JP5407705B2 - Time information acquisition device and radio clock - Google Patents

Description

  The present invention relates to a time information acquisition device that receives a standard time radio wave and acquires the time information, and a radio clock equipped with the time information acquisition device.

  Currently, in Japan, Germany, the United Kingdom, Switzerland, etc., long standard time radio waves are transmitted from transmitting stations. For example, in Japan, standard time radio waves with amplitude modulation of 40 kHz and 60 kHz are transmitted from transmitting stations in Fukushima Prefecture and Saga Prefecture, respectively. The standard time radio wave includes a sequence of codes constituting a time code indicating the year, month, day, hour and minute, and is transmitted in one cycle of 60 seconds. That is, the period of the time code is 60 seconds.

  A timepiece (radio timepiece) capable of receiving a standard time radio wave including such a time code, taking out the time code from the received standard time radio wave, and correcting the time has been put into practical use. The reception circuit of the radio clock accepts the standard time radio wave received by the antenna and demodulates the standard time radio signal amplitude-modulated by a band pass filter (BPF) for extracting only the standard time radio signal, envelope detection, etc. A demodulation circuit and a processing circuit that reads a time code included in the signal demodulated by the demodulation circuit are provided.

  The conventional processing circuit synchronizes at the rising edge of the demodulated signal, and then binarizes at a predetermined sampling period to obtain TCO data of a unit time length (1 second) which is a binary bit string. Further, the processing circuit measures the pulse width of the TCO data (that is, the time of bit “1” or the time of bit “0”), and codes “P”, “0” corresponding to the width. "Or" 1 "is determined, and time information is acquired based on the determined code sequence.

  In a conventional processing circuit, a process of second synchronization processing, minute synchronization processing, code acquisition, and matching determination is performed from the start of reception of standard time radio waves to acquisition of time information. When processing cannot be completed properly in each process, the processing circuit needs to start processing from the beginning. For this reason, processing may have to be performed again and again due to the influence of noise included in the signal, and the time until the time information can be acquired may be significantly increased.

  Second synchronization is to detect the rising of the code that arrives every second among the codes indicated by the TCO data. By repeating the second synchronization, it is possible to detect a portion where the position marker “P0” arranged at the end of the frame and the marker “M” arranged at the beginning of the frame are continuous. This continuous portion arrives every minute (60 seconds). The position of the marker “M” is the data of the first frame in the TCO data. Detecting this is called minute synchronization. Since the beginning of the frame is recognized by the above-mentioned minute synchronization, code acquisition is started thereafter, and after acquiring the data for one frame, the parity bit is checked, and an impossible value (year, month, day, and time cannot actually occur) Value) is determined (consistency determination). For example, minute synchronization finds the beginning of a frame and may take 60 seconds. Of course, it takes several times as long to detect the beginning of a frame over several frames.

  In Patent Document 1, TCO data obtained by binarizing a demodulated signal at a predetermined sampling interval (50 ms) is acquired, and a data group consisting of binary bit strings every second (20 samples) is listed. It becomes. The apparatus disclosed in Patent Document 1 includes this bit string, a binary bit string template representing the code “P: position marker”, a binary bit string template representing the code “1”, and a binary bit string representing the code “0”. Each of the templates is compared with each other to obtain the correlation, and it is determined by the correlation whether the bit string corresponds to the code “P”, “1”, or “0”.

JP 2005-249632 A

  In the technique disclosed in Patent Document 1, TCO data, which is a binary bit string, is acquired and matched with a template. In a state where the electric field strength is weak or a state where a lot of noise is mixed in the demodulated signal, many errors are included in the acquired TCO data. Therefore, it is necessary to finely adjust the filter for removing noise from the demodulated signal and the threshold of the AD converter to improve the quality of TCO data.

  In addition, if the data of the unit time length (1 second) only determines one of the codes “P”, “1”, and “0”, the start of the second, the start of the minute, etc. are based on the determination result. It is necessary to perform the determination process again. Here, if the beginning of the second or the beginning of the minute cannot be found properly, the processing must be performed again.

  The present invention is not affected by the state of electric field strength or signal noise, and can obtain the current time by accurately acquiring the code included in the standard time radio wave while reducing the amount of computation and computation time. It is an object to provide a time information acquisition device and a radio timepiece including the time information acquisition device.

An object of the present invention is to receive a standard time radio wave including a time code representing time information;
The signal including the time code output from the receiving means is sampled at a predetermined sampling period, and the data value at each sample point is either a first value indicating a low level or a second value indicating a high level. Input waveform data generating means for generating input waveform data having a unit time length of 1 or more,
The data value of each sample point takes either the first value or the second value, has the same time length as the input waveform data, and the waveform shape is a predetermined code in the time code Predicted waveform data generating means for generating predicted waveform data including:
Predicted waveform feature section extracting means for acquiring feature values of feature sections having unique and constant data values different from other codes in the predetermined code from the predicted waveform data;
Input waveform feature section extracting means for acquiring a data value of a feature section corresponding to a feature section of the predicted waveform data from the input waveform data;
Multiplying the data value of the feature section of the input waveform data extracted by the input waveform feature section extracting means by the feature value of the feature section of the predicted waveform data extracted by the predicted waveform feature section extracting means A correlation value calculating means for calculating a value and calculating a correlation value between the input waveform data and the predicted waveform data based on the sum of the calculated values;
A correlation value comparing means for comparing a plurality of correlation values calculated by the correlation value calculating means to calculate an optimum value of the correlation values;
This is achieved by a time information acquisition device comprising: control means for acquiring time information represented by the time code based on the optimum value.

  According to the present invention, the current time is obtained by accurately acquiring the code included in the standard time radio wave without being affected by the state of the electric field intensity or the signal noise, and reducing the calculation amount and the calculation time. It is possible to provide a time information acquisition device capable of performing the above and a radio timepiece including the time information acquisition device.

FIG. 1 is a block diagram showing the configuration of the radio timepiece according to the present embodiment. FIG. 2 is a block diagram showing a configuration example of the receiving circuit according to the present embodiment. FIG. 3 is a block diagram showing the configuration of the signal comparison circuit according to the present embodiment. FIG. 4 is a flowchart showing an outline of processing executed in the radio timepiece according to the present embodiment. FIG. 5 is a diagram for explaining the standard of the standard time radio signal. FIGS. 6A to 6C are diagrams illustrating examples of codes according to JJY, WWVB, and MSF standards, respectively. FIG. 7 is a diagram illustrating a portion of predicted waveform data used in the second synchronization according to the present embodiment. FIG. 8 is a flowchart showing in more detail the detection of the second pulse position (second synchronization) according to the present embodiment. FIG. 9 is a diagram schematically illustrating the second pulse position detection process according to the present embodiment. FIG. 10 is a diagram for explaining an example of calculating a covariance value according to the present embodiment. FIG. 11 is a flowchart showing in more detail the detection of minute head positions (minute synchronization) according to the present embodiment. FIG. 12 is a diagram for explaining input waveform data and predicted waveform data in the detection of the leading position according to the present embodiment. FIGS. 13A and 13B are diagrams for explaining the feature sections in each code of JJY. FIG. 14 is a flowchart showing an example of data value accumulation processing according to the present embodiment. FIG. 15 is a flowchart showing in more detail a fractional decoding process according to the present embodiment. FIG. 16 is a diagram schematically illustrating a decoding process according to the present embodiment. FIG. 17 is a flowchart showing in more detail the tens place detection process according to the present embodiment. FIGS. 18A to 18C are diagrams for explaining covariance values between input waveform data and predicted waveform data. FIG. 19 is a diagram for explaining a feature section in each code of WWVB. FIG. 20 is a diagram for explaining a feature section in each code of MSF. FIG. 21 is a flowchart showing in more detail the detection of the second pulse position (second synchronization) according to the second embodiment of the present invention. FIG. 22 is a diagram schematically illustrating second pulse position detection processing according to the second embodiment. FIG. 23 is a flowchart showing an example of a fractional detection process according to the third embodiment of the present invention.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. In the embodiment of the present invention, a standard time radio wave in a long wave band is received, the signal is detected, a sequence of codes indicating a time code included in the signal is extracted, and based on the sequence of the codes The radio timepiece for correcting the time is provided with the time information acquisition device according to the present invention.

  Currently, standard time radio waves are transmitted from a predetermined transmitting station in Japan, Germany, the United Kingdom, Switzerland, and the like. For example, in Japan, standard time radio waves with amplitude modulation of 40 kHz and 60 kHz are transmitted from transmitting stations in Fukushima Prefecture and Saga Prefecture, respectively. The standard time radio wave includes a string of codes constituting a time code indicating the year, month, day, hour and minute, and is transmitted in one cycle of 60 seconds. Since one code has a unit time length (1 second), one cycle can include 60 codes.

  FIG. 1 is a block diagram showing the configuration of the radio timepiece according to the present embodiment. As shown in FIG. 1, the radio timepiece 10 includes a CPU 11, an input unit 12, a display unit 13, a ROM 14, a RAM 15, a receiving circuit 16, an internal timing circuit 17, and a signal comparison circuit 18.

