method, when a deviation from the output clock due to fluctuations in the input clock when the destuffing process is executed is detected, the moving average value is obtained in a calculation process faster than the conventional method.

  According to the present invention, the moving average calculation processing method at the time of clock deviation change that can increase the response speed of the output clock to the input clock at the time of clock deviation change while maintaining high detection accuracy of the clock deviation that changes sequentially. Can be provided.

  Embodiments of the present invention will be described with reference to the accompanying drawings. The embodiments described below are examples of the present invention, and the present invention is not limited to the following embodiments. In the present specification and drawings, the same reference numerals denote the same components.

(Embodiment)
FIG. 2 is a block diagram showing a configuration of a signal mapping processing apparatus to which the moving average calculation processing method at the time of clock deviation change according to the embodiment of the present invention is applied.

  The signal mapping processing apparatus 10 shown in FIG. 2 includes a buffer memory 11, a clock extraction unit 12, and a clock deviation detection unit 13 having a moving average calculation processing function 13a. Although the clock is extracted from the input digital signal by the clock extraction unit 12 in FIG. 2, it may be an external clock input.

  The buffer memory 11 writes and stores the input digital signal with the input clock CK1 in order to convert the input digital signal into a frame format digital signal, and stores the stored digital signal with the output clock CK3 as an output digital signal. Device.

  The clock extraction unit 12 extracts a clock component from an input digital signal whose frequency changes, generates an input clock CK1 based on the extracted clock component, and uses the input clock CK1 as a buffer memory 11 and a clock deviation detection unit. 13 is output.

The clock deviation detector 13 includes a ^2n- stage shift register 17 shown in FIG. 3, detects a clock deviation a plurality of times in a fixed period, and sequentially holds the detected clock deviation detection values while moving average calculation. The data is output to the processing function 13a. However, S ₂ ⁿ of the clock deviation detection values held in the shift register 17 is the oldest detection value within a certain period, and is held in the 2 ^nth stage of the shift register 17. In FIG. 3, numerical values are shown as specific examples up to S ₆ .

The moving average calculation processing function 13a performs moving average calculation processing as described below. First, as shown in step F1 in FIG. 3, the clock deviation detection values S ₁ detected in the latest of a clock deviation detection value obtained sequentially from the shift register 17 with the latest clock deviation detection value T _1. Next, the sum total value (S ₂ ⁽ⁿ ₂ )) is the sum of the latest clock deviation detection value T ₁ and one or more 2n (n is an integer of 1 or more) clock deviation detection values detected before the latest detection. ^-1) ₊₁ +... + S ₂ ⁿ ).

Here, as shown in step F2, the latest clock deviation detection value T _1, and compares the detected one clock deviation detection value S ₂ of the sum value prior to the latest detection (S _2). As a result of the comparison, when the latest clock deviation detection value T ₁ is smaller than the total value S ₂ , that is, when T ₁ <S ₂ , the calculation of T ₂ = T ₁ × 2 + 1 is performed. When the latest clock deviation detection value T ₁ is equal to the total value S _2, that is, when T ₁ = S ₂ , the calculation of T ₂ = T ₁ × 2 is performed. Further, when the latest clock deviation detection value T ₁ is larger than the total value S ₂ , that is, when T ₁ > S ₂ , the calculation of T ₂ = T ₁ × 2-1 is performed.

Then, replace the value T ₂ obtained in any of the operations of the above steps F2 to date clock deviation detection value. Thereafter, as shown in step F3, the latest clock deviation detection value T _2, the two detected total value of the clock deviation detection value S _3, S ₄ before this latest detection and _(S 3 ₊ S ₄₎ When T ₂ <(S ₃ + S ₄ ), T ₃ = T ₂ × 2 + 1, and when T ₂ = (S ₃ + S ₄ ), T ₃ = T ₂ × 2 Further, when T ₂ > (S ₃ + S ₄ ), the calculation of T ₃ = T ₂ × 2-1 is performed.

