Information and communication system and information and communication device
The system addresses synchronization accuracy issues by calculating interval times and adjusting clock frequencies to synchronize devices, effectively reducing errors from propagation time fluctuations.
Patent Information
- Application Number
- JP2021153508
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-09-21
- Publication Date
- 2025-08-19
- Estimated Expiration
- 2041-09-21
AI Technical Summary
Existing time synchronization systems, such as IEEE1588 PTP, face challenges in maintaining accuracy due to fluctuations in propagation time caused by device buffering and network nodes, leading to increased costs with the use of boundary and transparent clocks, and independent calculation of minimum packet time failing to maintain symmetry.
An information communication system and device that calculates transmission and reception interval times, selects sets of communications to minimize time differences, and adjusts clock frequencies to synchronize devices, absorbing fluctuations in propagation time.
Achieves reduced error synchronization control by absorbing fluctuations in propagation time and maintaining accurate time synchronization.
Smart Images

Figure 0007725311000021 
Figure 0007725311000022 
Figure 0007725311000023
Abstract
Description
[Technical Field]
[0001] The present invention relates to an information communication system and an information communication device that synchronizes a plurality of information communication devices through communication. [Background technology]
[0002] A common time synchronization method between multiple information communication devices is, for example, the IEEE1588 Precision Time Protocol (PTP) described in Non-Patent Document 1. Non-Patent Document 1 defines a master device that has a reference time and a slave device that synchronizes with the time of the master device, and corrects the time of the slave device by periodically exchanging time synchronization packets between the master device and the slave device.
[0003] Specifically, the slave device estimates and corrects the time offset, which is the time difference between the master device and the slave device, using the master device's transmission time and slave device's reception time of a packet transmitted from the master device to the slave device, as well as the slave device's transmission time and master device's reception time of a packet transmitted from the slave device to the master device. [Prior art documents] [Non-patent literature]
[0004] [Non-Patent Document 1] “IEEE Standard for a Precision Clock Synchronization Protocol for Networked Measurement and Control Systems.”IEEE Standard 1588-2008. [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-225880 Summary of the Invention [Problem to be solved by the invention]
[0005] In a PTP system, when information is transmitted and received bidirectionally between a master device and a slave device, it is assumed that the propagation time from one device to the other is the same as the propagation time from the other device to the other device. However, the propagation time is affected by factors such as the buffering time of the devices that make up the network and the number of nodes through which the information passes during transmission and reception. Therefore, in reality, there may be a difference between the transmission and reception times, making it impossible to maintain symmetry.
[0006] To address this issue, PTP systems use network relay devices such as boundary clocks (BCs) and transparent clocks (TCs) to cancel out fluctuating propagation times and reduce the problem of reduced time synchronization accuracy caused by fluctuations in propagation time. BCs correct delays and fluctuations based on the time received from a higher-level master device and generate time as a master for lower-level slaves. TCs add the residence time within the relay device when relaying packets and send the time to the lower-level slaves.
[0007] However, BC and TC are generally expensive, which increases the cost of building a time synchronization system. For this reason, there is a need for alternative solutions that can satisfy the system requirements for time synchronization accuracy without introducing BC and TC.
[0008] For example, Patent Document 1 proposes a method of absorbing fluctuations in propagation time by using the minimum packet time. However, because the minimum packet time in each direction of information communication is calculated independently from moment to moment, the symmetry of propagation time is not evaluated, which leads to a decrease in time synchronization accuracy.
[0009] Furthermore, because the main cause of time difference fluctuations is the frequency deviation of the clocks between devices, some implementations of PTP provide an option to calculate and cancel this frequency deviation. However, this does not eliminate the error factor caused by fluctuations in propagation time, which leads to a decrease in time synchronization accuracy.
[0010] The present invention has been made to solve the above-mentioned problems, and aims to provide an information and communication system and an information and communication device that can absorb fluctuations in propagation time of information communication and perform synchronization control with reduced errors. [Means for solving the problem]
[0011] The present invention is an information communication system in which a plurality of information communication devices communicate information as master devices or slave devices, and in a first set of information communications which are a plurality of information communications performed within a predetermined time from a first timing between the master device and the slave devices, and a second set of information communications which are a plurality of information communications performed within the predetermined time from a second timing that is equal to or longer than the predetermined time from the first timing, For each pair of information communication from the first set of information communication and information communication from the second set of information communication, an interval time calculation unit that calculates a transmission interval time and a reception interval time; a selection unit that selects a set of information communications from the first set of information communications and the second set of information communications such that the difference between the reception interval time and the transmission interval time is smallest; a frequency deviation calculation unit that calculates a deviation in clock frequency between the master device and the slave device based on the set of information communications selected by the selection unit; and a synchronization control unit that synchronizes the clock frequencies or times of the master device and the slave device based on the deviation in clock frequency calculated by the frequency deviation calculation unit. The predetermined time is the minimum interval time at which the influence of frequency deviation appears in the information propagation time between the master device and the slave device. .