  The CPU 11 reads out a program stored in the ROM 14 at a predetermined timing or in response to an operation signal input from the input unit 12, expands the program in the RAM 15, and configures the radio clock 10 based on the program. Execute instructions and data transfer. Specifically, for example, by receiving the standard time radio wave by controlling the receiving circuit 16 every predetermined time, a sequence of codes included in the standard time radio signal is obtained from digital data based on the signal obtained from the receiving circuit 16. Specific processing is performed to correct the current time measured by the internal clock circuit 17 based on this sequence of codes, and to transfer the current time clocked by the internal clock circuit 17 to the display unit 13. In the present embodiment, one or more unit time lengths of predicted waveform data in a predetermined form are generated, and the predicted waveform data is compared with the input waveform data obtained from the standard time radio wave received by the receiving circuit. By doing so, the beginning of the second is specified.

  Further, in the present embodiment, predicted waveform data including a predetermined code having one or more unit time lengths is generated, and the beginning of the minute is specified. Furthermore, in the present embodiment, by using an operation value obtained by multiplying the accumulated value of the data value of a predetermined section (feature section) of the input waveform data and the multiplication value based on the predicted waveform data, The value of various codes including the date is specified. By specifying the date and time, the error in the internal clock circuit 17 is calculated, and the current time in the internal clock circuit 17 can be corrected.

  The input unit 12 includes a switch for instructing execution of various functions of the radio timepiece 10, and outputs a corresponding operation signal to the CPU 11 when the switch is operated. The display unit 13 includes a dial, an analog pointer mechanism controlled by the CPU 11, and a liquid crystal panel, and displays the current time measured by the internal clock circuit 17. The ROM 14 stores a system program, an application program, and the like for operating the radio timepiece 10 and realizing a predetermined function. The program for realizing the predetermined function includes a program for controlling the signal comparison circuit 18 for second pulse position detection processing, minute leading position detection processing, and code decoding processing, which will be described later. The RAM 15 is used as a work area for the CPU 11 and temporarily stores programs and data read from the ROM 14, data processed by the CPU 11, and the like.

  The reception circuit 16 includes an antenna circuit, a detection circuit, and the like, obtains a signal demodulated from the standard time radio wave received by the antenna circuit, and outputs the signal to the signal comparison circuit 18. The internal clock circuit 17 includes an oscillation circuit, counts clock signals output from the oscillation circuit, counts the current time, and outputs current time data to the CPU 11.

  FIG. 2 is a block diagram showing a configuration example of the receiving circuit according to the present embodiment. As shown in FIG. 2, the receiving circuit 16 includes an antenna circuit 50 that receives standard time radio waves, a filter circuit 51 that removes noise of a standard time radio signal (standard time radio signal) received by the antenna circuit 50, and a filter An RF amplification circuit 52 that amplifies a high-frequency signal that is an output of the circuit 51, and a detection circuit 53 that detects a signal output from the RF amplification circuit 52 and demodulates a standard time radio signal, and is demodulated by the detection circuit 53. The signal is output to the signal comparison circuit 18.

  FIG. 3 is a block diagram showing the configuration of the signal comparison circuit according to the present embodiment. As shown in FIG. 3, the signal comparison circuit 18 according to this exemplary embodiment includes an input waveform data generation unit 21, a reception waveform data buffer 22, a predicted waveform data generation unit 23, a waveform cutout unit 24, a correlation value calculation unit 25, and A correlation value comparison unit 26 is included. In addition, the signal comparison circuit 18 according to the present embodiment includes feature section extraction units 27 and 28, a data value accumulation unit 29, and an accumulation value buffer 30.

  The input waveform data generation unit 21 converts the signal output from the receiving circuit 16 into digital data whose value takes any one of a plurality of values at a predetermined sampling interval and outputs the digital data. For example, the sampling interval is 50 ms, and data of 20 samples per second can be acquired. The value of the digital data according to this embodiment will be described in detail later. The reception waveform data memory 22 sequentially stores the data generated in the input waveform data generation unit 21. The reception waveform data memory 22 can store a plurality of unit time length (1 second) data (for example, 10 unit time (10 seconds)). Erase.

  The predicted waveform data generation unit 23 generates predicted waveform data of a predetermined time length to be compared, which is used in a second pulse position detection process and a minute leading position detection process, which will be described later. The predicted waveform data generated by the predicted waveform data generation unit 23 will be described in detail in each detection process. The waveform cutout unit 24 extracts input waveform data having the same time length as the predicted waveform data from the received waveform data buffer 22.

  The correlation value calculation unit 25 calculates a correlation value between each of the plurality of predicted waveform data and the input waveform data. Further, in the code decoding process, the correlation value calculation unit 25 calculates a correlation value between the accumulated value stored in the accumulated value buffer 30 and the multiplication value based on the predicted waveform data. In the present embodiment, covariance is adopted to obtain a correlation as will be described later. The correlation value comparison unit 26 compares the correlation values (covariance values) calculated by the correlation value calculation unit 25 and identifies the optimum value.

  In the present embodiment, in the minute leading position detection process, the data value (feature value) of only the corresponding predetermined section (feature section) is used for calculation in both the input waveform data and the predicted waveform data. This feature section is determined by the type of code of the predicted waveform data. The feature section extraction unit 28 extracts the feature value of the feature section in the predicted waveform data according to the code type of the predicted waveform data. In addition, the feature section extraction unit 27 extracts data values of sections corresponding to the feature section extraction unit 28 in the input waveform data.

  In the present embodiment, a predetermined data value based on the input waveform data is acquired in advance for the calculation of the covariance value in the code decoding process during the minute leading position detection process. The data value accumulation unit 29 accumulates data values used for calculating the covariance. The accumulated data value is temporarily stored in the accumulated value buffer 30.

  FIG. 4 is a flowchart showing an outline of processing executed in the radio timepiece according to the present embodiment. The process shown in FIG. 4 is mainly executed by the CPU 11 and the signal comparison circuit 18 based on instructions from the CPU 11. As shown in FIG. 4, the CPU 11 and the signal comparison circuit 18 (hereinafter referred to as “CPU 11 etc.” for convenience of explanation) detect the second pulse position (step 401). The process of detecting the second pulse position is also referred to as second synchronization. FIG. 5 is a diagram for explaining the standard of the standard time radio signal. As shown in FIG. 5, the standard time radio signal is transmitted according to a predetermined standard. In the standard time radio signal, symbols indicating “P”, “1”, and “0” having a unit time length of 1 second are connected. The standard time radio wave has 60 seconds as one frame, and one frame includes 60 codes. Further, in the standard time radio wave, the position marker “P1”, “P2”,... Or the marker “M” arrives every 10 seconds, and the position marker “P0” and the frame arranged at the end of the frame. By detecting the portion where the marker “M” arranged at the head of the frame continues, the head of the frame that arrives every 60 seconds, that is, the head position of the minute can be found.

  FIGS. 6A to 6C are diagrams illustrating examples of codes according to JJY, WWVB, and MSF standards, respectively. As shown in FIG. 5 and FIG. 6A, JJY includes three codes indicating “0”, “1”, and “P”, respectively. The code of JJY is a code of unit time length (1 second), and rises from the low level to the high level at the head of the second. The code “0” of JJY becomes the high level only for the first 800 ms, and becomes the low level in the subsequent 200 ms. That is, the code “0” is an 80% duty signal. The code “1” becomes a high level only for the first 500 ms, and becomes a low level in the subsequent 500 ms. That is, the code “1” is a 50% duty signal. The code “P” is a code used as a position marker or a marker, and becomes a high level only for the first 200 ms and becomes a low level in the subsequent 800 ms. That is, the symbol “P” is a signal with a 20 percent duty.

  In the present embodiment, in order to accurately detect that the JJY signal rises from the low level to the high level at the beginning of the second, a predetermined number (units) of waveform data having a predetermined data value is obtained. In the embodiment, a plurality of pieces of predicted waveform data are generated such that they are continuous by 4) and are shifted by 50 ms. By calculating the covariance value between the multiple predicted waveform data and the input waveform data, the change point indicating the rise from the low level to the high level of the predicted waveform data indicating the optimal covariance value is set to the second pulse position. Judged as (first position of second).

  Next, the CPU 11 detects the start position of the minute, that is, the start position of the standard time radio signal of the one frame (step 402).

  After that, the CPU 11 executes a code decoding process (step 403). In step 403, various codes of the standard time radio signal (one-digit code (M1), ten-digit code (M10), other codes such as date, day of the week, etc.) Decoding is performed based on the comparison with the input waveform data.

  In the present embodiment, a case where a standard radio signal according to JJY is received and the second synchronization is performed will be described first. However, when a standard radio signal according to another standard, for example, WWVB or MSF is received. It can also be applied to. Here, the symbols according to WWVB and MSF are also briefly described.

  FIG. 6B is a diagram illustrating a code included in the WWVB of the United States. As shown in FIG. 6B, WWVB includes three codes indicating “0”, “1”, and “P”, respectively. The sign of WWVB falls from the high level to the low level at the beginning of the second. The code “0” of WWVB becomes a low level only for the first 200 ms, and becomes a high level in the subsequent 800 ms. The code “1” becomes a low level only for the first 500 ms, and becomes a high level in the subsequent 500 ms. Further, the code “P” becomes a low level only for the first 800 ms and becomes a high level in the subsequent 200 ms.