Similarly, as shown in step F4, the latest clock deviation detection value _Tn and the sum of the plurality of clock deviation detection values detected before this latest detection (S ₂ ^(n-1) ₊₁ +... + S ₂ ⁿ ), and if T _n <(S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), the calculation of T _{n + 1} = T _n × 2 + 1, and T _n = (S ₂ ^{(n − 1) In the case of} ₊₁ +... + S ₂ ⁿ ), T _{n + 1} = T _n × 2, and in the case of T _n > (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), T _{n + 1} = T _n × 2-1 calculation is performed.

Thus, the comparison and calculation are repeated, and as shown in step F5, the latest clock deviation detection value T _{n + 1} obtained after the calculation using all the clock deviation detection values detected a plurality of times in a certain period is set to all the clocks. The average value of the detected clock deviation values is obtained by dividing the detected number of deviation detection values by ²ⁿ .

Next, the operation of the signal mapping processing apparatus 10 having such a configuration will be described. However, as shown in FIG. 4, the clock deviation detection unit 13 includes an eight-stage shift register 17A, detects the clock deviation a plurality of times in a certain period, and detects the detected clock deviation value S ₁ ... S. ₈ sequentially hold and is adapted to output to the moving average calculation processing function 13a while.

  An input digital signal whose frequency is changed is written in the buffer memory 11 by the input clock CK1 output from the clock extraction unit 12, and the written digital signal is read out by the output clock CK3 input from the outside to be framed. Output digital signal in the format.

  In this mapping process, the clock deviation between the frequency of the input clock CK1 and the frequency of the output clock CK3 is detected, and the moving average calculation process at the time of the clock deviation change is performed, so that the response speed of the mapping amount adjustment is higher than the conventional one. Can also be improved.

Changes in the clock deviation are sequentially detected eight times by the clock deviation detector 13 in a certain period, and these eight clock deviation detection values S ₁ ... S ₈ are sequentially held in the shift register 17A and output to the moving average arithmetic processing function 13a. Is done. Here, it is assumed that all of the eight clock deviation detection values S ₁ ... S ₈ change from “0” to “1” as shown in FIG.

First, as shown in (b) lines from (a) line in FIG. 5, the clock deviation detection values S ₁ holding value of the shift register 17A is assumed to have changed from "0" to "1".
In the moving average calculation processing by the moving average calculation processing function 13a, first, in step F11 of FIG. 4, the clock deviation detection value S detected most recently among the clock deviation detection values S ₁ ... S ₈ sequentially obtained from the shift register 17A. ₁ = “1” is the latest clock deviation detection value T ₁ = “1”.

Next, in step F12, the latest clock deviation detection value T ₁ = “1” and the total value (S ₂ = “0”) of one clock deviation detection value S ₂ detected before the latest detection are obtained. To be compared. Since this comparison result is T ₁ > S _2, T ₂ = T ₁ × 2-1 is calculated, and T ₂ = “1” is obtained as shown in FIG.

Next, in step F13, the value T ₂ = “1” obtained by the calculation in step F12 is replaced with the latest clock deviation detection value. Thereafter, the latest clock deviation detection value T ₂ = “1” and the sum of the two clock deviation detection values S ₃ = “0” and S ₄ = “0” detected before this latest detection (S ₃ + S ₄ = “0”). Since this comparison result is T ₂ > (S ₃ + S ₄ ), the calculation of T ₃ = T ₂ × 2-1 is performed, and T ₃ = “1” is obtained.

Next, in Step F14, the value T ₃ = “1” obtained by the calculation in Step F13 is replaced with the latest clock deviation detection value, and thereafter, the latest clock deviation detection value T ₃ = “1”, Four clock deviation detection values S ₅ = “0”, S ₆ = “0”, S ₇ = “0”, S ₈ = “0” detected before the latest detection (S ₅ + S ₆ + S ₇ + S ₈ = “0”). Since this comparison result is T ₃ > (S ₅ + S ₆ + S ₇ + S ₈ ), an operation of T ₄ = T ₃ × 2-1 is performed, and T ₄ = “1” is obtained.