[0012] The present invention is an information communication device that communicates information with another information communication device, and includes a first set of information communications that are multiple pieces of information communications that are performed within a predetermined time from a first timing, and a second set of information communications that are multiple pieces of information communications that are performed within the predetermined time from a second timing that is equal to or longer than the predetermined time from the first timing, For each pair of information communication from the first set of information communication and information communication from the second set of information communication,an interval time calculation unit that calculates a transmission interval time and a reception interval time; a selection unit that selects a set of information communications from the first set of information communications and the second set of information communications such that the difference between the transmission interval time and the reception interval time is smallest; a frequency deviation calculation unit that calculates a deviation in clock frequency with another information communications device based on the set of information communications selected by the selection unit; and a synchronization control unit that synchronizes a clock frequency or a time based on the clock frequency deviation calculated by the frequency deviation calculation unit. The predetermined time is the minimum interval time at which the influence of frequency deviation appears in the propagation time of information between other information communication devices. . [Effects of the Invention]
[0013] According to the present invention, it is possible to provide an information communication system and an information communication device that can absorb fluctuations in the propagation time of information communication and perform synchronization control with reduced errors. [Brief explanation of the drawings]
[0014] [Figure 1] 1 is a schematic diagram of an information communication system according to an embodiment. [Figure 2] 1 is a functional block diagram of an information communication device constituting an information communication system according to an embodiment. [Figure 3] FIG. 1 is a diagram illustrating a mode of communication between information communication devices. [Figure 4] FIG. 2 is a functional block diagram of a control unit according to the embodiment. [Figure 5] FIG. 1 illustrates a communication aspect with propagation time symmetry. [Figure 6] FIG. 1 illustrates a mode of communication that does not have propagation time symmetry. [Figure 7] FIG. 10 is a diagram showing a synchronization state vector without errors due to variations in propagation time and time difference. [Figure 8] FIG. 10 is a diagram showing a synchronization state vector including errors due to fluctuations in propagation time and time difference. [Figure 9] 10 is a flowchart illustrating a procedure for a synchronization control process according to an embodiment. [Figure 10] FIG. 1 is a diagram illustrating a set of information communications in an information communication system according to an embodiment. [Figure 11] 10A and 10B are diagrams illustrating shift amounts of statistics of a set of information communications at different timings. [Figure 12] FIG. 1 is a diagram plotting synchronization state vectors for a set of multiple information communications. [Figure 13] FIG. 2 is a diagram showing a set of two-way information communications in the information communication system according to the embodiment. [Figure 14] FIG. 1 is a plot of synchronization state vectors in a set of asymmetric two-way information communications. DETAILED DESCRIPTION OF THE INVENTION
[0015] Hereinafter, an information communication system and an information communication device according to an embodiment will be described with reference to the drawings.
[0016] [composition] Fig. 1 is a schematic diagram of an information communication system 100 according to an embodiment. Fig. 2 is a functional block diagram of an information communication device 1 constituting the information communication system 100 according to the embodiment. Fig. 3 is a diagram showing a mode of communication between the information communication devices 1, and Fig. 4 is a functional block diagram of a control unit 70 in Fig. 2.
[0017] An information communication system 100 according to this embodiment is made up of a plurality of information communication devices 1, and each information communication device 1 synchronizes through information communication. An information communication device 1 serving as a slave device synchronizes with an information communication device 1 serving as a master device. A master device is an information communication device 1 that is to be synchronized with another information communication device 1 in the information communication system 100. A slave device is an information communication device 1 that synchronizes with another information communication device 1 that is a master device in the information communication system 100. A slave device synchronizes with the master device through transmission and reception of synchronization information, which is information for achieving time synchronization from the master device to the slave device.
[0018] The information communication device 1 that will be the master device will be referred to as master device CLa, and the information communication device 1 that will be the slave device will be referred to as slave device CLb. There must be at least one slave device CLb for the master device CLa, but as shown in FIG. 1, the system may include multiple slave devices CLb.
[0019] The master device CLa and the slave device CLb are connected via a network capable of communicating information via a wire. In this embodiment, an example will be described in which the information communication device 1 achieves synchronization by transmitting and receiving information via a wire.
[0020] (Information and communication devices) The information communication device 1 includes a computer, and performs necessary calculations by having a processor including a CPU or the like execute a program that is stored in advance in a storage unit such as an HDD or SSD.
