JP6674400B2 - Time synchronization system and time synchronization method - Google Patents

Description

  The present invention relates to a time synchronization system and a time synchronization method for synchronizing time between devices connected via a network.

2. Description of the Related Art Conventionally, as a technique for synchronizing a clock in a device with a standard time (UTC: Universal Time), a method of receiving time information from a GPS (Global Positioning System) satellite is known (for example, Non-Patent Document 1 below). reference).
Also, in order to achieve time synchronization between terminals connected to a LAN or the like, a master clock and a slave device (terminal) are connected, and a standard time is transferred from the master clock to a slave device under the control using a PTP (Precision Time Protocol). Has been put to practical use.
In such a time synchronization system using a master clock, high accuracy and high reliability of the time synchronization are required. In particular, a network configuration in which a plurality of master clocks are distributed and arranged in response to the high reliability is studied. Have been. By distributing a plurality of master clocks, time synchronization over a wider range can be achieved.

Takeyasu Sakai, "Basic Knowledge of GPS / GNSS", [online], [Search February 13, 2017], Internet, <URL: http://www.enri.go.jp/~sakai/pub/symp07a .pdf # search =% 27gps +% E5% 9F% BA% E7% A4% 8E% 27>

However, the standard time received by each master clock is true due to factors such as GPS antenna installation conditions (interference and multipath) in each master clock, supplementary GPS satellite conditions, environmental disturbances (weather, ionospheric effects), and the like. There may be a difference from the standard time.
If the difference between the true standard time and the standard time distributed by each master clock becomes large, it affects applications and time services that synchronize with the standard time. For this reason, a problem is to reduce the absolute time error between the true standard time and the standard time distributed by each master clock.

  The present invention has been made in view of such circumstances, and has an object to reduce a difference between a standard time distributed by each master clock and a true standard time in a configuration in which a plurality of master clocks are distributed. I do.

  In order to achieve the above object, an invention according to claim 1 is a time synchronization system including a plurality of master clocks for distributing a standard time obtained by using a GPS (Global Positioning System) to a slave device. The master clock includes a GPS receiving unit that receives a first standard time candidate value from a GPS satellite, a synchronization receiving unit that receives a second standard time candidate value from another master clock in the time synchronization system. A standard time processing unit that calculates a standard time estimated value estimated as a true standard time using the first standard time candidate value and the second standard time candidate value; A time holding unit that holds the calculated standard time estimated value as a standard time to be delivered to the slave device, and the standard time that is held by the time holding unit A transmission unit for transmitting the estimated time value as the second standard time candidate value in the another master clock.

  The invention according to claim 5 is a time synchronization method using a plurality of master clocks for distributing a standard time obtained by using a GPS (Global Positioning System) to a slave device, wherein each of the master clocks is Receiving a first standard time candidate value from a GPS satellite, a second receiving step receiving a second standard time candidate value from another master clock, and the first standard time candidate. Using the value and the second standard time candidate value, a standard time estimate estimated as a true standard time is calculated, and the calculated standard time estimate is stored as a standard time for distribution to the slave device. And transmitting the estimated standard time value as the second standard time candidate value in the other master clock. And a transmitting step.

  By doing so, it is possible to improve the accuracy of the distribution time in the master clock as compared to simply transmitting the time received from the GPS satellite to the slave device. In addition, the standard time held by the master clocks distributed in each area as a whole network can be substantially unified, and time synchronization can be performed with high accuracy.

  Also, in the invention according to claim 2, the standard time processing unit is configured such that: the reception state and environment information of the GPS receiving unit at each master clock; and the second standard time candidate received from the other master clock. Weighting each of the first standard time candidate value and the second standard time candidate value using at least one of a value variation amount and a time and frequency synchronization state with the other master clock And calculating the standard time estimated value.

  Also, in the invention according to claim 6, in the standard time processing step, the reception state and environment information of the first standard time candidate value from the GPS satellite in each of the master clocks, from the other master clocks Using at least one of the received variation amount of the second standard time candidate value and the synchronization state of time and frequency with the other master clock, the first standard time candidate value and the second And weighting each of the standard time candidate values to calculate the standard time estimated value.