Since the calculation using all the clock deviation detection values S ₁ ... S ₈ detected a plurality of times in a certain period in steps F11 to F14 is performed, in step F15, the latest clock deviation detection obtained after the calculation is performed. the value _{T 4} = "1", is divided by all the detection number ^{2 n} = "8" clock deviation detection value _S 1 ... _{S 8,} whereby, first of all the clock deviation detection value _S 1 ... _{S 8} Moving average value = “1/8”.

Next, assume that a change to "1" (b) as shown in (c) from one line, the clock deviation detection value S ₂ of the value held in the shift register 17A from "0".
In this case, in step F11, the latest detected clock deviation value S ₁ = “1” among the detected clock deviation values S ₁ ... S ₈ sequentially obtained from the shift register 17A is the latest detected clock deviation value T _1. = “1”.

Next, in step F12, the latest clock deviation detection value T ₁ = “1” and the total value (S ₂ = “1”) of one clock deviation detection value S ₂ detected before the latest detection are obtained. To be compared. The result of this _comparison, computation of T 1 = _{S 2} since _T 2 = _{T 1} × 2 is performed, _{T 2} = "2" is obtained.

Next, in step F13, the value T ₂ = “2” obtained in the calculation of step F12 is replaced with the latest clock deviation detection value, and then the latest clock deviation detection value T ₂ = “2” The two clock deviation detection values S ₃ = “0” and the sum of S ₄ = “0” (S ₃ + S ₄ = “0”) detected before the detection of the above are compared. Since the comparison result is T ₂ > (S ₃ + S ₄ ), the calculation of T ₃ = T ₂ × 2-1 is performed, and T ₃ = “3” is obtained.

Next, in step F14, after the value T ₃ = “3” obtained by the calculation in step F13 is replaced with the latest clock deviation detection value, the latest clock deviation detection value T ₃ = “3” 4 clock deviation detection values S ₅ = “0”, S ₆ = “0”, S ₇ = “0”, S ₈ = “0” detected before the detection of (S ₅ + S ₆ + S ₇ + S ₈ = “0”). Since this comparison result is T ₃ > (S ₅ + S ₆ + S ₇ + S ₈ ), T ₄ = T ₃ × 2-1 is calculated, and T ₄ = “5” is obtained.

Since the calculation using all the clock deviation detection values S ₁ ... S ₈ detected a plurality of times in a certain period in steps F11 to F14 is performed, in step F15, the latest clock deviation detection obtained after the calculation is performed. the value _{T 4} = "5" is divided by all the detection number ^{2 n} = "8" clock deviation detection value _S 1 ... _{S 8,} thereby, the second all the clock deviation detection value _S 1 ... _{S 8} The moving average value of “5/8” is obtained.

Next, the clock deviation detection value S ₃ shown in (d) of the line (c) to the line from "0" in the case of changes to "1" also, the third of every clock deviation detection value S ₁ ... S ₈ The moving average value = “5/8” is obtained.

Next, when the clock deviation detection value S ₄ changes from “0” to “1” as shown in the (d) line to the (e) line, the latest detected clock deviation is detected in the same manner as described above in Step F11. The value S ₁ = “1” is set to the latest clock deviation detection value T ₁ = “1”.

In step F12, the latest clock deviation detection value T ₁ = “1” is compared with the total value (S ₂ = “1”) of one clock deviation detection value S ₂ detected before the latest detection, An operation of T ₂ = T ₁ × 2 corresponding to the result of T ₁ = S ₂ is performed, and T ₂ = “2” is obtained.

In step F13, after the value T ₂ = “2” obtained in the calculation of step F12 is replaced with the latest clock deviation detection value, the latest clock deviation detection value T ₂ = “2” is obtained. The two previously detected clock deviation detection values S ₃ = “1” and the total value of S ₄ = “1” (S ₃ + S ₄ = “2”) are compared. Since this comparison result is T ₂ = (S ₃ + S ₄ ), an operation of T ₃ = T ₂ × 2 is performed, and T ₃ = “4” is obtained.