[0021] Specifically, as shown in Fig. 2, the information communication device 1 has a communication unit 10, a clock 20, a timepiece 30, a storage unit 50, an external interface 60, and a control unit 70. For example, each of the units 10 to 70 is configured as hardware. Each unit of the control unit 70 may be configured as software including programs and data. Which parts of the control unit 70 are configured as software can be changed as needed.
[0022] The communication unit 10 transmits and receives information to and from other information communication devices 1. That is, the communication unit 10 transmits information to the outside of the information communication device 1, receives information from the outside of the information communication device 1, or performs both of these. The communication unit 10 has a transmitter 11, a receiver 12, a transmission timing detection unit 13, and a reception timing detection unit 14.
[0023] The transmitter 11 is a device that transmits input information. Specifically, the transmitter 11 breaks down the information into its smallest components in chronological order and transmits the information to the outside. The packet length of the information (amount of communication information) is arbitrary and may be different for each communication. For example, an input unit that converts an audio signal input from a microphone into audio data is connected to the information communication device 1, and the audio data as information is input from the input unit to the information communication device 1.
[0024] The receiver 12 is a device that receives information from outside. Specifically, the receiver 12 reconstructs information that has been decomposed into minimum components in time series and received from outside the information communication device 1, and outputs the reconstructed information to other components within the information communication device 1. For example, a playback unit such as an ear receiver or speaker that outputs audio is connected to the information communication device 1, and the receiver 12 converts audio data into an audio signal and outputs the audio signal to the playback unit. In this way, the information that has been decomposed into minimum components in time series and transmitted and received allows the position of each element in time series, i.e., the information element position, to be identified.
[0025] Here, the information communicated by the information communication device 1, that is, the information transmitted by the transmitter 11 and the information received by the receiver 12, includes synchronization information for synchronizing the time between the master device CLa and the slave device CLb. The synchronization information in this embodiment includes at least information on the time corresponding to the transmission timing.
[0026] The information communication device 1 has one or both of a transmitter 11 and a receiver 12. When the information communication device 1 has both, the transmitter 11 and the receiver 12 may be selectively operated exclusively, or both may be operated simultaneously.
[0027] The transmission timing detection unit 13 detects the transmission timing. The transmission timing is the timing at which a predetermined information element position of the information transmitted by the transmitter 11 is transmitted to the outside of the information communication device 1. This transmission timing is detected based on the clock period of a clock 20 (described later) (in other words, the period of the pulses oscillated by the clock 20). In other words, the transmission timing is expressed based on an integer multiple of the clock period.
[0028] Furthermore, the transmission timing detection unit 13 outputs the detection result to other components within the information communication device 1. The information referred to here is, for example, a packet, and in this case, the predetermined information element position (hereinafter referred to as the predetermined information element position) is a bit position. For example, when the information is broken down in time series into eight minimum component bits, the transmission timing detection unit 13 detects the timing at which the third information element position (bit position) is transmitted, and outputs the timing at which the third information element position (bit position) is transmitted to the outside.
[0029] The reception timing detection unit 14 detects the reception timing. The reception timing is the timing at which a predetermined information element position of the information received by the receiver 12 is received from outside the information communication device 1. This reception timing is detected based on the clock period of a clock 20, which will be described later. In other words, the reception timing is expressed as an integer multiple of the clock period.
[0030] Furthermore, the reception timing detection unit 14 outputs the detection result to other components within the information communication device 1. For example, when information is broken down in time series into bits, which are the eight smallest components, the reception timing detection unit 14 detects the timing at which the third information element position (bit position) is received, and outputs the timing at which the third information element position (bit position) is received to the outside.
[0031] In the above examples of transmission timing and reception timing, the position of the same smallest element is detected in both cases, but for example, the transmission timing detection unit 13 of the information communication device 1 acting as the transmitter detects the fifth element, and the reception timing detection unit 14 of the information communication device 1 acting as the receiver detects the eighth element. As long as a predetermined relationship (for example, an interval (here, three)) is maintained between the transmitting and receiving information communication devices 1 in each timing detection, it is not necessary to detect the same position, and any deviation in the detected element position can be ignored or corrected.
[0032] FIG. 3 shows a configuration for wired communication as an example, in which, in packet transmission between information communication devices 1, specific bit positions within a packet are detected as transmission and reception timings, and the transmission interval and reception interval are calculated from adjacent transmission and reception timings, respectively.
[0033] Furthermore, the transmission timing may be detected a predetermined time before or after the actual transmission timing, and the reception timing may be detected a predetermined time before or after the actual reception timing. These predetermined values may be fixed within the communication unit, or may be set statically or dynamically from the outside.