  By doing so, it is possible to calculate a standard time estimated value reflecting the reliability of each standard time candidate value, and it is possible to further reduce the error between the standard time estimated value and the true standard time.

  Further, in the invention according to claim 3, the master clock distributes the standard time to the slave device by using a Precision Time Protocol (PTP), and the synchronization transmitting unit and the synchronization receiving unit are configured to transmit the standard time to the slave device. It comprises a transmitter and a receiver connected to another master clock, and a PTP processing unit for transmitting and receiving time information using the PTP. The operation of the PTP processing unit is controlled by a master state and a slave at regular intervals. The time synchronization system is characterized in that the second standard time candidate value is transmitted / received to / from the other master clock by switching between the states.

  Also, in the invention according to claim 7, the master clock distributes the standard time to the slave device using a PTP (Precision Time Protocol), and the transmitting step and the second receiving step perform And a PTP processing unit that transmits and receives time information using the PTP. The operation of the PTP processing unit is switched between a master state and a slave state at regular intervals. The time synchronization method is characterized in that the transmission step and the second reception step are sequentially executed by switching.

  By doing so, it is possible to transmit and receive the second standard time candidate value with another master clock using a single PTP processing unit, and separately provide a master-side PTP processing unit and a slave-side PTP processing unit This is advantageous in simplifying the configuration of the master clock as compared with the case.

  Further, in the invention according to claim 4, the master clock distributes the standard time to the slave device using a Precision Time Protocol (PTP), and the synchronization transmitting unit connects to the other master clock. A first transmitter and a first receiver, and a master-side PTP processing unit that functions as a master side of the PTP and transmits the standard time estimation value held by the time holding unit to the other master clock. The synchronization receiving unit is configured to function as a slave side of the PTP with a second transmitter and a second receiver connected to the other master clock, and to perform synchronization from the other master clock. A slave PTP synchronization processing unit for receiving a second standard time candidate value. Was.

  Also, in the invention according to claim 8, the master clock distributes the standard time to the slave device using a Precision Time Protocol (PTP), and the transmitting step is connected to the other master clock. A first transmitter and a first receiver, and a master side PTP processing unit which functions as a master side of the PTP and transmits the standard time estimation value held in the standard time processing step to the other master clock. Being executed, the second receiving step includes a second transmitter and a second receiver connected to the other master clock, and a second transmitter and a second receiver that function as a slave side of the PTP. And a slave PTP synchronization processing unit that receives the standard time candidate value of Was.

  By doing so, the second standard time candidate value can be constantly received from all the connected master clocks, which is advantageous in improving the accuracy of the standard time estimated value.

  According to the present invention, in a configuration in which a plurality of master clocks are distributed and arranged, it is advantageous in reducing the difference between the standard time distributed by each master clock and the true standard time.

FIG. 1 is a diagram illustrating a configuration of a time synchronization system according to an embodiment. FIG. 3 is a block diagram showing a configuration of a master clock. FIG. 3 is a diagram illustrating a method of switching port states in a master clock. FIG. 14 is a block diagram showing another configuration of the master clock. 5 is a flowchart illustrating processing of a master clock. It is an explanatory view showing a time synchronization process by a PTP processing unit.

(Embodiment)
Hereinafter, a time synchronization system and a time synchronization method according to the present invention will be described in detail with reference to the accompanying drawings.
FIG. 1 is an explanatory diagram illustrating a configuration of a time synchronization system 10 according to the embodiment.
The time synchronization system 10 includes a plurality of master clocks 20 (20A to 20F) that deliver the standard time received from the GPS satellites 40 to the slave devices 50 (50A, 50B).
In the time synchronization system 10, a plurality of master clocks 20 are connected to each other, and use PTP-based synchronization to reduce the time error and minimize the absolute time error from the true standard time.