  FIG. 6C is a diagram showing symbols included in the UK MSF. Unlike JJY and WWVB, MSF has five codes, and four of them can represent the values of two bits (A, B). The sign of MSF falls from high level to low level at the beginning of the second. The code corresponding to “A = 0, B = 0” becomes a low level only for the first 100 ms, and becomes a high level in the subsequent 900 ms. The code corresponding to “A = 1, B = 0” becomes a low level only for the first 200 ms, and subsequently becomes a low level in 800 ms. Further, the code “M” corresponding to the marker becomes a low level only for the first 500 ms, and becomes a high level in the subsequent 500 ms. A code corresponding to “A = 0, B = 1” sequentially becomes a low level, a high level, and a low level by 100 ms in the first 300 ms, and then becomes a high level in 700 ms thereafter. Further, the code corresponding to “A = 1, B = 1” becomes a low level only for the first 300 ms, and becomes a high level in the subsequent 700 ms.

  Hereinafter, the second pulse position detection process (step 401) according to the present embodiment will be described in more detail. FIG. 7 is a diagram illustrating a portion of predicted waveform data used in the second synchronization according to the present embodiment. In FIG. 7, one second corresponding to the initial unit time length of each predicted waveform data is displayed. A broken line indicated by reference numeral 700 indicates the beginning of the predicted waveform data. Actually, in the present embodiment, four unit time lengths, ie, four unit time lengths, that is, four seconds of predicted waveform data shown in FIG. 23. Further, in the present embodiment, 20 pieces of predicted waveform data P (1, j) to P (20, j) in which the position of the head of data (rise from low level to high level) is shifted by 50 ms each. Is generated by the predicted waveform data generation unit 23.

  As shown in FIG. 7, the first predicted waveform data P (1, j) rises from a low level to a high level at the beginning of the data (see reference numeral 700). In the predicted waveform data according to the present embodiment, the low level is set only for a predetermined interval in front of the point rising from the low level to the high level (the point indicated by reference numeral 700 in the first predicted waveform data). And a second value indicating a high level only for a predetermined interval behind the point (a new side in time). Other than the predetermined section before and after the point, another third value indicating “0” is provided. In the example of FIG. 7, “−1” is used as the first value, “1” is used as the second value, and “0” is used as the third value. In the present embodiment, the predetermined interval of the first value indicating the low level and the predetermined interval of the second value indicating the high level are both 50 ms.

  In the actual calculation, in the portion having the first value and the portion having the second value, that is, in the first predicted waveform data P (1, j) (see reference numeral 701), reference numerals 711 and 712 are given. In the portion shown, the second predicted waveform data P (2, j) to the twentieth predicted waveform data P (20, j), only the operations indicated by the reference numerals 721, 731, 741, and 751 are substantially performed. It becomes effective.

  As can be understood from FIG. 7, the second predicted waveform data P (2, j) (see reference numeral 702) rises from the low level to the high level at a position where 50 ms has elapsed from the top of the data. Hereinafter, in the third predicted waveform data P (3, j), the fourth predicted waveform data P (4, j),..., The rising position from the low level to the high level corresponds to 50 ms. Only the position goes back sequentially.

  FIG. 8 is a flowchart showing in more detail the detection of the second pulse position (second synchronization) according to the present embodiment. FIG. 9 is a diagram schematically illustrating the second pulse position detection process according to the present embodiment. As shown in FIG. 8, the predicted waveform data generation unit 23 rises from the low level to the high level by 50 ms each in 4 unit time length (4 seconds) as described above according to the instruction of the CPU 11. Twenty pieces of predicted waveform data P (1, j) to P (20, j) whose positions are shifted are generated (step 801, reference numeral 901 in FIG. 9). As described with reference to FIG. 7, the value of the predicted waveform data is any one of the first value, the second value, and the third value.

  Next, in accordance with an instruction from the CPU 11, the waveform cutout unit 24 cuts out 4 unit time length (4 seconds) data (see reference numeral 910 in FIG. 9) from the received waveform data buffer 22 to input waveform data Sn (j). (Step 802, reference numeral 911 in FIG. 9). In this embodiment, since 20 samples of data are acquired per second, Sn (j) is data including 80 samples. In order to increase the processing speed or reduce the size of the received waveform data buffer 22, the waveform cutout unit 24 is configured so that all of the data of four unit time lengths are not stored in the received waveform data buffer 22. Sample data may be taken out in order of (1), Sn (2),.

  In the present embodiment, the input waveform data generation unit 21 binarizes the analog signal (see reference numeral 910 in FIG. 9) output from the receiving circuit 16 with a predetermined level as a threshold value. At the time of binarization, the first value “−1” is given as the data value when the level is low, and the second value “1” is given as the data value when the level is high. Therefore, the reception data buffer 22 stores digital data composed of the first value and the second value. Therefore, the data value of the input waveform data Sn (j) extracted by the waveform cutout unit 24 also takes either the first value or the second value.

  Next, the correlation value calculation unit 25 initializes a parameter p for specifying the predicted waveform data to “1” in accordance with an instruction from the CPU 11 (step 803), the input waveform data Sn (j), and the predicted waveform data P (p, The correlation value (covariance value) C (p) between each of j) is calculated (step 804).

  In the present embodiment, the correlation value calculation unit 25 includes the data value Sn (j) of the input waveform data, its average value Sm, the data value P (p, j) value of the predicted waveform data, and its average value Pm. Is used to calculate the covariance value C (p) according to the following equation. In FIG. 9, by calculating the covariance values of Sn (j) and each of P (1, j), P (2, j),..., P (20, j), C (1), C (2),..., C (20) are obtained.

C (p) = (1 / N) * Σ ((Sn (j) −Sm) * (P (p, j) −Pm))
Sm = (1 / N) * Σ (Sn (j)), Pm = (1 / N) * Σ (P (p, j))
The sigma is for j = 1 to N. As described above, when the waveform cutout unit 24 sequentially extracts sample data in the order of Sn (1), Sn (2),. All Sn (j) (j = 1 to N) is not acquired at the beginning of step 703. Therefore, at the initial stage of step 703, the average value Sm = (1 / N) * Σ (Sn (j)) cannot be obtained.

However, the above C (p) is
C (p) = (1 / N) Σ (Sn (j) * P (p, j)) − Sm * Pm
And transformed. Therefore, every time the waveform cutout unit 24 acquires the sample data Sn (j), the correlation value calculation unit 25 calculates Sn (j) * P (p, j) and obtains the calculated value as the multiplication result. When the last sample data Sn (N) is obtained, the correlation value calculation unit 25 calculates the average value Sm, and calculates Sm * Pm from the accumulation result. Subtract.

  Further, in the present embodiment, when Sn (j) is the first value “−1” and P (p, j) is the first value “−1”, Sn (j) * P (p , J) becomes “1”. Similarly, when Sn (j) is the second value “1” and P (p, j) is the second value “1”, Sn (j) * P (p, j) is “ 1 ". In other words, when P (p, j) takes the first value or the second value and Sn (j) and P (p, j) match, the predetermined value indicates a positive correlation. (“1” in this embodiment).

  On the other hand, when Sn (j) is the first value “−1” and P (p, j) is the second value “1”, Sn (j) * P (p, j) is “− 1 ". Similarly, when Sn (j) is the second value “1” and P (p, j) is the first value “−1”, Sn (j) * P (p, j) is “−1”. That is, when P (p, j) takes the first value or the second value and Sn (j) and P (p, j) do not match, a negative predetermined value indicating a negative correlation (In this embodiment, “−1”).

  Further, when P (p, j) is the third value “0”, Sn (j) * P (p, j) affects the calculation of the covariance value regardless of the value of Sn (j). The value not to be “0”.

  FIG. 10 is a diagram for explaining an example of calculating a covariance value according to the present embodiment. In the example of FIG. 10, the first one second (j = 1 to 20) is shown in the input waveform data and the predicted waveform data. Further, C (1) to C (6) displayed on the right side indicate the accumulated value of Sn (j) * P (p, j), not the covariance value itself. In the input waveform data Sn (j), when j = 4, 5, the input waveform data Sn (j) is at the high level and indicates the second value “1”. Otherwise, it is at a low level, indicating the first value “−1”.

  In the predicted waveform data P (1, j) to P (3, j), C (1) to C (3) are “0”, respectively. On the other hand, in the predicted waveform data P (4, j) whose waveform rising position matches the waveform rising position of the input waveform data Sn (j), C (4) is “2”. On the other hand, in the predicted waveform data P (6, j) in which the rising position of the waveform matches the falling position of the waveform of the input waveform data Sn (j), C (6) is “−2”. Thus, in this embodiment, when the rising position of the input waveform data matches the rising position of the predicted waveform, a larger covariance value appears to indicate that there is a stronger correlation. In addition, when the falling position of the input waveform data matches the rising position of the predicted waveform, that is, when the waveform has an opposite phase, a negative covariance value indicating a negative correlation appears. On the other hand, a value “0” that does not affect the calculation result appears except for the rising position of the predicted waveform data.

  If the parameter p is smaller than “20”, the parameter p is incremented (step 806) and the process returns to step 804. When the covariance values C (1) to C (20) are acquired for all the parameters p, Yes is determined in step 805. In this case, the correlation value comparison unit 26 compares the covariance values C (1) to C (20) to find the optimum value (in this case, the maximum value) C (x) (step 807). The CPU 11 receives the optimum value C (x) and determines whether or not the optimum value is valid (step 808).