[0034] The clock 20 oscillates at a predetermined frequency and outputs a signal that provides operation timing to each unit of the information communication device 1. As a result, each unit in the information communication device 1 operates in synchronization with the clock 20. The entire information communication device 1 may be synchronized simultaneously with a single clock 20, or multiple clocks 20 may be used to independently synchronize multiple functional units. This clock 20 has an inherent finite oscillation frequency tolerance. In other words, the clock 20 has an error (e.g., 20 ppm) relative to a predetermined oscillation frequency (e.g., 10 MHz). For example, a fixed-frequency oscillator such as a quartz crystal resonator can be used as the clock 20.
[0035] Even though the nominal frequencies of the clocks 20 of the master device CLa and the slave device CLb are the same, there are actual individual differences. That is, there is a frequency deviation between the frequencies of the clocks 20 of the master device CLa and the slave device CLb. The clock 20 may receive an external frequency control signal and variably control the oscillation frequency in response to this signal.
[0036] The clock 30 uses the output signal of the clock 20 as an oscillation source to keep time and outputs the relative time since the information communication device 1 was started up. The clocking may be synchronized with a divided frequency of the input clock signal. The clocking may also have a variable advance per clock, or the clock frequency may be fixed but the clock drive frequency may be intermittently controllable. The time is referenced in a specified unit, but this specified value may be fixed within the clock or may be set statically or dynamically from outside. The clock 30 outputs the relative time in response to a reference request from outside, for example.
[0037] The storage unit 50 is a recording medium such as an HDD, SSD, memory, or register. The storage unit 50 stores information necessary for the control unit 70 to perform calculations. The time of the clock 30 corresponding to the transmission timing or reception timing described below is preferably stored in a recording medium that can be accessed only by hardware, without the intervention of a CPU or software. This is because jitter caused by software can be eliminated. It is important that the correspondence between the transmission and reception timing and time is not affected by software jitter, and after the transmission and reception timing and time are associated, the time may be stored in a slow-access area.
[0038] The memory serving as the storage unit 50 inputs and outputs any information and stores the information in a designated storage area. Information is stored in response to an external storage request, at which time the information to be stored and the storage area are input. Information is referenced in response to an external reference request, at which time the storage area for the reference information is input, and the information in the storage area specified by that input is output. Information may be stored only while the device is operating, or may be retained permanently, even when the device is stopped.
[0039] The external interface 60 (hereinafter also referred to as external I / F 60) connects the inside of the information communication device 1 with the outside, and inputs and outputs any information. The information includes transmitted and received data and the time of the clock 30. For example, the external I / F 60 acquires information to be stored in the memory unit 50 from the outside. The external I / F 60 may also acquire from the outside the time from the reception timing detected by the reception timing detection unit 14 to the transmission timing detected by the transmission timing detection unit 13, and use this time in the scheduler 74 described below.
[0040] Master device CLa has at least an external I / F 60 that transmits synchronization information. Slave device CLb has at least an external I / F 60 that receives synchronization information. However, master device CLa may have an external I / F 60 that receives synchronization information, and slave device CLb may have an external I / F 60 that transmits synchronization information. Furthermore, information communication device 1 may have a user interface such as a display device that can control and display the internal state.
[0041] The control unit 70 controls the overall operation of each unit of the information communication device 1. Fig. 4 is a functional block diagram of the control unit 70. As shown in Fig. 4, the control unit 70 has a main control unit 71, a transmission / reception data I / F 72, a communication control unit 73, a scheduler 74, a time recording unit 75, an interval time calculation unit 76, a selection unit 77, a frequency deviation calculation unit 78, a time difference calculation unit 79, a synchronization control unit 80, and a timing information reference unit 81.
[0042] The main control unit 71 is linked to each unit within the control unit 70 and controls the operation of each unit within the control unit 70. The transmission / reception data I / F 72 converts information from the memory unit 50 and the external I / F 60 into a format that can be transmitted outside the device. The transmission / reception data I / F 72 also converts information received from outside the device into a format suitable for the control unit 70 and the memory unit 50.
[0043] The communication control unit 73 controls the operation of the communication unit 10. The communication control unit 73 inputs and outputs transmission and reception information between the communication unit 10 and the control unit .
[0044] The scheduler 74 controls the schedule (time) for transmitting or receiving information. For example, when communicating information one-way, the scheduler 74 on the transmitting side controls the interval between information transmissions or sets a schedule in which the timing of transmitting information is detected at a predetermined time. The scheduler 74 may also control the time from when information is received to when it is transmitted. When transmitting information, the communication control unit 73 can include in the information a time corresponding to the transmission timing of a preset schedule or the detected transmission timing. In this embodiment, the scheduler 74 controls the first timing and second timing described below and the timing of transmitting synchronization information in accordance with these timings.