In step F14, after the value T ₃ = “4” obtained by the calculation in step F13 is replaced with the latest clock deviation detection value, the latest clock deviation detection value T ₃ = “4” is obtained. The previously detected four clock deviation detection values S ₅ = “0”, S ₆ = “0”, S ₇ = “0”, and S ₈ = “0” (S ₅ + S ₆ + S ₇ + S ₈ = “0”). Since this comparison result is T ₃ > (S ₅ + S ₆ + S ₇ + S ₈ ), T ₄ = T ₃ × 2-1 is calculated, and T ₄ = “7” is obtained.

In step F15, the latest clock deviation detection value T ₄ = “7” obtained at the end is divided by the number of detections 2 ⁿ = “8” of all the clock deviation detection values S ₁ ... S _8. The fourth moving average value of clock deviation detection values S ₁ ... S ₈ = “7/8” is obtained.

Next, even when the clock deviation detection values S ₅ , S _6, and S ₇ change from “0” to “1” as shown in lines (e) to (h), all the clock deviation detection values S are detected. ₁ ... The moving average value of S _{8 for} the fifth to seventh times = “7/8” is obtained.

Finally, if the clock deviation detection value S ₈ as shown (h) from the line in the (i) line is changed from "0" to "1", in step F11～F13, in the same manner as described above for the latest clock deviation detection value T ₁ = “1”, T ₂ = “2” and T ₃ = “4” are obtained.

In step F14, after the value T ₃ = “4” obtained above is replaced with the latest clock deviation detection value, the latest clock deviation detection value T ₃ = “4” is detected before this latest detection. Furthermore, the sum of the four clock deviation detection values S ₅ = “1”, S ₆ = “1”, S ₇ = “1”, S ₈ = “1” (S ₅ + S ₆ + S ₇ + S ₈ = “4”) ) Is compared. Since this comparison result is T ₃ = (S ₅ + S ₆ + S ₇ + S ₈ ), T ₄ = T ₃ × 2 is calculated, and T ₄ = “8” is obtained.

In step F15, the latest clock deviation detection value T ₄ = “8” obtained at the end is divided by the number of detections 2 ⁿ = “8” of all clock deviation detection values S ₁ ... S _8. The clock deviation detected value S ₁ ... S ₈ of the eighth moving average value = “8/8” is obtained.

  Thus, according to the signal mapping processing apparatus 10 to which the moving average calculation processing method at the time of clock deviation change according to the present embodiment is applied, the following effects can be obtained.

For example, if the clock deviation is detected, for example, eight times in a certain period, and the detected clock deviation detected values S ₁ ... S ₈ are all changed from “0” to “1”, the conventional moving average calculation processing method may be used. For example, the average value sequentially obtained by the calculation is “0/8”, “1/8”, “2/8”, “3/8”, “4/8”, “5/8”, “6/8” ”,“ 7/8 ”, and“ 8/8 ”.

  However, in the moving average calculation processing by the moving average calculation processing function 13a of the present embodiment, the average values sequentially obtained by calculation are “0/8”, “1/8”, “5/8”, “5/8”. , “7/8”, “7/8”, “7/8”, “7/8”, “8/8”.

  As can be seen from this result, in the moving average calculation processing method of the present embodiment, the calculation time until the average value becomes “1” is not different from the conventional method, but the average value is changed from “0/8” in the conventional method. The second calculation time until the change to “2/8” can be changed to “5/8”, which conventionally takes the calculation time up to the fifth. Therefore, it is possible to speed up the response when the clock deviation changes for a certain period.

  According to the moving average calculation processing method at the time of clock deviation change according to the present embodiment, the response speed of the output clock CK3 to the input clock CK1 at the time of clock deviation change is increased while maintaining the detection accuracy of the sequentially changing clock deviation high. can do. Therefore, it is possible to improve the response speed of the mapping process in which the input digital signal is written in the buffer memory 11 in the frame format with the input clock CK1, and the written digital signal is read with the output clock CK3.