The GPS satellite 40 has an atomic clock 42 mounted thereon and transmits a signal including a standard time (UTC: Coordinated Universal Time).
The master clock 20 extracts time information (first standard time candidate value) from a signal received from the GPS satellite 40. Further, the second standard time candidate value is received from another master clock 20. Then, a standard time estimated value is calculated using the first standard time candidate value and the second standard time candidate value, and the calculated standard time estimated value is delivered to the slave device 50 under the control as the standard time.
It is assumed that the respective master clocks 20 are connected to each other (master clock network N), and frequency synchronization with the same frequency is established between the respective master clocks 20.
In FIG. 1, for simplification of description, one master clock 20 is connected to two adjacent master clocks 20. However, in the case of a mesh configuration or the like, the number of connections is increased.
Examples of the slave device 50 include a boundary clock (BC) 50A and a service node 50B. Although details will be described later, standard time is distributed between the master clock 20 and the slave device 50 using PTP.

FIG. 2 is a block diagram showing a configuration of the master clock 20.
The master clock 20 includes a GPS antenna (GPS receiving unit) 202, a synchronization processing unit 210 (210A to 210n), a standard time processing unit 220, a time holding unit 222, a switching control unit 224, a port state switching unit 226, and a time distribution unit 230. Is provided.
Since the basic function of the master clock 20 is to distribute the standard time to the slave device 50 by the time distribution unit 230, the time distribution unit 230 will be described first.

The time distribution unit 230 includes a time distribution PTP processing unit (PTP processing unit) 232, a time distribution transmitter (transmitter) 234, a time distribution receiver (receiver) 236, and a time distribution port (PTP port) 238. It is comprised including.
The time distribution PTP processing unit 232 exchanges a time stamp with the slave device 50 via the time distribution transmitter 234, the time distribution receiver 236, and the time distribution port 238, so that the time The held standard time estimate is delivered to the slave device 50 as the standard time. More specifically, by exchanging time stamps, the slave device 50 calculates the offset between the clock on its own device side and the clock on the master clock 20 side, thereby obtaining the standard time at which the master clock 20 holds the clock of the slave device 50. Synchronize to estimates.

FIG. 6 is an explanatory diagram illustrating the time synchronization processing by the PTP processing unit.
In the processing shown in FIG. 6, the delay time in the path from the master clock 20 (hereinafter, referred to as “master side”) to the slave device 50 (hereinafter, referred to as “slave side”) and the delay time in the path from the slave side to the master side Using the time, an offset of the slave side time with respect to the master side time is calculated.

First, the delay time in the route from the master side to the slave side is calculated.
First, the master transmits a Sync message to the slave (S1). The Sync message includes a predicted value of the transmission time T1 of the Sync message. Next, the master side transmits a Sync follow-up message including the transmission time T1 of the actual Sync message to the slave side (S2).
The slave records the time T2 when the Sync message is received from the master (S3).
From the time T1 and the time T2, the delay time in the path from the master side to the slave side can be calculated. That is, the delay time on the path from the master side to the slave side = T2-T1.

Next, the delay time in the path from the slave side to the master side is calculated.
First, the slave transmits a delay request message to the master (S4), and the slave records the transmission time T3 of the delay request message (S5). Next, a delay response message including the time T4 at which the master side received the delay request message is transmitted from the master side to the slave side (S6).
From the time T3 and the time T4, the delay time in the path from the slave side to the master side can be calculated. That is, the delay time in the path from the slave side to the master side = T4−T3.

Using the delay time (first delay time) in the path from the master side to the slave side and the delay time (second delay time) in the path from the slave side to the master side, a one-way delay time between the master and the slave is calculated. , Can be calculated using the following equation (1).
One-way delay time = (first delay time + second delay time) / 2 = ((T2-T1) + (T4-T3)) / 2 (1)
Then, the offset (time shift) of the slave side time with respect to the master side time is calculated using the following equation (2).
Slave offset = first delay time-one-way delay time = ((T2-T1)-(T4-T3)) / 2 (2)

Returning to the description of FIG. 2, the GPS antenna 202 receives the first standard time candidate value from the GPS satellite 40. The reception of the first standard time candidate value by the GPS antenna 202 is performed at a predetermined cycle.
Originally, the time transmitted from the GPS satellite 40 is considered to be the standard time itself. However, there is a possibility that an error from the true standard time is included depending on the signal reception status of the GPS antenna 202 and the like. Therefore, the master clock 20 handles the time received by the GPS antenna 202 as a standard time candidate value.