[0045] The time recording unit 75 associates the transmission timing of a predetermined information element position in the information transmitted by the transmission timing detection unit 13 with the time of the clock 30 at that transmission timing, and stores the association in the memory of the storage unit 50. For example, when the time recording unit 75 receives a signal from the transmission timing detection unit 13 indicating that the transmission timing of a predetermined information element position in the transmitted information has been detected, the time recording unit 75 refers to the time of the clock 30 and associates that time with the transmission timing.
[0046] The time recording unit 75 also functions as a time adding unit that adds a time corresponding to the transmission timing to the information to be transmitted. As in transmission, the time recording unit 75 associates the reception timing of a predetermined element position in the information received by the reception timing detection unit 14 with the time of the clock 30 at the reception timing, and stores the association in memory. The time recording unit 75 also adds a time corresponding to the reception timing to the information and causes the transmitter 11 to transmit it.
[0047] Thus, in this embodiment, "time" refers to the time of the clock 30 corresponding to the detected transmission timing or reception timing of a predetermined information element position in the information, and "hours" refers to the difference between the times.
[0048] The interval time calculation unit 76 calculates a transmission interval time and a reception interval time for a first set of information communications and a second set of information communications between master device CLa and slave device CLb. The first set of information communications is a plurality of information communications that are performed within a predetermined time from a first timing. The second set of information communications is a plurality of information communications that are performed within a predetermined time from a second timing that is at least a predetermined time from the first timing.
[0049] The predetermined time is an interval time during which the influence of clock frequency deviation between master device CLa and slave device CLb does not appear in the propagation time of information between master device CLa and slave device CLb. Therefore, if the interval time between the first timing and the second timing exceeds the predetermined time, the influence of frequency deviation appears in the propagation time.
[0050] The transmission interval time is the difference between the transmission time of the first timing and the transmission time of the second timing. The transmission interval time is calculated by referencing the transmission time included in the received information and deriving the time difference between the two transmissions. The reception interval time is calculated by referencing the time corresponding to the reception timing from the clock 30 and deriving the time difference between the two receptions. The two transmission times and the two reception times may be read directly from the received information or may be read from the memory unit 50.
[0051] The selector 77 selects a set of information communications from the first set of information communications and the second set of information communications, which has the smallest difference between the reception interval time and the transmission interval time. This set can be considered to have the smallest fluctuation in propagation time, as will be described later.
[0052] The frequency deviation calculation unit 78 calculates the deviation of the clock frequencies between the master device CLa and the slave device CLb based on the set of information communications selected by the selection unit 77. That is, the frequency deviation calculation unit 78 calculates the deviation between the clock frequency of the information communication device 1 to be synchronized (master device CLa) and the clock frequency of the information communication device 1 to be synchronized (slave device CLb). This deviation of the clock frequencies (hereinafter also simply referred to as "frequency") is a value expressed as a frequency ratio. In other words, calculating the frequency deviation is synonymous with calculating the frequency ratio. The frequency deviation calculation unit 78 of this embodiment calculates the frequency ratio by, for example, dividing the reception interval time in the set of information communications selected by the selection unit 77 by the transmission interval time.
[0053] In this embodiment, the frequency ratio is the ratio of the clock frequency of the slave device to the clock frequency of the master device (hereinafter, sometimes simply referred to as the frequency ratio). In other words, the frequency deviation is the difference between the clock frequency of the master device and the clock frequency of the slave device.
[0054] Time difference calculation unit 79 finds the time difference corresponding to the largest number of sets of information communications between master device CLa and slave device CLb within a predetermined time period, among sets with the smallest time difference errors.
[0055] The synchronization control unit 80 controls the clock 20 or the timepiece 30 based on the frequency ratio calculated by the frequency deviation calculation unit 78. That is, the synchronization control unit 80 controls the oscillation frequency of the clock 20 so that the frequency ratio calculated by the frequency deviation calculation unit 78 becomes 1. For example, if the clock 20 is a voltage-controlled crystal oscillator including a crystal resonator, the frequency ratio becomes 1 by controlling the voltage applied to the oscillator, thereby synchronizing the clock frequency with the information communication device 1 that is the synchronized device (master device).