  The application to the mapping process has been described above, but it can also be applied to the demapping process. That is, in demapping, an input clock is generated based on a clock component extracted from a frame format digital signal, and the frame format digital signal is written to the buffer memory with this input clock, and then the written frame format The original digital signal is restored by reading the digital signal with the output clock input from the outside. Also in this case, the moving average calculation processing method at the time of clock deviation change can be applied as described above.

(Example)
Next, an example will be described in which the moving average calculation processing method at the time of clock deviation change of the above-described embodiment is applied to a wavelength division multiplexing (WDM) optical transmission system.
FIG. 6 is a block diagram of a stuff synchronous multiplex conversion device on the transmission side to which the moving average calculation processing method at the time of clock deviation change according to this embodiment, and FIG. 7 shows the moving average calculation processing method at the time of clock deviation change according to this embodiment. 2 is a block configuration diagram of a reception side stuff synchronous demultiplexing device to which is applied. FIG.

  However, the optical transmission system transmits signals in the signal format and SDH (synchronous digital hierarchy) frame format shown in FIG. That is, IP (Internet Protocol) -NW using Point-to-Point-Protoco1 (hereinafter, “Point-to-Point-Protoco1” is abbreviated as “PPP”) or the like in the router 21 having the SDH interface. A packet 22 in the (network) is mapped in correspondence with a virtual container having a frame format of SDH-NW, so that a client payload is formed, and POH (path overhead) is added to the packet payload so that SDH is indicated by reference numeral 22a. Store in frame 23.

  Along with this storage, a pointer 23P indicating the head position of the SDH frame 23 is inserted into the SOH (section overhead) to indicate the frame phase. A plurality of SDH frames 23 configured as described above are wavelength-multiplexed by the WDM unit 24.

  In such an optical transmission system, in order to transmit a digital signal economically to a destination, a plurality of low-speed digital signals are time-division multiplexed to form one high-speed digital signal (also referred to as a high-speed transmission signal), This high-speed transmission signal is transmitted to the optical fiber. In order to time-division multiplex a plurality of low-speed digital signals, it is necessary that the frequencies of the low-speed digital signals are exactly the same. As a method of synchronizing the frequencies of the respective low-speed digital signals, there are a stuff synchronization method and a network synchronization method.

The principle of the staff synchronization method will be described. As shown in FIGS. 9A and 9D, first, each low-speed digital signal having waveform frequencies f _i and f _j is temporarily stored. As shown in FIGS. 9B and 9C, all low-speed digital signals are stored. Reading is performed at a frequency f ₀ slightly higher than the frequencies f _i and f _{j of the} digital signal. Further, if necessary, it corresponds to each frequency difference f ₀ −f _i , f ₀ −f _j, etc., which is the difference between the frequency f ₀ and the frequencies f _i and f _{j of} each low-speed digital signal, and information the stuffing process for adding a stuff pulse 26 is extra pulses that do not have to synchronize the low speed digital signals to a frequency f _0.

  Further, the synchronized low-speed digital signals are time-division multiplexed and transmitted as one high-speed digital signal. At this time, by separately transmitting information regarding the addition of the stuff pulse 26, the stuff pulse 26 is removed by destuffing processing on the receiving side to restore the original low-speed digital signal.

Next, stuff synchronization processing and moving average calculation processing by the stuff synchronization multiple conversion apparatus shown in FIG. 6 will be described.
The stuff synchronous multiple conversion device 30 includes a clock extraction unit 31, a buffer memory 32, a timing generation unit 33, a stuff control unit 34, a multiple conversion unit 35, and a clock deviation detection having the moving average calculation processing function 13 a described above. And a portion 36.

  The clock extraction unit 31 extracts a clock component from the low-speed digital signal and generates a write clock WCK for writing data to the buffer memory 32. The low-speed digital signal data is written into the buffer memory 32 in accordance with the write clock WCK. Low-speed digital signals written in the buffer memory 32 are sequentially read according to the output clock input from the outside. However, the write clock WCK corresponds to the input clock CK1 and the read clock RCK corresponds to the output clock CK3.