The synchronization processing section 210 (210A to 210n) exchanges time information (time stamp) with another master clock 20, receives a second standard time candidate value, and transmits a second standard time candidate value to the other master clock 20. The standard time candidate value is transmitted.
The synchronization processing unit 210 is provided for each of the other master clocks 20 connected to the master clock 20. The example in FIG. 2 shows that n other master clocks 20 are connected to the master clock 20.

The synchronization processing unit 210 (210A to 210n) includes a synchronization PTP processing unit (PTP processing unit) 212 (212A to 212n), a synchronization transmitter (transmitter) 214 (214A to 214n), and a synchronization receiver (reception). 216 (216A to 216n) and a synchronization port (PTP port) 218 (218A to 218n).
The processing of the synchronization PTP processing unit 212 is the same as that of the time distribution PTP processing unit 232 described with reference to FIG. 6, except that the time distribution PTP processing unit 232 performs only the operation on the master side. The synchronization PTP processing unit 212 can switch the operation on the master side and the operation on the slave side under the control of the port state switching unit 226. Hereinafter, when the operation of the synchronization PTP processing unit 212 is on the master side, it is referred to as “master state”, and when it is on the slave side, it is referred to as “slave state”.
Further, when the operation state of the synchronization PTP processing unit 212 is switched, the port state of the synchronization port 218 is switched. That is, when the synchronization PTP processing unit 212 is in the master state, the corresponding synchronization port 218 becomes the master-side port, and when the synchronization PTP processing unit 212 is in the slave state, the corresponding synchronization port 218 becomes the master port. Slave side port.

When the synchronization PTP processing unit 212 has been switched to the master state, a message including a time stamp based on the standard time held by the time holding unit 222 is transmitted to the slave side (the other master clock 20). That is, when the synchronization PTP processing unit 212 has been switched to the master state, the synchronization standard transmission unit transmits the standard time estimated value held by the time holding unit 222 as the second standard time candidate value in another master clock. The synchronization processing unit 210 functions as a unit.
On the other hand, when the synchronization PTP processing section 212 is switched to the slave state, the master PTP is processed using the time stamp transmitted from the master side (the other master clock 20) and the standard time candidate value held in the time holding section 222. Calculate the offset of the own device with respect to the local time. That is, when the synchronization PTP processing unit 212 is switched to the slave state, the synchronization processing unit 210 serves as a synchronization reception unit that receives the second standard time candidate value from another master clock 20 in the time synchronization system 10. Works.

The standard time processing unit 220 estimates the true standard time using the first standard time candidate value received from the GPS satellite 40 and a plurality of second standard time candidate values received from other master clocks. Calculate the estimated standard time.
The standard time processing unit 220 compares the first standard time candidate value and a plurality of second standard time candidate values with each other, and weights each standard time candidate value based on various parameters. Then, the standard time estimated value is calculated by averaging (weighted average) the weighted standard time candidate values or by selecting the standard time candidate value with the smallest offset.
Note that, out of the standard time candidate values, a clear failure value and a standard time candidate value with extremely low reliability of the master clock 20 of the transmission source are excluded from the standard time estimation value calculation target samples.
Further, as a method of calculating the standard time estimated value based on the standard time candidate value, various conventionally known methods can be applied.

The following are examples of the weighting parameters.
The information shown below is received together with the time information when the synchronization PTP processing unit 212 of the own device is switched to the master state and the synchronization PTP processing unit 212 of another master clock 20 is switched to the slave state. deep.
1. 1. Reception state and environment information of GPS antenna 202 at each master clock 20 1-1. GPS satellite information, number of GPS satellites being captured, year of manufacture of GPS satellites being captured, on-board oscillator of GPS satellites being captured (Cs or Rb)
・ Navigation message information from GPS satellites (satellite orbit (ephemeris), errors in internal clocks of GPS satellites, etc.)
-Position information of the GPS satellites with respect to the GPS antenna 202 (zenith, elevation, etc.)
-Errors due to Doppler shift of GPS satellites and relativistic effects 1-2. Environmental condition / weather monitor value (surrounding weather condition)
-Troposphere and ionospheric delay 1-3. Receiver (GPS antenna 202) status / GPS radio wave reception status, noise value (SNR (Signal-to-Noise Ratio, etc.))
-GPS antenna 202 installation conditions (such as the magnitude of multipath effects due to reflection)
1. Cable delay correction information between the GPS antenna 202 and the receiver 2. Offset of standard time (second standard time candidate value) by PTP received from another master clock 20 and its variation amount Synchronization state of time and frequency between master clocks 20 in master clock network N. State such as normal / holdover / free run. Quantitative numerical values such as frequency deviation and drift amount of a transmitter in master clock 20.