Among the obtained covariance values C (p), C (x) indicating the maximum value is a predicted waveform having the highest correlation. However, in a covariance value obtained from a sample with insufficient parameters, noise The maximum value may appear due to an accidental factor. For the purpose of eliminating such a case, for example, in step 705, for example, the following criteria are provided to avoid erroneous detection.
(1) The number of input waveform data used for covariance calculation is greater than or equal to a predetermined number. (2) The value of x indicating C (x) appears multiple times, and the value of the multiple times x is equal. The frequency is higher than others. (X is the mode)
(3) The value of x is continuously equal to the predetermined number of times. (Continuity of mode)
When applying the determinations (1) to (3) above, the set of processes in steps 802 to 807 in FIG. 8 is executed a plurality of times.
(4) The variance of C (p) is below a specified value,
(5) Calculate kurtosis or skewness, which is a statistic of C (p), or an evaluation function equivalent thereto, and determine whether the result reaches a specified value.

  Of course, the judgment of effectiveness is not limited to the above-described method, and the average value or standard deviation of the covariance values is used, for example, even if the maximum value of the covariance values is smaller than the average value, it is not significant. It may be determined that there is not, or a general significance level (for example, 5 percent) may be used in statistics.

  If the optimum value C (x) is valid (Yes in step 808), the CPU 11 starts from the change point of the signal level in the predicted waveform data indicated by the optimum value C (x), that is, the first value indicating the low level. The position that changes to the second value indicating the high level is determined as the second pulse position (step 809). The CPU 11 stores second pulse position information in the RAM 15. This second pulse position is used in processing such as detection of the leading position as described below.

  When the second pulse position is detected (step 401), that is, when the second synchronization is completed, the minute leading position is detected (step 402). Detection of the minute start position is also referred to as minute synchronization. In step 401, the second pulse position (first position of the second) has already been determined. In JJY, in one frame, a code indicating the position marker “P0” is arranged at the end, and a code indicating the marker “M” is arranged at the top. Therefore, in the minute synchronization, the CPU 11 and the signal comparison circuit 18 indicate that the code indicating the position marker “P0” arranged at the end of the frame and the code indicating the marker “M” arranged at the top of the frame are continuous. to decide.

  Next, the minute leading position detection will be described in detail. Detection of the minute start position is also referred to as minute synchronization. FIG. 11 is a flowchart showing in more detail the detection of minute head positions (minute synchronization) according to the present embodiment. FIG. 12 is a diagram for explaining input waveform data and predicted waveform data in the detection of the leading position according to the present embodiment. With the second synchronization, the second pulse position (first position of the second) has already been determined. Also, as shown in FIG. 5, at the minute start position, the code “P” with a duty of 20% is continuous before and after (60 seconds and 1 second). Therefore, in the minute synchronization, the prediction waveform data having a unit time length of 2 units in a form in which the code “P” having a duty of 20% is generated is generated. In addition, 60 pieces of input waveform data of 2 unit time length (2 seconds) starting from the second pulse position (second start position) are generated. By calculating the correlation value between the predicted waveform data and each of the 60 input waveform data, 60 correlation values (covariance values) C (1) to C (60) can be obtained.

  As shown in FIG. 11, the predicted waveform data generation unit 23 generates predicted waveform data P (j) having a two unit time length in a form in which two symbols “P” are connected in accordance with an instruction from the CPU 11 (step S <b> 11). 1101). As shown in FIG. 12, this predicted waveform data (see reference numeral 1200) has two waveforms in which the first 200 ms (20%) is high level and the rest is low level in unit time length (1 second). It is a thing.

  The feature section extraction unit 28 extracts data values (feature values) belonging to the feature section of the predicted waveform data P (j) (step 1102). FIGS. 13A and 13B are diagrams for explaining the feature sections in each code of JJY. As shown in FIG. 13 (a), in code “P” (see symbol 1301), code “1” (see symbol 1302), and code “0” (see symbol 1303), the value is the value of another code. There are sections with unique values that are different from For example, in the code “P”, the section from 200 ms to 500 ms (see the code 1311) is a low level (data value “−1”), and in this section, has a unique value “−1” different from others. ing.

  Therefore, in the present embodiment, the section of 200 ms to 500 ms is the feature section in the code “P”. In the prediction code data P (j) shown in FIG. 11, if the number of samples of unit time length is 20, j = 1 to 40. In this case, the feature section extraction unit 28 uses the data values of the sections of 200 ms to 500 ms and the sections of 1200 ms to 1500 ms as the feature values, that is, P (5) to P (10) and P (25) to P ( Only 30) is extracted.

  As will be described later, the feature value of the feature section of the predicted waveform data is also used in the code decoding process. In the code decoding process, it is only necessary to determine whether the code is “0” or “1”. That is, since the minute synchronization has been completed, it is not necessary to determine the code “P”.

  As shown in FIG. 13B, in the code “1”, the section (see the code 1312) extending from 500 ms to 800 ms is at the low level (data value “−1”). And a unique value “−1”. On the other hand, in the code “0”, the section from 500 ms to 800 ms (see the code 1313) is at the high level (data value “1”). In this section, the unique value “1” different from the other codes “1” is used. "have. Therefore, in the codes “1” and “0”, a section of 500 ms to 800 ms is a feature section. Further, the feature value in the feature section with the code “1” is “−1”, and the feature value in the feature section with the code “0” is “1”.

Next, the parameter i for specifying the second head position is initialized, and in accordance with an instruction from the CPU 11, the waveform cutout unit 24 inputs an input waveform having a unit time length (2 seconds) from the second head position from the received waveform data buffer 22. Data Sn (i, j) is generated (step 1104). Further, the feature section extraction unit 27 extracts data values belonging to the feature section of Sn (i, j) so as to correspond to the feature section of the predicted waveform data according to the information from the feature section extraction unit 28 (step 1105). ). The feature section extraction unit 27 extracts Sn (1,5) to Sn (1,10) and Sn (1,25) to Sn (1,30) as data values belonging to the feature section when i = 1. To do. In general,
Sn (i, 5) -S (i, 10), and
Sn (i, 25) to S (i, 30) are extracted as data belonging to the feature section.

  Next, the correlation value calculation unit 25 calculates a correlation value (covariance value) C (i) between the input waveform data Sn (i, j) belonging to the feature section and the predicted waveform data P (j) belonging to the feature section. (Step 1106). Since the calculation of the covariance value is the same as that of the second synchronization process, only the balance with the feature section will be described. In the present embodiment, Sn (i, 5) to S (i, 10) and Sn (i, 25) to S (i, 30) are input waveform data as data values belonging to the feature section. Extracted. Further, P (5) to P (10) and P (25) to (P30) are also extracted from the predicted waveform data.

Therefore, when calculating the covariance value C (i), the sum of the product values of the input waveform data and the data value of the predicted waveform data ΣSn (i, j) * P (j), j = 5 -10, 25-30. Here, as described with reference to FIG. 13A, the data value (feature value) of the feature section of the symbol “P” is “−1”. Therefore, the prediction waveform data P (j) (j = 5 to 10, 25 to 30) in the prediction section is “−1”. Therefore, when calculating the total sum of the multiplication values, Sn (i, j) (j = 5 to 10, 25 to 30) may be obtained and multiplied by the feature value “−1”.

  Next, the CPU 11 determines whether or not the parameter i is 60 (step 1107). If it is determined No in step 1107, the CPU 11 increments the parameter i (step 1108). In the subsequent step 1104, according to the instruction from the CPU 11, the waveform cutout unit 24 sets the length of 2 unit times (2) from the next second start position (that is, the position 20 samples behind the second start position of the previous input waveform data). Second) input waveform data Sn (i, j). Hereinafter, a covariance value is calculated between the newly acquired input waveform data Sn (i, j) and the predicted waveform data P (j).

  As shown in FIG. 12, the input waveform data Sn (1, j) is composed of data 1201 and 1202 having a length of 2 unit time from a certain second head position. The next input waveform data Sn (2, j) is composed of data 1202 and 1203 of 2 unit time length from the next second head position. Thus, Sn (n-1, j) and Sn (n, j) are data in which the second head position is shifted by the unit time length (1 second). The last input waveform data Sn (60, j) is composed of data 1259 and 1260 of 2 unit time length shifted by 59 seconds from the first input waveform data Sn (1, j).

  The data values belonging to the feature sections of the input waveform data S (1, j), S (2, j), S (3, j),..., S (60, j) and the feature section space of the predicted waveform data Are used to calculate the respective covariance values. In FIG. 12, for the convenience of illustration, the prediction for calculating the covariance among Sn (1, j), Sn (2, j), Sn (3, j),..., Sn (60, j). The waveform data is P (1, j), P (2, j), P (3, j),..., P (60, j), but these are actually the same value P (j). It is.

  When all the correlation values (covariance values) C (1) to C (60) are acquired, the correlation value comparison unit 26 compares the covariance values C (1) to C (60) to determine the optimum value. (Maximum value in this case) C (x) is found (step 1109). The CPU 11 accepts the optimum value C (x) and determines whether or not the optimum value is valid (step 1110). Whether it is valid or not is also the same as in the case of the second synchronization process (step 808 in FIG. 8). If NO in step 1110, the process returns to step 1103, and the waveform cutout unit 24 is different from the data used for the previous processing stored in the reception waveform buffer 22 in accordance with the instruction from the CPU 11. Get input waveform data.

  When it is determined Yes in step 1110, the CPU 11 changes the head position of the second code “P” in the input waveform data indicated by the optimum value C (x), that is, from the second low level to the high level. Is determined to be the leading position of the minute (step 1111). The CPU 11 stores information on the start position of the minute in the RAM 15.