In the multiplex converter 35, the read low-speed digital signal and the stuff pulse are combined to generate a synchronization signal, and a plurality of synchronization signals are time-division multiplexed to generate a high-speed digital signal.
Further, in the clock deviation detection unit 36, the phase difference between the write clock WCK and the read clock CK3 is detected a plurality of times in a certain period, and this clock deviation detection value is output to the moving average calculation processing function 13a. The moving average calculation processing function 13a performs the same moving average calculation processing as described above to obtain the moving average value of the write clock WCK and the read clock CK3, and outputs this moving average value to the stuff control unit 34. Is done.

  In this stuff synchronization method, the stuff processing is performed using the moving average value obtained in the calculation process faster than the conventional method by the moving average calculation processing method at the time of clock deviation change of this embodiment, so the response speed of the stuff processing Can be made faster.

Next, destuffing processing by the stuff synchronous demultiplexing device shown in FIG. 7 will be described.
The stuff synchronous demultiplexing device 40 includes a demultiplexing unit 41, a clock extraction unit 42, a buffer memory 43, a destuff control unit 44, a phase comparison unit 46 having a moving average calculation processing function 13a, an LPF 47, and a voltage controlled oscillation. And a PLL circuit 45 composed of a unit 48.

  In the demultiplexing unit 41, the multiplexed signal, which is a high-speed digital signal, is demultiplexed into a plurality of synchronization signals. In the clock extraction unit 42, a clock component is extracted from the synchronization signal, and a write clock WCK to the buffer memory 43 is generated. When a stuff pulse is present in the transmission frame, the write clock WCK is delayed by one bit at the stuff pulse insertion position by the destuff control unit 44, and as a result, the write clock WCK is not output at the stuff pulse insertion position. Therefore, the stuff pulse is not written in the buffer memory 43. That is, the destuffing process in which the stuff pulse is removed is performed.

  This destuffing process causes a gap in the write clock WCK and a phase difference from the read clock RCK. However, a synchronization process is performed in the PLL circuit 45, and the read clock RCK responds to the write clock WCK. That is, when the write clock WCK capped in the destuffing process is input to the phase comparison unit 46, the phase difference between the write clock WCK and the read clock RCK is detected a plurality of times in a certain period, and this clock deviation detection The value is output to the moving average calculation processing function 13a. In the moving average calculation processing function 13a, a moving average calculation process similar to that described above is performed, whereby a moving average value between the write clock WCK and the read clock RCK is obtained in a short time, and the phase difference that is the moving average value is obtained. Is output to the voltage controlled oscillator 48 via the LPF 47.

  By this processing, the read clock RCK having the same frequency as that of the original low-speed digital signal is reproduced, and data is read in the order in which the data is written from the buffer memory 43 in accordance with the read clock RCK, thereby restoring the low-speed digital signal. In this way, time division multiplexed transmission of digital signals can be performed by the stuff synchronization method.

  Next, the principle of the network synchronization method described above will be described. The network synchronization method means that the processing functions of each device are simplified by supplying a common clock in the network to multiple conversion devices, exchanges, terminal devices, etc. in the network, improving the economics and flexibility of the network. It is a scheme to plan. SDH has been standardized as a multiplexing method that realizes synchronization at all transmission speeds up to the gigabit / second range.

  10 and 11 are diagrams showing SDH STM (synchronous transfer module) frames. The STM frame is composed of a section overhead 51 defined for network operation and maintenance, a payload 52 for storing user information, and a pointer 53 that points to the start position 55 of the user information 54 in the payload 52.

  SDH is premised on network synchronization, but is stable even when information is transmitted through multiple telecommunications carriers' networks and the common clock frequency of each telecommunications carrier's network is independent In order to provide such communication quality, a staff synchronization function using a pointer is employed.