  The time holding unit 222 holds the estimated standard time calculated by the standard time processing unit 220 as the standard time to be delivered to the slave device 50. The time holding unit 222 holds the standard time by synchronizing a built-in clock with the estimated standard time. The time held by the time holding unit 222 is updated every time the standard time estimated value is calculated by the standard time processing unit 220 (if offset = 0, there is no update).

  The switching control unit 224 controls switching of the operation state of the synchronization PTP processing unit 212 by the port state switching unit 226 (that is, switching of the port state of the synchronization port 218). The switching control unit 224 shares the port state switching timing with the switching control unit 224 of another master clock 20 in advance. Then, by synchronizing with the switching control unit 224 of another master clock 20 and switching the operation state of each synchronization PTP processing unit 212, transmission / reception of the standard time candidate value between the master clocks 20 is enabled.

  Here, three patterns shown in FIG. 3 are given as a port state switching method.

<Pattern 1>
In the pattern 1 shown in FIG. 3A, the port states (master / slave) of all master clocks 20 in the time synchronization system 10 are simultaneously changed at regular intervals.
In FIG. 3A, three master clocks (a master clock 20A, a master clock 20B, and a master clock 20C, hereinafter referred to as "master clock A", "master clock B", and "master clock C") are shown. Each is connected.
The master clock A has a port Pab connected to the master clock B and a port Pac connected to the master clock C. The master clock B has a port Pba connected to the master clock A and a port Pbc connected to the master clock C. The master clock C has a port Pca connected to the master clock A and a port Pcb connected to the master clock B.
In the initial state of FIG. 3A, the port Pab of the master clock A is in the slave state S, the port Pac is in the master state M, the port Pba of the master clock B is in the master state M, and the port Pbc is in the slave state. S, the port Pca of the master clock C is in the slave state S, and the port Pcb is in the master state M.
At a predetermined port state switching timing, the switching control unit 224 of each master clock 20 transmits a control signal for instructing each synchronization PTP processing unit 212 to switch the port state from the port state switching unit 226. Each synchronization PTP processing unit 212 interrupts the currently executing PTP session and switches the operation state.

3A, the port Pab of the master clock A is in the master state M, the port Pac is in the slave state S, the port Pba of the master clock B is in the slave state S, and the port Pbc. Is in the master state M, the port Pca of the master clock C is in the master state M, and the port Pcb is in the slave state S.
As described above, by simultaneously changing the port states of all the master clocks 20 at regular intervals, time information (second standard time candidate value) from all other master clocks 20 can be constantly obtained in all the master clocks 20. Can be.
In pattern 1, when there is a possibility that an error has occurred in some master clocks 20 and erroneous time information is transmitted, the master clock 20 of the communication destination transmits the time information received by its own GPS antenna 202. The normality is determined by comparing the (first standard time candidate value) with time information (second standard time candidate value) from the master clock 20 which may have an abnormality. If it is determined that there is an abnormality, the time information received from the master clock 20 having the abnormality is excluded from the standard time estimated value calculation target sample.

<Pattern 2>
In the pattern 2 shown in FIG. 3B, the port status (master / slave) is changed in units of one link (between two opposing devices) in order at regular intervals.
The port configuration in Pattern 2 is the same as in Pattern 1.
In FIG. 3B, the port state is changed in the order of a link L1 connecting the master clock B and the master clock C, a link L2 connecting the master clock A and the master clock C, and a link L3 connecting the master clock A and the master clock B. Shall be.
In the initial state of the left diagram of FIG. 3B, the port Pab of the master clock A is in the slave state S, the port Pac is in the master state M, the port Pba of the master clock B is in the master state M, and the port Pbc is in the slave state. S, the port Pca of the master clock C is in the slave state S, and the port Pcb is in the master state M.
When the port state switching timing of the link L1 comes, the switching control unit 224 of the master clock B and the master clock C transmits a control signal instructing switching of the port states of the ports Pbc and Pcb, respectively.
As a result, the port Pbc of the master clock B is switched to the master state M and the port Pcb of the master clock C is switched to the slave state S, as shown in the center diagram of FIG. In the master clock B, both ports are in the master state M, and in the master clock C, both ports are in the slave state S.