  In the present embodiment, when the detection of the second pulse position is completed and the detection of the minute leading position is started, in parallel with this, the characteristics of the input waveform data are synchronized every second so as to synchronize with the second leading position. Accumulate data values belonging to the interval. FIG. 14 is a flowchart showing an example of data value accumulation processing according to the present embodiment. When the detection of the second pulse position is completed, the CPU 11 gives a data accumulation instruction to the signal comparison circuit 18. When the data value accumulating unit 29 of the signal comparison circuit 18 receives the above instruction (step 1401), the parameter k for specifying the accumulated value S per second is initialized to “1” (step 1402).

When the detected second pulse position arrives, the CPU 11 outputs the signal to the data value accumulation unit 29. When the second pulse position has arrived, that is, when the input waveform data corresponds to the beginning of the second (step 1403), the data value accumulation unit 29 obtains the data value belonging to the feature section of the input waveform data. These are sequentially accumulated to obtain an accumulated value S (k) (step 1404). The feature section here is a section from 500 ms to 800 ms from the start position of the second in the input waveform data every second. That is, the input waveform data Sn (i, j) (i is a parameter specifying the input waveform data, j = 20), and the number of samples of the unit time length (1 second) is 20 and the data length is the unit time length. 1-20)
Sn (i, 11) to Sn (i, 16) are extracted as data belonging to the feature section, and an accumulated value of Sn (i, 11) to Sn (i, 16) is obtained as S (k).

  The data value accumulation unit 29 stores the obtained accumulation value S (k) in the accumulation value buffer 30 (step 1405). If a signal indicating the end of code decoding is received from the CPU 11 (Yes in step 1406), the process is terminated. On the other hand, if it is determined No in step 1406, the data value accumulation unit 29 increments the parameter k (step 1407) and returns to step 1403.

  Next, a description will be given of a decoding process of codes constituting the time code. By determining the start position of the minute, the position of various codes such as year, day, day of the week, hour, minute in the time code is determined. In the present embodiment, as described above, after the second pulse position is determined, the input waveform data for each second is referred to, and the data value S (k) of the feature section is accumulated for each second. Thus, the accumulated value is held in the accumulated value buffer 30.

  Therefore, in the present embodiment, a correlation value (covariance value) is calculated between the set of predicted values of the feature interval of the predicted waveform data whose code is predicted and the set of accumulated values in the accumulated value buffer. Then, the value indicated by the sign such as the year, day, day of the week, hour, and minute included in the time code can be determined from the value of the sign corresponding to the predicted waveform data with the optimal covariance value. Note that the feature value of the feature section is also referred to as a multiplication value because it is a coefficient by which the accumulated value is multiplied.

  First, the decoding of the decimal place (M1) will be described. The decimal place takes any value from “0” to “9”. In the time code, this is represented by a 4-bit BCD code. That is, the input waveform data of four unit time lengths from the leading position of the fractional sign indicates any value from “0000” to “1001”. Therefore, in the present embodiment, the value of the characteristic interval of the predicted waveform data corresponding to each bit of BCD = “0000” to “1001” is acquired as the multiplication value, and the four multiplication values and the corresponding accumulation are obtained. The covariance value is calculated by calculating the value.

  FIG. 15 is a flowchart showing in more detail a fractional decoding process according to the present embodiment. FIG. 16 is a diagram schematically illustrating a decoding process according to the present embodiment. When the detection of the minute head position is completed, the CPU 11 gives an instruction to start decoding the code to the signal comparison circuit 18. When the above instruction is received (step 1501), the correlation value calculating unit 35 refers to the accumulated value buffer 30 and accumulates the value S (m) corresponding to the code head position of the fractional position in the accumulated value buffer 30. Is identified (step 1502). When the process of FIG. 15 is started, the minute start position is fixed. Therefore, the correlation value calculation unit 25 first identifies the cumulative value corresponding to the minute start position among the cumulative values stored in the cumulative value buffer 30 according to the processing shown in FIG. 14, and then follows the time code standard. It is possible to specify the accumulated value S (m) corresponding to the leading position of the fractional sign at a position deviated by a predetermined number from the minute leading position.

  Next, starting from the accumulated value S (m), the correlation value calculating unit 25 acquires accumulated values S (m) to S (m + 3) corresponding to four unit time lengths (step 1503). Thereafter, the correlation value calculation unit 25 initializes a parameter i for specifying the BCD code to “1” (step 1504).

  The correlation value calculation unit 25 acquires the multiplication values PM (i, m) to PM (i, m + 3) for multiplying the accumulated value from the feature section extraction unit 28 (step 1505). As described with reference to FIG. 13B, in the codes “1” and “0”, sections of 500 ms to 800 ms are feature sections. Further, the data value (feature value) in the feature section with the code “1” is a constant value “−1”, and the data value (feature value) in the feature section with the code “0” is a constant value “1”. Therefore, the multiplication values PM (i, m) to P (i, m + 3) each take a value of “1” or “−1”.

  For example, if i = 1 (BCD = 0000), PM (1, m) to P (1, m + 3) are all “1”. If i = 2 (BCD = 0001), PM (2, m) to P (2, m + 2) are “1” and P (2, m + 3) are “−1”. If i = 3 (BCD = 0010), P (3, m), P (3, m + 1) and P (3, m + 3) are “1”, and P (3, m + 2) is “−1”. (See reference numerals 1630 to 1632 in FIG. 16).

  Next, the correlation value calculator 25 multiplies the accumulated values S (m) to S (m + 3) and the corresponding multiplied values PM (i, m) to P (i, m + 3), respectively, to obtain the calculated value. A covariance value C (i) is calculated based on the sum Σ (S (q) * P (i, q)) (q = m to m + 3) of the operation values (step 1506).

  As described above, the covariance value C (p) between the input waveform data Sn (j) and the predicted waveform data P (p, j) is obtained as follows. Sn (j) * P (p, j) is a product of data values belonging to the feature section of the predicted waveform data.

C (p) = (1 / N) Σ (Sn (j) * P (p, j)) − Sm * Pm
In order to facilitate understanding, the product of average values Sm * Pm is considered as “0”. When this is divided in units of 1 second,
C (p) = (1 / N) (ΣSn * P) + ΣSn * P + ΣSn * P...
It can be rewritten as Here, Sn and P are data of unit time lengths of Sn (j) and S (p, j), respectively.

As described above, the value of the feature section of the predicted waveform data of unit time length is the constant value “1” or “−1”. Therefore, the covariance value C (p) is
C (p) = (1 / N) (P * ΣSn + P * ΣSn + P * ΣSn...)
Can be rewritten. That is, P that is a constant value can be output outside the sigma (Σ). Here, ΣSn in the above equation is a feature section in the unit time length, and corresponds to the accumulated value S stored in the accumulation buffer. P corresponds to the multiplication value PM. As described above, in the present embodiment, if the accumulated value S of the data value of the unit time length is acquired in advance, it is multiplied by “1” or “−1” and further added. Thus, the sum Σ (Sn (j) * P (p, j)) of the operation values (multiplication values) of the data values in the covariance value operation can be obtained.

  As described above, the accumulated value accumulated by the data value accumulating unit 29 and accumulated in the accumulated value buffer 30 is multiplied by the multiplication value that is either “1” or “−1”. In addition, the covariance value C (p), that is, the covariance value C (i) can be obtained based on the sum of the operation values obtained by multiplication. For convenience of explanation, parameter p and parameter i are used, but the meanings of the parameters are the same.

  If the parameter i is smaller than 10 (No in Step 1507), the parameter i is incremented (Step 1508), and the process returns to Step 1505. When it is determined Yes in step 1507, the correlation value comparison unit 26 compares the obtained covariance values C (1) to C (10) and determines the optimum value (maximum value in this case) C. Find (x) (step 1509). The CPU 11 receives the optimum value C (x) and determines whether or not the optimum value is valid (step 1510).

  If the optimum value C (x) is valid (Yes in step 1510), the CPU 11 determines the BCD value indicating the optimum value C (x) as a fractional value (step 1511). The CPU 11 stores the fractional value in the RAM 15. If NO in step 1510, the process returns to step 1502. In the second and subsequent steps 1502, the accumulated value stored in the accumulated value buffer 30 is referred to find an accumulated value S (m) corresponding to another code head position, and S (m) ˜ What is necessary is just to acquire S (m + 3). In addition, when calculating the covariance value C (i), a new covariance value using the newly found accumulated value may be calculated. The final covariance value may be obtained based on the added value by adding the covariance value obtained by the previous processing.

  With reference to FIG. 16, the fractional calculation process will be described again. In a group of input waveform data 1600, input waveform data having a unit time length from the second head position is indicated by reference numerals 1601 to 1603, for example. In the data accumulation process shown in FIG. 14, during the detection of the minute start position, the data value Sn of each characteristic section (for example, reference numerals 1611 to 1613) of the input waveform data of unit time length is accumulated, and ΣSn , And the obtained accumulated value (see, for example, reference numeral 1621 to reference numeral 1623) is stored in the accumulated value buffer 30.

  In the decoding process of the code, for example, the value of the one's place shown in FIG. 16, among the accumulated values stored in the accumulated value buffer 30, values S (m) to S from the code head position are stored. (M + 3) is read and used for the calculation.

  The detection of other codes, for example, the tenths of a minute, can be performed in the same manner as the detection of a tenths digit. FIG. 17 is a flowchart showing in more detail the tens place detection process according to the present embodiment. The process of FIG. 17 can be executed in parallel with the process of FIG. Step 1701 is the same as step 1501 in FIG.