[0056] Furthermore, the synchronization control unit 80 calculates the time difference, i.e., the time difference, with respect to the information communication device 1 that is the synchronized device (master device CLa) based on the frequency ratio calculated by the frequency deviation calculation unit 78, and corrects the value kept by the clock 30 based on that time difference. In this embodiment, the time difference can be calculated by the time difference calculation unit 79, and the synchronization control unit 80 can correct the time based on the calculated time difference. For example, the correction may be performed by controlling the clock 30 itself to correct the time. Alternatively, the correction may be performed based on the time difference when the clock 30 outputs the time. In other words, the time kept by the clock 30 itself may not be corrected, but the clock 30 may be controlled so that when the clock 30 outputs the time, it outputs a time corrected for the difference.
[0057] The timing information reference unit 81 references the timing information output from the synchronization control unit 80 and executes desired processing in synchronization with the timing information. The timing information is composed of a periodic timing signal and time information that uniquely corresponds to a specified timing. The synchronization control unit 80 can synchronize with other information communication devices 1 based on the timing information.
[0058] [Measurement of time difference when propagation times are symmetric] Next, measurement of propagation time and time difference when there is symmetry in the propagation time will be explained with reference to Fig. 5. Fig. 5 shows the relationship between the time difference between the clocks and the propagation time of information between a pair of master device CLa and slave device CLb, and the transmission and reception timing observed through information communication, using wired communication as an example.
[0059] The synchronization information transmitted by the master device CLa and the other synchronization information transmitted by the slave device CLb are intended to determine the timing for synchronization, and the synchronization information includes the transmission time of the master device CLa, but the contents of the synchronization information and the other synchronization information are arbitrary.
[0060] As shown in FIG. 5, the master device CLa transmits information at time t a,T and after a propagation time of d, the signal is transmitted to the b,R The slave device CLb receives the signal at Δtb Later, the slave device CLb sends another piece of information at time t b,T After the same propagation time d, the master device CLa transmits the a,R Receive at.
[0061] Here, the interval from transmission to reception in the master device CLa is Δt a and the interval from reception to transmission in the slave device CLb is Δt b The transmission timing of the master device CLa is t a,T The same timing is t' in the slave device CLb. a,T Therefore, the time difference g between the clock 20 of the master device CLa and the clock of the slave device CLb at a certain moment can be calculated using equation (1). Furthermore, the time difference between the clock of the master device CLa and the clock of the slave device CLb at the same moment can be expressed as -g.
[0062] TIFF0007725311000001.tif11161 Note that, as described above, it is assumed that the propagation time d from master device CLa to slave device CLb is the same as the propagation time d from slave device CLb to master device CLa, that is, they are symmetrical in the direction of information communication.
[0063] From these conditions, the propagation time d can be calculated as shown in equation (2). TIFF0007725311000002.tif15159
[0064] Therefore, the time difference g between the clocks 20 can be calculated using equation (2) as shown in equation (3). TIFF0007725311000003.tif14158
[0065] Synchronization control unit 80 of slave device CLb can synchronize its time with master device CLa by repeatedly adjusting the clock frequency and time so that the time difference g between clocks 20 becomes zero.
[0066] If the clock frequencies of master device CLa and slave device CLb were the same, i.e., if the clock domain were single, time synchronization would be possible with just a single time adjustment. However, since there is generally a frequency deviation between the two clocks, clock frequency adjustment is essential for time synchronization. Even in systems where the clock frequency cannot be adjusted directly, time synchronization can be achieved by repeatedly adjusting the advance of time per clock, thereby controlling the drive frequency of clock 30. In these frequency adjustments, closed-loop control is periodically performed, with the synchronization phase as input and the frequency adjustment value as output, and the output value is controlled so that the input synchronization phase becomes zero.
[0067] It should be noted that propagation time and internal delays between transmission and reception timings can be ignored or corrected depending on system requirements, etc. Also, instead of master device CLa initiating transmission of synchronization information, slave device CLb may transmit synchronization information. Furthermore, although one interface is sufficient for information communication, multiple interfaces may be used simultaneously, for example, by implementing separate transmission and reception interfaces.
[0068] [Effects of time difference and propagation time variations] The above explanation is based on the assumption that the time difference is constant and that the propagation time is symmetrical in the direction of information communication. However, the actual frequency deviation from the nominal frequency of the clock 20 between the information communication devices 1 varies and fluctuates from moment to moment. Therefore, the time difference between the information communication devices 1 also fluctuates, and depending on the network system to which the master device CLa and the slave device CLb are connected, the propagation time is generally asymmetrical due to the influence of the information propagation path and communication arbitration, as shown in Figure 6.
[0069] That is, the time difference g between the master device CLa and the slave device CLb at a certain moment b / a Then, after a certain period of time has passed, the time difference g between the same master device CLa and slave device CLb a / bIn addition, the propagation time d of information from the master device CLa to the slave device CLb is b / a is the propagation time d of information from the slave device CLb to the master device CLa a / b In general, the time difference in Figure 6 is expressed as g b / a and g a / b have the same sign, and g b / a ·g a / b ≧0. Furthermore, as mentioned above, the time difference and propagation time fluctuate from moment to moment, and the time difference g b / a [m] and propagation time d b / a [m] is the time difference g observed at different moments b / a [n] and propagation time d b / a It is common for [n] to be different from each other.