Next, at the timing of switching the port state of the link L2, the switching control unit 224 of the master clock A and the master clock C transmits a control signal instructing the switching of the port states of the ports Pac and Pca, respectively.
As a result, the port Pac of the master clock A is switched to the slave state S and the port Pca of the master clock C is switched to the master state M, as shown in the right diagram of FIG.
Although not shown, at the port state switching timing of the link L3, the switching control unit 224 of the master clock A and the master clock B transmits a control signal for instructing switching of the port state of the ports Pab and Pba, respectively.
Thereby, the port Pab of the master clock A is switched to the master state M and the port Pba of the master clock B is switched to the slave state S, and the port states of all the ports are inverted from the initial state shown in the left diagram of FIG. I do.

In pattern 2, for example, the port state change order of each link is set in advance by EMS (Element Management System) / NMS (Network Management System). Alternatively, for example, when some ports are changed from the master state to the slave state in a certain master clock, the other (other link) port is changed from the slave state to the master state at the next timing, so that the order is changed. The port status may be changed for each link.
When some master clocks are transmitting abnormal times, the normality is determined by comparing the time of its own GPS antenna and the time of the master clock to be newly referenced at the next port state change timing. .
In pattern 2, when there is a possibility that an error has occurred in some master clocks 20 and erroneous time information is transmitted, the master clock 20 of the communication destination transmits the time information received by its own GPS antenna 202. The normality is determined by comparing the (first standard time candidate value) with the time information (second standard time candidate value) of the master clock 20 which is newly referred to at the next port state change timing.

<Pattern 3>
In pattern 3 shown in FIG. 3C, a master-only port and a slave-only port are provided between each master clock, and the port state is always fixed.
In the case of pattern 3, each master clock 20 has the configuration shown in FIG.
4, the same components as those in FIG. 2 are denoted by the same reference numerals, and detailed description thereof will be omitted. Although the illustration of the internal configuration of the time distribution unit 230 is omitted in FIG. 4 due to space restrictions, it is the same as the configuration shown in FIG.

In the configuration of pattern 3 shown in FIG. 4, master synchronization processing section 250 (250A to 250n) and slave synchronization processing section 240 (240A to 240n) are provided instead of synchronization processing section 210 (210A to 210n). I have.
Each master-side synchronization processing unit 250 (250A to 250n) includes a master-side PTP processing unit 252 (252A to 252n), a synchronization transmitter (transmitter) 254 (254A to 254n), and a synchronization receiver (receiver). 256 (256A to 256n), and a synchronization port (master side port) 258 (258A to 258n).
Further, each slave side synchronization processing section 240 (240A to 240n) includes a slave side PTP processing section 242 (242A to 242n), a synchronization transmitter (transmitter) 244 (244A to 244n), and a synchronization receiver (reception). 246 (246A to 246n) and a synchronization port (master side port) 248 (248A to 248n).
The configurations of the master-side synchronization processing section 250 and the slave-side synchronization processing section 240 are basically the same as those of the synchronization processing section 210 of FIG. 2, but the operation of each synchronization PTP processing section is in the master state (master-side PTP processing section 252). ) Or slave state (slave-side PTP processing unit 242).
Further, since the operation state (port state) of the synchronization PTP processing unit is fixed as described above, the port state switching unit 226 and the switching control unit 224 are not provided.
According to the configuration of Pattern 3, since the time information (second standard time candidate value) can be constantly obtained from all of the other connected master clocks 20, the estimation accuracy of the standard time can be improved.