  Next, the correlation value calculation unit 25 refers to the accumulated value buffer 30 and specifies the accumulated value S (k) corresponding to the code start position of the tenth digit in the accumulated value buffer 30 (step 1702). The sign leading position of the tens place of the minute can also be specified by the same method as the sign leading position of the position of the minute position in step 1502 of FIG. 15, and therefore, the accumulated value S corresponding to the sign leading position. (K) can also be specified.

  Next, starting from the accumulated value S (k), the correlation value calculating unit 25 acquires accumulated values S (k) to S (k + 2) corresponding to three unit time lengths (step 1703). Since the tens place of the minute is composed of three codes, “0 (BCD = 000)” to “5 (BCD = 101)”, an accumulated value of 3 unit time length is acquired. Thereafter, the correlation value calculation unit 25 initializes a parameter i for specifying the BCD code to “1” (step 1704).

  The correlation value calculation unit 25 calculates multiplication values PM (i, k) to PM (i, k + 2) for multiplying the accumulated value (step 1705). Similarly to step 1505 in FIG. 15, the multiplication value is a value corresponding to each of BDD = 000 to 101. Next, the correlation value calculator 25 multiplies the accumulated values S (k) to S (k + 2) and the corresponding multiplied values PM (i, k) to P (i, k + 2), respectively, to obtain the calculated value. A covariance value C (i) is calculated based on the sum of the calculated values (step 1706). The calculation of the covariance value is the same as step 1506 in FIG.

  If the parameter i is smaller than 6 (No in Step 1707), the parameter i is incremented (Step 1708), and the process returns to Step 1705. When it is determined Yes in step 1707, the correlation value comparison unit 26 compares the obtained covariance values C (1) to C (6) to obtain an optimum value (in this case, the maximum value) C. Find (x) (step 1709). The CPU 11 accepts the optimum value C (x) and determines whether or not the optimum value is valid (step 1710).

  If the optimum value C (x) is valid (Yes in Step 1710), the CPU 11 determines the BCD value indicating the optimum value C (x) as the value of the tens place (Step 1711). The CPU 11 stores the tens value of this amount in the RAM 15. If NO in step 1711, the process returns to step 1702.

  In this way, by obtaining the fractional and tenth values, it is possible to determine the “minute” of the “hour / minute”.

  The value of the tens place and the tens place of the hour can be specified in substantially the same manner as the tens place of the minute. When the correlation value calculation unit 25 detects the value of the first place, the correlation value calculation unit 25 obtains a cumulative value corresponding to 4 unit time lengths, calculates a multiplication value corresponding to 4 unit time lengths, The covariance value is calculated by multiplying each. In addition, when detecting the ten values of time, a cumulative number corresponding to 2 unit time lengths is acquired, and a multiplication value corresponding to 2 unit time lengths is calculated and multiplied. Calculate the variance value.

  Similarly, the values can be obtained for other codes (total day from January 1 and year) by specifying a value for each digit. By specifying any value from “0” to “6” for the day of the week, the value (which day of the week) can be obtained.

  When the decoding of the minute, hour, day (total date from January 1), and year (year) is completed, the CPU 11 can obtain an accurate current time. Actually, the current time is usually acquired when the minute and hour decoding is completed. The CPU 11 corrects the current time measured internally by the internal clock circuit 17 with the accurate current time acquired by decoding (step 404 in FIG. 4). The corrected current time is displayed on the display unit 13.

  Although second synchronization, minute synchronization, and code decoding have been described for the JJY signal, second synchronization, minute synchronization, and code decoding are also possible for other standards such as WWVB and MSF.

  Second synchronization in the case of WWVB or MSF will be described. As shown in FIGS. 6B and 6C, WWVB and MSF fall from the high level to the low level at the beginning of the second. Consider the case where the predicted waveform data applied to JJY is applied to the signal falling at the head in this way. FIGS. 18A to 18C are diagrams for explaining covariance values between input waveform data and predicted waveform data. In FIG. 18A, the input waveform data Sn (j) rises from a low level to a high level at the first position of the second (see reference numeral 1800). Predicted waveform data P (j) that takes a first value corresponding to a low level before the rising point and takes a second value corresponding to a high level after the rising point; Considering the covariance value C with the input waveform data Sn (j), the maximum value is “2”. Note that the covariance value is simply the sum of the product of the data value of the predicted waveform data and the data value of the input waveform data for ease of explanation.

  On the other hand, as shown in FIG. 18B, consider input waveform data S′n (j) that falls from a high level to a low level at the start position of the second. Considering the covariance value C between the input waveform data S′n (j) and the predicted waveform data P (j), the minimum value is “−2”. That is, for the data that falls from the high level to the low level at the leading position of the second, such as WWBV or MSF, in order to detect the leading position of the second, the covariance is used as the optimum value in the processing of FIG. What is necessary is just to select the minimum value.

  Alternatively, the above-described predicted waveform data P (j) is inverted in order to detect the leading position of the second of the data that falls from the high level to the low level at the leading position of the second, such as WWVB or MSF. Other predicted waveform data may be applied. In FIG. 18C, the input waveform data S′n (j) is the same as that in FIG. The predicted waveform data P ′ (j) is obtained by inverting the predicted waveform data P (j) shown in FIG. In this case, the covariance value C (2) when p = 2 takes the maximum value. Therefore, the second position of the second may be determined based on the predicted waveform data P (2) and the input waveform data Sn (2).

  Next, minute synchronization and code decoding in the case of WWVB and MSF will be described. Also in WWVB, it is possible to calculate a covariance value using only the data value of the feature section by extracting the data value belonging to the feature section. FIG. 19 is a diagram for explaining a feature section in each code of WWVB. As shown in FIG. 19, also in the WWVB code, the value of the marker (see reference numeral 1901), the reference numeral “0” (see reference numeral 1902), and the reference numeral “1” (see reference numeral 1903) There are different sections with unique values. In the marker, an interval from 500 ms to 800 ms (see reference numeral 1911) is a low level (data value “−1”), and this interval has a unique value “−1” different from others. In the present embodiment, a section of 500 ms to 800 ms is a feature section in the WWVB marker. The data value of the feature section is “−1” indicating a low level.

  In addition, in the code “0”, the section from 200 ms to 500 ms (refer to the code 1912) is the high level (data value “1”), and in this section, a unique value “1” different from the other codes “1”. have. On the other hand, in the code “1”, the section from 200 ms to 500 ms (refer to the code 1913) is the low level (data value “−1”). In this section, the unique value “ -1 ". Therefore, in the codes “0” and “1”, a section of 500 ms to 800 ms is a feature section. The data values in the feature section are “1” and “−1”, respectively.

  Therefore, in each of the minute head position detection and code decoding processes, it is possible to execute the process using the data value of the feature section as in the case of JJY.

  FIG. 20 is a diagram for explaining a feature section in each code of MSF. As shown in FIG. 20, also in the MSF code, a marker (see reference numeral 2001), a reference numeral “00” (see reference numeral 2002), a reference sign “01” (see reference numeral 2003), a reference sign “10” (see reference numeral 2004), and In the code “11” (see the code 2005), there is a section having a unique value whose value is different from the values of other codes.

  In the marker, an interval from 300 ms to 500 ms (see reference numeral 2011) is a low level (data value “−1”), and this interval has a unique value “−1” different from others. In the present embodiment, a section of 300 ms to 500 ms is a feature section in the MSF marker. The data value of the feature section is “−1” indicating a low level.

  In addition, in the code “00”, the section (see the code 2012) from 100 ms to 300 ms is a high level (data value “1”), and in this section, has a unique value “1” different from other codes. ing. In the code “01”, a high level (data value “1”) in a section from 100 ms to 200 ms (see 2013), and a low level (data value “−1”) in a section from 200 ms to 300 ms (see 1014). In the above section, there are combinations of values different from other codes (the first half is “1” and the second half is “−1”). In the code “10”, the low level (data value “−1”) in the section from 100 ms to 200 ms (see the code 2015), and the high level (data value “1”) in the section from the 200 ms to 300 ms (see the code 1016). In the above section, there is a combination of values different from other codes (the first half is “−1” and the second half is “1”). Further, in the code “11”, the section from 100 ms to 300 ms (see the code 2017) is the low level (data value “−1”), and in this section, a unique value “−1” different from other codes is set. Have. Therefore, a section extending from 100 ms to 300 ms is a feature section. Each data value is as shown in FIG.

  In the MSF, the data values of the feature sections of the code “00” to the code “11” are not constant. For example, for the code “01”, the first half data value of the feature section is “1” and the second half data value is “−1”. Therefore, in the decoding process of the code according to the MSF standard, the data value belonging to the feature section of the input waveform data and the corresponding data value belonging to the feature section of the predicted waveform data, as in the detection of the minute start position, To obtain a calculated value, find a sum of the calculated values, and obtain a covariance value based on the obtained sum.

  Next, a second embodiment of the present invention will be described. In the above-described embodiment, when the second pulse position is detected, the first value “−1” indicating the low level only for a predetermined interval in front of the point rising from the low level to the high level (the oldest side in time). And a second value “1” indicating a high level only for a predetermined section behind the point (a new side in time), and “0” for other than the predetermined section before and after the point. Predicted waveform data having another third value indicating the above is used. However, the present invention is not limited to this, and predicted waveform data in a form in which a code constituting a time code, for example, a code “0” is repeated may be used.