[0070] Here, from the relationship between the times in Figure 6, the propagation time d b / a ,d a / b and time difference g b / a ,g a / b Using this, equations (2) and (3) are rearranged into equations (4) and (5), respectively. TIFF0007725311000004.tif15162TIFF0007725311000005.tif13162
[0071] Equation (2) allows us to calculate the propagation time assuming symmetry, but equation (4) shows that the average propagation time includes a variable component of the time difference. Similarly, equation (3) allows us to calculate the time difference assuming the time difference is constant, but equation (5) shows that the average time difference includes a variable component of the propagation time. In this way, the inclusion of variable components of the time difference and propagation time leads to a decrease in the accuracy of time synchronization.
[0072] [Errors due to time difference and fluctuations in propagation time] We will explain the time synchronization error caused by fluctuations in time difference and propagation time in a time synchronization system. First, we combine equations (4) and (5) and transform them into a matrix form as shown in equation (6). TIFF0007725311000006.tif15162
[0073] Here, the two-dimensional rotation matrix R θ is expressed as equation (7). TIFF0007725311000007.tif15160
[0074] Then, by focusing on the common factor in equation (6), R in equation (7) θ By using this, equation (6) can be transformed into equation (8). TIFF0007725311000008.tif14163
[0075] As shown in Figure 7, the horizontal axis represents the time difference g when communication is performed from master device CLa to slave device CLb. b / a and propagation time d b / a The vertical axis represents the time difference g when communication is performed from slave device CLb to master device CLa. a / b and propagation time d a / b Consider a two-dimensional space with components of (g b / a - g a / b ) T The time difference vector g, (d b / a d a / b ) T If we let d be the propagation time vector, then we can obtain the synchronization state vector s = (g d d g ) T can be interpreted as the sum of the time difference vector g and the propagation time vector d rotated by +π / 4 on the same space and scaling each dimension.
[0076] In the operation of equation (8), element g of the time difference vector g b / a , g a / b There is no fluctuation in the element d of the propagation time vector d. b / a , d a / b There is no change in g b / a =g a / b And d b / a =d a / b As shown in Figure 7, when gIngredients and g d It can be seen that by orthogonally decomposing the components, it is possible to obtain the time difference and delay time without including any error.
[0077] However, in general, the time difference vector g and the propagation time vector d each contain a fluctuation error in their respective elements, as shown in Figure 8. Because the errors in each of the orthogonal components are compounded, it is not possible to remove the errors from the time difference and propagation time by simply manipulating equation (8), which leads to a decrease in time synchronization accuracy.
[0078] Since equations (4) and (5) are linearly dependent, an analytical solution cannot be found without introducing other independent operations, but by measuring only the transmission and reception timing of synchronization information, it is not possible to find new independent operations that result in the vectors shown in Figure 7. Therefore, a different approach must be introduced to remove the fluctuation error.
[0079] [Operation] Based on the above, the operation of the information communication system 100 of this embodiment will be described. First, an overview of the operation of this embodiment will be described with reference to the flowchart of FIG. 9. The communication control unit 73 performs the first information communication and the second information communication in accordance with the scheduler 74 (steps S101 and S102). As will be described later, this is because the master device CLa transmits a plurality of pieces of synchronization information in a burst manner from the first timing, i.e., timing m, and then transmits L g This corresponds to transmitting multiple pieces of synchronization information in a burst fashion starting from timing n, which is the first timing separated by a time interval of [n] (see FIG. 10). Hereinafter, transmitting multiple pieces of synchronization information consecutively in a relatively short period of time like this is also referred to as burst transmission.
[0080] Next, the interval time calculation unit 76 calculates the transmission interval time and the reception interval time in the first set of information communication and the second set of information communication (step S103). This is done based on the transmission time t recorded by the time recording unit 75 in the storage unit 50 or read from the received information. a,T [n],t a,T [m], Reception time t b,R [n], tb,R [m] to the transmission interval time (t a,T [n]-t a,T [m], reception interval time (t b,R [n]-t b,R This is equivalent to finding the
[0081] The selection unit 77 selects the combination of information communication that minimizes the difference between the transmission interval time and the reception interval time (step S104), the frequency deviation calculation unit 78 calculates the frequency ratio between the master device CLa and the slave device CLb, that is, the frequency deviation, based on the selected combination of information communication (step S105), and the time difference calculation unit 79 calculates the time difference (step S106). These will be described in detail later.