FIG. 5 is a flowchart showing the processing of the master clock 20.
The master clock 20 distributes the estimated standard time held by the time holding unit 222 to the slave device 50 (Step S500). Note that the distribution of the standard time to the slave device 50 is continuously performed during the processing of this flowchart.
Further, the master clock 20 receives time information (first standard time candidate value) from the GPS satellite 40 by the GPS antenna 202 at predetermined intervals (step S502). The received first standard time candidate value is held by the standard time processing unit 220 until the standard time estimated value calculation timing (step S512: Yes).
If the port state of the synchronization processing unit 210 is the master state (step S504: Yes), the standard time estimation value held by the time holding unit 222 is transmitted to the other master clock 20 of the connection destination (step S504). S506). On the other hand, if the port state of the synchronization processing unit 210 is not the master state (step S504: No), the process proceeds to the next step S508.
If the port state of the synchronization processing unit 210 is in the slave state (step S508: Yes), the synchronization processing unit 210 receives the standard time estimated value (second standard time candidate value) in the other master clock 20 of the connection destination. (Step S510). On the other hand, if the port state of the synchronization processing unit 210 is not in the slave state (step S508: No), the process proceeds to the next step S512. The standard time processing unit 220 holds the second standard time candidate value received in step S510 until the standard time estimated value calculation timing (step S512). That is, the standard time processing unit 220 waits until the calculation (update) timing of the standard time estimated value is reached (Step S512 → No).

When it is time to calculate (update) the standard time estimated value (step S512: Yes), the standard time processing unit 220 calculates the standard time estimated value from the held standard time candidate value (step S514), and The time estimation value is stored in the time storage unit 222 (step S516). That is, the clock in the time holding unit 222 is synchronized with the estimated standard time. As a result, the estimated standard time value transmitted to the slave device 50 is updated.
Subsequently, the switching control unit 224 waits until a preset port state switching timing is reached (Step S518: No). Then, at a preset port state switching timing (step S518: Yes), the switching control unit 224 transmits a port state switching control signal to the synchronization PTP processing unit 212 to be switched. The state switching unit 226 is controlled (step S520). When the pattern 3 shown in FIG. 3C is used as the port state switching method, the processing in steps S518 and S520 is not required.
After that, the master clock 20 continues the processing of the flowchart in FIG.

As described above, the time synchronization system 10 according to the embodiment exchanges time information held between a plurality of master clocks 20 and uses the master information to generate a standard time estimated value estimated as a true standard time. calculate.
By doing so, it is possible to improve the accuracy of the distribution time in the master clock 20 as compared with simply transmitting the time received from the GPS satellite 40 to the slave device 50. In addition, the standard times held by the master clocks 20 distributed in each area of the entire network can be substantially unified, and the accuracy of time synchronization can be improved.
In addition, since the time synchronization system 10 calculates the standard time estimated value reflecting the reliability of each standard time candidate value, it is possible to further reduce the error between the standard time estimated value and the true standard time.
Further, in the time synchronization system 10, the operation of the synchronization PTP processing unit 212 is switched between the master state and the slave state at regular intervals as in pattern 1 or 2 of FIG. By transmitting and receiving the second standard time candidate value, it is possible to transmit and receive the second standard time candidate value to and from another master clock 20 using a single PTP processing unit. This is advantageous in simplifying the configuration of the master clock 20 as compared with the case where the slave side PTP processing unit is separately provided.
Further, in the time synchronization system 10, if the master-side PTP processing unit and the slave-side PTP processing unit are provided separately as shown in pattern 3 in FIG. Candidate values can be received, which is advantageous in improving the accuracy of the standard time estimation value.

10 Time Synchronization System 20 (20A-20F) Master Clock 202 GPS Antenna (GPS Receiver)
210 (210A to 210n) synchronization processing unit (transmission unit for synchronization, reception unit for synchronization)
212 (212A to 212n) PTP processing unit for synchronization (PTP processing unit)
214 (214A to 214n) Synchronization transmitter 216 (216A to 216n) Synchronization receiver 218 (218A to 218n) Synchronization port 220 Standard time processing unit 222 Time holding unit 224 Switching control unit 226 Port state switching unit 230 Time distribution Unit 232 Time distribution PTP processing unit 234 Time distribution transmitter 236 Time distribution receiver 238 Time distribution port 40 GPS satellite 50 (50A, 50B) Slave device N Master clock network