  The positive stuff process and the negative stuff process using the pointer will be described with reference to FIGS. 12 and 13 are diagrams showing SDH STM frames similar to those in FIG.

  As shown in FIG. 12, when the frequency of the low-speed digital signal to be multiplexed is slightly lower than the frequency of the payload of the STM frame, a normal stuff process is performed in which a stuff byte 57 is inserted immediately after the pointer byte of the pointer 56.

  On the contrary, as shown in FIG. 13, when the frequency of the low-speed digital signal to be multiplexed is slightly higher than the frequency of the payload of the STM frame, negative stuffing processing for storing user information 59 in the last byte of the pointer 58 is performed. If the frequency of the low-speed digital signal to be multiplexed matches the frequency of the payload of the STM frame, the stuff process is not performed. As described above, frequency synchronization is performed by positive / negative stuff processing and synchronous digital transmission is performed, whereby asynchronous digital signals are time-division multiplexed and transmitted with stable communication quality.

  The clock deviation detection method of the present invention is an optical transmission system that employs a wavelength division multiplexing method or the like, while maintaining high detection accuracy of the deviation between the input clock and the output clock required when mapping client data to the frame format. This can be applied to improving the response speed of the output clock to the input clock when the clock deviation changes.

It is a correlation diagram of JC processing amount and time. It is a block diagram which shows the structure of the signal mapping processing apparatus to which the moving average arithmetic processing method at the time of the clock deviation change by embodiment of this invention is applied. It is a figure for demonstrating the moving average calculating process using the clock deviation detection value hold | maintained at the ²ⁿ stage shift register of a clock deviation detection part. It is a figure for demonstrating the moving average calculating process using the clock deviation detection value hold | maintained at the 8-stage shift register of a clock deviation detection part. It is a figure showing the clock deviation detection value hold | maintained at the 8-stage shift register of a clock deviation detection part, and the moving average value calculated | required by the moving average calculation process using this clock deviation detection value. It is a block diagram which shows the structure of the stuff synchronous multiplex conversion apparatus of the transmission side to which the moving average arithmetic processing method at the time of the clock deviation change of this embodiment is applied. It is a block diagram which shows the structure of the stuff synchronous demultiplexing apparatus by the side of the transmission to which the moving average arithmetic processing method at the time of the clock deviation change of this embodiment is applied. It is a figure which shows the signal format and SDH frame format in an optical transmission system. It is a wave form diagram for demonstrating the principle of a staff synchronization system. It is a 1st figure which shows the STM frame of SDH. It is a 2nd figure which shows the STM frame of SDH. It is a 3rd figure which shows the STM frame of SDH. It is a 4th figure which shows the STM frame of SDH.

Explanation of symbols

10: Signal mapping processor 11, 32, 43: Buffer memory 12: Clock extraction unit 13, 36: Clock deviation detection unit 46: Phase comparison unit 13a: Moving average calculation processing function 15: LPF
16, 48: Voltage controlled oscillators 17, 17A: Shift register 21: Router 22: Packet 22a: SDH client payload 23: SDH frame format 23P, 53, 56, 58: Pointer 24: WDM unit 26: Stuff pulse 30: Staff synchronous demultiplexing devices 31, 42: clock extraction unit 33: timing generation unit 34: staff control unit 35: demultiplexing unit 40: stuff synchronous demultiplexing device 41: demultiplexing unit 44: destuff control unit 47: low-pass Filter 51: Section overhead 52: Payload 54, 59: User information 55: User information head position 57: Stuff byte CK1: Input clock CK2, CK3: Output clock WCK: Write clock RCK: Read clock

Claims (5)