  FIG. 21 is a flowchart showing in more detail the detection of the second pulse position (second synchronization) according to the second embodiment of the present invention. FIG. 22 is a diagram schematically illustrating second pulse position detection processing according to the second embodiment. As shown in FIG. 21, the predicted waveform data generation unit 23, according to the instruction of the CPU 11, has a unit time length of 4 units (4 seconds) as described above, 50 ms each, and the head (low level) of the code “0”. 20 pieces of predicted waveform data P (1, j) to P (20, j) whose positions are shifted from each other (step 2101, reference numeral 2201 in FIG. 22).

  Next, the feature section extraction unit 28 extracts feature sections of the predicted waveform data P (1, j) to P (20, j) (step 2102). As described above, in the code “0”, as shown in FIG. 13B, the feature section extends from the head of the code to the position of 500 ms to the position of 800 ms, and the feature value in the feature section is “1”. Is. Therefore, the feature section extraction unit 28 specifies the position of 500 ms to 800 ms from the beginning of the code “0” in each of the predicted waveform data P (1, j) to P (20, j).

For example, the feature section of the predicted waveform data P (1, j) is
P (1,10) -P (1,16), P (1,30) -P (1,36), P (1,50) -P (1,56), and P (1,70) To P (1,76), and the characteristic section of the predicted waveform data P (2, j) is
P (1,11) -P (1,17), P (1,31) -P (1,37), P (1,51) -P (1,57), and P (1,71) This is a section of data values of .about.P (1, 77).

In general, the characteristic section of the predicted waveform data (p, j) is
P (p, 10 + (p-1)) to P (p, 16 + (p-1)), P (p, 30 + (p-1)) to P (p, 36 + (p-1)), P ( p, 50 + (p-1)) to P (p, 56 + (p-1)) and P (p, 70 + (p-1)) to P (p, 76 + (p-1)) It becomes the section. In P (p, j), when the right number j> 80, (j−80) is obtained. Further, the data value (feature value) of the feature section is “1” as shown in FIG.

  Further, in accordance with an instruction from the CPU 11, the waveform cutout unit 24 cuts out the data of 4 unit time length (4 seconds) from the received waveform data buffer 22 to generate the input waveform data Sn (j) (step 2103, FIG. 22 reference 2211). Also in the second embodiment, since 20 samples of data are acquired per second, Sn (j) is data including 80 samples. Next, the CPU 11 initializes a parameter p for specifying the predicted waveform data to “1” (step 2104).

Next, the feature section extraction unit 27 acquires the data value of the feature section corresponding to the feature section of the predicted waveform data in the input waveform data Sn (j) (step 2105). In step 2105, corresponding to the predicted waveform data P (p, j), the data value of the feature section of the input waveform data Sn (j) is
Sn (10+ (p-1)) to Sn (16+ (p-1)), Sn (30+ (p-1)) to Sn (36+ (p-1)), Sn (50+ (p-1)) to Sn (56+ (p-1)) and Sn (70+ (p-1)) to Sn (76+ (p-1)).

  Next, the correlation value calculation unit 27 obtains an operation value by multiplying the data value of the feature section of Sn (j) and the corresponding data value (feature value) of the feature section of P (p, j). A covariance value is calculated based on the calculated value (step 2106). As described above, the data value of the feature section of the predicted waveform data P (p, j) is “1”. Therefore, in the multiplication of the data value of the feature section of Sn (j) and the corresponding data value (feature value) of the feature section of P (p, j), the data value of the feature section of Sn (j) is accumulated. It is only necessary to calculate and obtain ΣSn (j).

  If the parameter p is smaller than 20, the parameter p is incremented (step 2108) and the process returns to step 2106. Steps 2109 to 2111 are the same as steps 807 to 809 in FIG.

  In the embodiment, in the data accumulation process (FIG. 14), the data values of the input waveform data belonging to the feature section are accumulated, and the accumulated values are stored in the accumulated value buffer 30. In the decoding process, in order to obtain a covariance value, the accumulated value and the multiplied value are multiplied by using the feature value in the feature section of a specific code as a multiplication value. However, instead of accumulating the data values of the input waveform data belonging to the feature sections in advance, the input waveform data is acquired, the data values of the feature sections in the input waveform data are extracted, and the extracted data values are used. May be.

  FIG. 23 is a flowchart showing an example of a fractional detection process according to the third embodiment of the present invention. In this embodiment, the waveform cutout unit 24 acquires the input waveform data Sn (j) having a length of 4 unit time from the head position of the code (fractional place) (step 2301). If it is 20 samples per second, input waveform data of 80 samples (j = 1 to 80) can be obtained.

Next, the feature section extraction unit 27 extracts a feature section of the input waveform data Sn (j) (step 2302). As shown in FIG. 13B, the feature sections for the codes “0” and “1” based on JJY are positions from the position of 500 ms from the head of the code to the position of 800 ms. Therefore, the characteristic section of the input waveform data Sn (j) is
It is a section of data values of Sn (10) to Sn (16), Sn (30) to Sn (36), Sn (50) to Sn (56), and Sn (70) to Sn (76).

Next, the parameter BCD that specifies the value of the fractional part represented by BCD is initialized to “0000” (step 2303). The predicted waveform data generation unit 23 generates predicted waveform data P (j) having a 4-unit time length corresponding to the code indicated by the parameter BCD (step 2304). The feature section extraction unit 28 extracts a feature section of the predicted waveform data P (j) (step 2305). This feature section is
This is a section of data values P (10) to P (16), P (30) to P (36), P (50) to P (56), and P (70) to P (76).

Next, the correlation value calculation unit 25 obtains an operation value by multiplying the data value of the feature section of Sn (j) and the corresponding data value (feature value) of the feature section of P (j). A covariance value C (p) is calculated based on the calculated value (step 2306). In the process of FIG. 23, there are four characteristic sections of the predicted waveform data P (j), and each data value is “1” or “0”. Therefore, the multiplication of data values can be obtained as follows.
(Feature value of feature section (P (10) to P (16))) × Σ (Sn (10) to Sn (16)) + (Feature value of feature section (P (30) to P (36))) × Σ (Sn (30) to Sn (36)) + (feature value of feature section (P (50) to P (56))) × Σ (Sn (50) to Sn (56)) + (feature section ( Characteristic value of P (70) to P (76)) × Σ (Sn (70) to Sn (76))
As described above, it is not necessary to obtain the sum of multiplication values of all data values. If the parameter is smaller than “BCD = 1001” (step 2307), the parameter BCD is incremented (step 2308) and the process returns to step 2304. If it is determined YES in step 2307, the process proceeds to step 2309. Steps 2309 to 2311 are the same as steps 1509 to 1511 in FIG.

  In the embodiment of the present invention, the input waveform data generation unit 21 takes one of the first value indicating the low level and the second value indicating the high level as the data value of each sample point, and 1 Input waveform data having the above unit time length is generated. Further, the predicted waveform data generation unit 23 takes either the first value or the second value as the data value of each sample point, has the same time length as the input waveform data, and Predicted waveform data whose waveform shape includes a predetermined code in the time code is generated. The feature section extraction unit 28 uses a unique and constant data value different from other codes for the predetermined code (for example, a code “P” or a code “0” in accordance with JJY) in the predicted waveform data. The feature value of the feature section is acquired. Further, the feature section extraction unit 27 acquires the data value of the feature section corresponding to the feature section of the predicted waveform data in the input waveform data. The correlation value calculation means calculates the operation value by multiplying the data value of the feature section of the input waveform data by the feature value of the feature section of the predicted waveform data, and based on the sum of the operation values, predicts the input waveform data and A correlation value between the waveform data is calculated. The correlation value comparison unit calculates an optimum value of the correlation value, and the CPU 11 acquires time information included in the time code based on the optimum value.

  As described above, since the calculation is performed between the feature value of the feature section of the predicted waveform data and the corresponding data value of the feature section of the input waveform data, the amount of calculation can be reduced. In addition, since the characteristic section of the predicted waveform data has a unique and constant characteristic value different from other codes, it is possible to obtain a highly accurate correlation value by using the characteristic value of the characteristic section for calculation. Become.

  Further, in the present embodiment, in order to detect the start position of the second, the predicted waveform data generation unit 23 generates a plurality of predicted waveform data whose waveform shapes are sequentially shifted by a predetermined sample, and the correlation value The calculation unit 35 multiplies the data value of the feature section of the input waveform data by the feature value of the feature section of each of the plurality of predicted waveform data, and calculates a calculated value. The CPU 11 identifies the second position in the time code based on the predicted waveform data indicating the optimum value. As described above, in the present embodiment, the feature value of the feature section is used in the calculation to obtain a highly accurate correlation value, and based on this, the second head position, that is, the second pulse position is specified with high accuracy. Is possible.

  In the present embodiment, in order to detect the minute start position, the input waveform data generation unit 21 has a plurality of input waveform data having a plurality of unit time lengths starting from each of the second start positions. And the predicted waveform data generation unit 23 has the same time length as the input waveform data whose waveform shape includes the code of the start position of the minute in the time code (for example, two consecutive codes “P” in JJY) And the correlation value calculation unit 25 multiplies the data value of the feature section of each of the plurality of input waveform data by the feature value of the feature section of the predicted waveform data, and calculates the calculated value. To do. The CPU 11 specifies the minute start position in the time code based on the input waveform data indicating the optimum value. As described above, in the present embodiment, by using the feature value of the feature section for the calculation, it is possible to obtain a highly accurate correlation value, and to specify the minute start position with high accuracy based on the correlation value.