Claims (8)

A time synchronization system including a plurality of master clocks for distributing a standard time obtained by using a GPS (Global Positioning System) to a slave device,
Each of the master clocks is
A GPS receiver for receiving a first standard time candidate value from a GPS satellite;
A synchronization receiving unit that receives a second standard time candidate value from another master clock in the time synchronization system;
A standard time processing unit that calculates a standard time estimated value estimated as a true standard time using the first standard time candidate value and the second standard time candidate value;
A time holding unit that holds the standard time estimated value calculated by the standard time processing unit as a standard time to be delivered to the slave device,
A synchronization transmission unit that transmits the standard time estimation value held by the time holding unit as the second standard time candidate value in the another master clock.
A time synchronization system characterized in that: The standard time processing unit includes: a reception state and environment information of the GPS receiving unit in each of the master clocks; a variation amount of the second standard time candidate value received from the other master clock; Calculating the standard time estimated value by weighting each of the first standard time candidate value and the second standard time candidate value using at least one of a time and frequency synchronization state between
The time synchronization system according to claim 1, wherein: The master clock distributes the standard time to the slave device using PTP (Precision Time Protocol),
The synchronization transmission unit and the synchronization reception unit are configured by a transmitter and a receiver connected to the other master clock, and a PTP processing unit that transmits and receives time information using the PTP.
Transmitting and receiving the second standard time candidate value to and from the other master clock by switching the operation of the PTP processing unit between a master state and a slave state at regular intervals;
3. The time synchronization system according to claim 1, wherein: The master clock distributes the standard time to the slave device using PTP (Precision Time Protocol),
The synchronization transmission unit is configured to transmit a first transmitter and a first receiver connected to the other master clock and the standard time estimation value that functions as a master side of the PTP and that is held by the time holding unit. A master-side PTP processing unit for transmitting to the other master clock,
The synchronization receiver includes a second transmitter and a second receiver connected to the other master clock, and a second standard time candidate from the other master clock that functions as a slave side of the PTP. A slave PTP synchronization processing unit that receives the value.
3. The time synchronization system according to claim 1, wherein: A time synchronization method using a plurality of master clocks for distributing a standard time obtained by using a GPS (Global Positioning System) to a slave device,
Each of the master clocks is
A first receiving step of receiving a first standard time candidate value from a GPS satellite;
A second receiving step of receiving a second standard time candidate value from another master clock;
The first standard time candidate value and the second standard time candidate value are used to calculate a standard time estimate that is estimated as a true standard time, and the calculated standard time estimate is used as the slave device. A standard time processing step to hold as standard time to be delivered to
Transmitting the standard time estimated value as the second standard time candidate value in the another master clock.
A time synchronization method, characterized in that: In the standard time processing step, the reception state and environment information of the first standard time candidate value from the GPS satellite in each of the master clocks, and the second standard time candidate value received from the other master clock Weighting each of the first standard time candidate value and the second standard time candidate value using at least one of a variation amount and a synchronization state of time and frequency with the other master clock. Calculating the standard time estimate;
6. The time synchronization method according to claim 5, wherein: The master clock distributes the standard time to the slave device using PTP (Precision Time Protocol),
The transmitting step and the second receiving step are executed by a transmitter and a receiver connected to the other master clock, and a PTP processing unit that transmits and receives time information using the PTP,
The transmission step and the second reception step are sequentially performed by switching the operation of the PTP processing unit between a master state and a slave state at regular intervals.
7. The time synchronization method according to claim 5, wherein: The master clock distributes the standard time to the slave device using PTP (Precision Time Protocol),
The transmitting step includes a first transmitter and a first receiver connected to the other master clock, and the standard time estimation value that functions as a master side of the PTP and is held in the standard time processing step. Executed by a master-side PTP processing unit that transmits to another master clock,
The second receiving step includes a second transmitter and a second receiver connected to the other master clock, and a second standard time from the other master clock that functions as a slave side of the PTP. Executed by the slave-side PTP synchronization processing unit that receives the candidate value,
7. The time synchronization method according to claim 5, wherein:

Images