An input clock is generated based on the clock component extracted from the digital signal whose frequency changes, and the digital signal is written to the buffer memory in order to map the digital signal to the frame format with this input clock. When reading with the input output clock, the deviation between the input clock and the output clock is detected multiple times in a certain period, and the moving average calculation processing of the clock deviation detection value obtained sequentially by this detection is performed. In the moving average calculation processing method of
The moving average calculation process is:
Of the clock deviation detection values sequentially obtained by the detection, the latest clock deviation detection value, which is the latest clock deviation detection value, and one or more clock deviation detection values detected before the latest detection. Comparing the sum total value, which is a sum, and calculating the weight using the latest clock deviation detection value according to each of the cases where the latest clock deviation detection value is smaller than, equal to, or greater than the sum total value After replacing the value obtained by this calculation with the latest clock deviation detection value, performing the comparison according to the comparison using the latest clock deviation detection value after the replacement and the result of the comparison. Repeatedly, the latest clock deviation detection value obtained after the calculation using all the clock deviation detection values detected a plurality of times in the certain period is used as all the clock deviations. Moving average calculation processing method when the clock deviation change, characterized by dividing the number of detected error detection value. An input clock is generated based on the clock component extracted from the frame format digital signal, and the frame format digital signal is written to the buffer memory for demapping with the input clock, and then the written frame format When reading out a digital signal with an externally input output clock, the deviation between the input clock and the output clock is detected a plurality of times in a certain period, and the moving average calculation processing of the clock deviation detection values sequentially obtained by this detection is performed. In the moving average calculation processing method at the time of clock deviation change to be performed,
The moving average calculation process is:
Of the clock deviation detection values sequentially obtained by the detection, the latest clock deviation detection value, which is the latest clock deviation detection value, and one or more clock deviation detection values detected before the latest detection. Comparing the sum total value, which is a sum, and calculating the weight using the latest clock deviation detection value according to each of the cases where the latest clock deviation detection value is smaller than, equal to, or greater than the sum total value After replacing the value obtained by this calculation with the latest clock deviation detection value, performing the comparison according to the comparison using the latest clock deviation detection value after the replacement and the result of the comparison. Repeatedly, the latest clock deviation detection value obtained after the calculation using all the clock deviation detection values detected a plurality of times in the certain period is used as all the clock deviations. Moving average calculation processing method when the clock deviation change, characterized by dividing the number of detected error detection value. The latest clock deviation detection value which is the latest clock deviation detection value among the clock deviation detection values sequentially obtained by the detection is defined as T _n (n is an integer of 1 or more), and is detected before the latest detection. In addition, when one or more of the plurality is set to 2n and the total value is (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), the latest clock deviation detection value T _n and the total value (S ₂ ^(n-1) ₊₁ +... + S ₂ ⁿ ), and the latest clock deviation detection value T _n is smaller than the total value (S ₂ ^(n-1) ₊₁ +... + S ₂ ⁿ ), When the calculation of T _{n + 1} = T _n × 2 + 1 is performed and the latest clock deviation detected value T _n is _equal to the total value (S ₂ ⁽ⁿ⁻¹⁾ ₊₁ +... + S ₂ ⁿ ), T _{n + 1} = T _n × for 2 operations, the latest clock deviation detection value T _n previous If the sum value ^{_{(S 2 (n-1)}} +1 + ... + S 2 n) greater _{than, T n} + 1 = performs a computation of _{T n} × 2-1, the latest clock deviation values _{T n + 1} obtained by the calculation after replacing the detected value T _n, repeatedly to perform the operation in accordance with the result of the comparison as well as the comparison with the most recent clock deviation detection value T _n after the replacement, is detected a plurality of times in the predetermined period The latest clock deviation detection value T _{n + 1} obtained after the calculation using all clock deviation detection values is divided by the detection number 2 ⁿ of all the clock deviation detection values. Moving average calculation processing method when clock deviation changes.   The stuff control method which inserts a stuff bit according to the moving average value obtained by the moving average calculation processing method at the time of clock deviation change according to claim 1 or 3.   When performing the demapping in the moving average calculation processing method at the time of clock deviation change according to claim 2 or 3, when the destuffing process is performed to remove stuff bits from the digital signal, the destuffing process is performed. A destuff control method for performing moving average calculation processing by the moving average calculation processing method using a clock deviation detection value with respect to the output clock corresponding to the fluctuation of the input clock caused by the input clock.

Images