  Further, in the present embodiment, the input waveform data generation unit 21 generates input waveform data having a plurality of unit time lengths starting from a position corresponding to the start position of a predetermined code string in the time code. The predicted waveform data generation unit 23 indicates a sequence of codes indicating the first value or the second value (for example, BCD = 0000 to 1001), and a plurality of predicted waveform data having the same time length as the input waveform data Is generated. The correlation value calculation unit 25 calculates a calculated value by multiplying the data value of the feature section of the input waveform data by the feature value of each of the plurality of predicted waveform data. The CPU 11 acquires time information included in the time code, such as hours and minutes, based on the code string indicating the optimum value. As described above, in the present embodiment, it is possible to obtain a highly accurate correlation value by using the feature value of the feature section for the calculation, and to specify time information including time and minute accurately based on the correlation value. Become.

  Further, in the present embodiment, the correlation value calculation unit 25 accumulates the data value of the input waveform data to be multiplied by the feature value of the specific feature section, and the accumulated value of the data value of the input waveform data And the feature value of the predicted waveform data. As a result, each multiplication of the data value and the feature value can be omitted, so that the amount of calculation can be significantly reduced.

  Further, in the present embodiment, the calculated value is calculated from the first value in the input waveform data and the first value in the predicted waveform data, or the second value in the input waveform data and the first value in the predicted waveform data. A positive first calculation value “1” indicating a positive correlation is obtained by calculation with the value of 2, while calculating the first value in the input waveform data and the second value in the predicted waveform data, or The negative second calculation value “−1” indicating a negative correlation is obtained by calculating the second value in the input waveform data and the first value in the predicted waveform data. As a result, the correlation value calculation unit 25 can specify the code having the greatest correlation with the input waveform data by selecting the maximum correlation value as the optimum value.

  In the present embodiment, time information constituting the time code is acquired, and the current time is calculated based on the acquired time information. Therefore, it is possible to correct the time obtained in the internal clock circuit with the calculated current time.

  The present invention is not limited to the above embodiments, and various modifications can be made within the scope of the invention described in the claims, and these are also included in the scope of the present invention. Needless to say.

  In the detection of the second pulse position according to the above-described embodiment, the predicted waveform data has a first value indicating a low level only for a predetermined interval in front of a point that rises from a low level to a high level (on the older side in time). And a second value indicating a high level only for a predetermined interval behind the point (a new side in time). Other than the predetermined section before and after the point, another third value indicating “0” is provided. However, the present invention is not limited to this. For example, a range other than a predetermined interval before and after the point may be excluded from calculation. That is, in a section other than the predetermined section, the third value “0” may be indefinite or invalid.

  In this case, only the data of a predetermined section before and after the change point of the signal level of the predicted waveform data, that is, the point rising from the low level to the high level, is extracted, and the data value of the predetermined section and the corresponding input waveform data Calculate the covariance value of the data values. In FIG. 7, the predetermined section in the first predicted waveform data P (1, j) is a section indicated by reference numerals 711 and 712. Second predicted waveform data P (2, j), third predicted waveform data P (3, j), fourth predicted waveform data P (4, j),..., Twentieth predicted waveform data P The predetermined section in each of (20, j) is a section indicated by reference numerals 721, 731, 731,.

  With the configuration as described above, the calculation of the data value of the input waveform data and the data value of the predicted waveform data in the section other than the predetermined section can be omitted, and the number of calculations can be reduced. .

In the second embodiment, in the detection of the second pulse position, the first value indicating the low level of the input waveform data and the predicted waveform data is set to “−1”, and the second value indicating the high level is set to “1”. The other third value was set to “0”. In the calculation of the minute head position and the decoding of the code, the first value indicating the low level of the input waveform data and the predicted waveform data is set to “−1”, and the second value indicating the high level is set to “1”. . However, it is not limited to these values. The first value and the second value may be as follows.
(1) When the first value appears in the input waveform data and the first value appears in the predicted waveform data, or the second value appears in the input waveform data and the second value in the predicted waveform data A predetermined positive calculation value indicating that there is a positive correlation by calculating the first value and the first value or the second value and the second value when the value of Is calculated. That is, when the same value appears in the input waveform data and the predicted waveform data, a predetermined positive calculation value is calculated.
(2) When the first value appears in the input waveform data and the second value appears in the predicted waveform data, or the second value appears in the input waveform data and the first value in the predicted waveform data When a value of 現 れ appears, a predetermined negative calculation value indicating that there is a negative correlation is calculated by calculating the first value and the second value. That is, when different values appear in the input waveform data and the predicted waveform data, a predetermined negative calculation value is calculated. Further, it is desirable that the predetermined negative calculation value is obtained by setting the sign of the predetermined positive value to “− (minus)”.

  In the calculation of the second pulse position, the third value may not be “0”, but when the first value is calculated and when the second value is calculated, the correlation such as the covariance value is correlated. It must be a value that does not affect the value.

  Moreover, in the said embodiment, although the covariance value was utilized as a correlation value, it is not limited to this. For example, a residual that is the sum of absolute values of differences may be used as the correlation value. Alternatively, a cross-correlation coefficient may be used instead of covariance or residual.

10 radio time clock 11 CPU
12 Input unit 13 Display unit 14 ROM
15 RAM
16 Reception Circuit 17 Internal Clock Circuit 18 Signal Comparison Circuit 21 Input Waveform Data Generation Unit 22 Received Waveform Data Buffer 23 Predicted Waveform Data Generation Unit 24 Waveform Extraction Unit 25 Correlation Value Calculation Unit 26 Correlation Value Comparison Units 27 and 28 Feature Section Extraction Unit 29 Data value accumulation part 30 Accumulation value buffer

Claims (7)

Receiving means for receiving a standard time radio wave including a time code representing time information;
The signal including the time code output from the receiving means is sampled at a predetermined sampling period, and the data value at each sample point is either a first value indicating a low level or a second value indicating a high level. Input waveform data generating means for generating input waveform data having a unit time length of 1 or more,
The data value of each sample point takes either the first value or the second value, has the same time length as the input waveform data, and the waveform shape is a predetermined code in the time code Predicted waveform data generating means for generating predicted waveform data including:
Predicted waveform feature section extracting means for acquiring feature values of feature sections having unique and constant data values different from other codes in the predetermined code from the predicted waveform data;
Input waveform feature section extracting means for acquiring a data value of a feature section corresponding to a feature section of the predicted waveform data from the input waveform data;
Multiplying the data value of the feature section of the input waveform data extracted by the input waveform feature section extracting means by the feature value of the feature section of the predicted waveform data extracted by the predicted waveform feature section extracting means A correlation value calculating means for calculating a value and calculating a correlation value between the input waveform data and the predicted waveform data based on the sum of the calculated values;
A correlation value comparing means for comparing a plurality of correlation values calculated by the correlation value calculating means to calculate an optimum value of the correlation values;
And a control unit that acquires time information represented by the time code based on the optimum value. The predicted waveform data generating means generates a plurality of predicted waveform data whose waveform shape is sequentially shifted by a predetermined sample,
The correlation value calculation means calculates the operation value by multiplying the data value of the feature section of the input waveform data by the feature value of each of the plurality of predicted waveform data,
2. The time information acquisition apparatus according to claim 1, wherein the control unit specifies a second head position in the time code based on predicted waveform data indicating an optimum value. The input waveform data generating means generates a plurality of input waveform data having a plurality of unit time lengths starting from each of the first positions of the seconds;
The predicted waveform data generating means generates predicted waveform data having the same time length as the input waveform data, the waveform shape including a sign of the start position of the minute in the time code, and the correlation value calculating means includes the plurality of correlation value calculating means Each of the input waveform data is multiplied by the data value of the feature section and the feature value of the feature section of the predicted waveform data to calculate a calculated value,
3. The time information acquisition apparatus according to claim 1, wherein the control unit specifies a minute start position in the time code based on input waveform data indicating the optimum value. 4. The input waveform data generating means generates input waveform data having a plurality of unit time lengths, starting from a position corresponding to a start position of a predetermined code string in the time code;
The predicted waveform data generating means generates a plurality of predicted waveform data having the same time length as the input waveform data, indicating a sequence of codes indicating the first value or the second value,
The correlation value calculation means calculates the operation value by multiplying the data value of the feature section of the input waveform data by the feature value of each of the plurality of predicted waveform data,
The time information acquisition according to any one of claims 1 to 3, wherein the control means acquires time information included in the time code based on a code string indicating the optimum value. apparatus.   The correlation value calculation means accumulates the data value of the input waveform data to be multiplied with the feature value of a specific feature section, and multiplies the accumulated value of the data value by the feature value. The time information acquisition device according to claim 1. The calculated value is a calculation of a first value in the input waveform data and a first value in the predicted waveform data, or a second value in the input waveform data and a second value in the predicted waveform data. The first calculation value indicating a positive correlation is obtained by the calculation of the above, while the calculation of the first value in the input waveform data and the second value in the predicted waveform data, or the input waveform data By calculating the second value in the first waveform and the first value in the predicted waveform data, a negative second calculated value indicating a negative correlation is obtained.
6. The time information according to claim 1, wherein the correlation value comparison unit compares the calculated correlation values and selects a maximum correlation value as an optimum value. Acquisition device. The time information acquisition device according to any one of claims 1 to 6,
An internal time measuring means for measuring the current time by an internal clock;
Time correction means for correcting the current time measured by the internal time measurement means according to the current time acquired by the time information acquisition device;
A radio-controlled timepiece comprising time display means for displaying the current time measured by the internal time measuring means or corrected by the time adjusting means.

Images