[0082] By correcting the frequency deviation and time difference, synchronization control unit 80 controls synchronization of the clock frequency or time between master device CLa and slave device CLb (step S107). Note that synchronization control by synchronization control unit 80 is repeatedly performed by feedback until the frequency deviation and time difference converge to 0 (frequency ratio becomes 1).
[0083] [Frequency deviation estimation under propagation time fluctuations] Next, the selection of a set of information communication by the selection unit 77 (step S104) and the calculation of the frequency deviation by the frequency deviation calculation unit 78 (step S105) will be described in detail.
[0084] First, the oscillation frequency f of the clock 20 of each information communication device 1 used for time synchronization is the nominal frequency f N , which includes a frequency deviation e. Furthermore, since the frequency deviation e varies from moment to moment depending on the mechanical and electromagnetic operating environment of the clock 20, the relationship between the transmission frequency f[n] and the frequency deviation e[n] at the observation timing n can be expressed as in equation (9).
[0085] The frequency deviation is sufficiently smaller than 1, and r[n] represents the frequency ratio. Each clock has a set allowable frequency deviation from the nominal frequency, but as mentioned above, the actual frequency deviation fluctuates from moment to moment, so it is not possible to accurately grasp it. TIFF0007725311000009.tif11161
[0086] Figure 10 shows the relationship between the transmission and reception timing, time difference, and propagation time when communicating information from master device CLa to slave device CLb. The nominal frequency of the clock 20 used for time synchronization in both devices is the same, but the frequency deviation is a [n], slave device CLb is e b [n]. When a large number of packets are sent from master device CLa to slave device CLb, a time difference gradually occurs over time due to frequency deviation e. There is also variation in each propagation time. The frequency deviation e is calculated from this situation.
[0087] At this time, the frequency deviation of the slave device CLb as seen from the master device CLa is (e b [n]‐e a The frequency ratio r b / a [n] is r b / a [n]=1+(e b [n]‐e a Similarly, the frequency deviation of master device CLa as seen from slave device CLb can be expressed as (e a [n]‐e b The frequency ratio r a / b [n] is r a / b [n]=1+(e a [n]‐e b The frequency ratio is 1 when the frequency deviation is 0, and (e b [n]‐e a [n]), (e a [n]‐e b [n]) represents the deviation from the state of 1 where there is no frequency deviation.
[0088] The clocks 20 of the information communication devices 1 are each specified with a tolerance for the nominal frequency, and the clocks 20 change from moment to moment within this range. If the clock frequencies of the information communication devices 1 differ, the advance of the clocking time also differs, resulting in a time difference from moment to moment. b / a [m]≠g a / b There is also an observation timing of [n], and correction taking into account the frequency deviation is required in time calculations between the information communication devices 1. If correction is not performed, errors will occur in the time calculations, leading to a decrease in the accuracy of time synchronization.
[0089] In FIG. 10, the average frequency ratio of the slave device CLb as seen from the master device CLa from the first timing m to the second timing n is defined as r b / a If [n, m], the relationship in equation (10) holds. TIFF0007725311000010.tif20162
[0090] From equation (10), g b / a [n]=g b / a [m]+(r b / a [n,m]-1)(t a,T [n]-t a,T It can be seen that the time difference changes due to the influence of frequency deviation.
[0091] In addition, in equation (10), the frequency deviation is not applied to the propagation time, so correction is not required. b / a [n,m] to d b / a It is assumed that the cases of multiplying by [n] and adding without multiplying are almost equal. This is because the clock 20 has a nominal frequency f N , the minimum time L that causes a time difference between the observation times of the information communication devices 1 with apparent frequency deviation e[n] g [n] can be expressed as in equation (11), but in general, the propagation time is this minimum time L g This is because it is smaller than [n]. N =1 / f NIn other words, since the propagation time is a very small value, even if frequency deviation is applied to it, the effect is not apparent. In other words, the predetermined time for transmitting synchronization information from timing m and the predetermined time for transmitting synchronization information from timing n are the minimum time L g A time smaller than [n]. TIFF0007725311000011.tif13153
[0092] On the other hand, the interval between the selected communication pairs is L g If it is less than [n], even if you try to find it using Equation 10, the effect of the frequency deviation will not appear, and the frequency deviation will be 1, that is, there will be no deviation. Therefore, in order to observe the frequency deviation from the transmission and reception timing of information communication, L g It is necessary to select a pair of information communications separated by a time interval of at least [n]. In other words, the interval between timing m and timing n must be at least a predetermined time (a predetermined time has elapsed), that is, the minimum time L at which the influence of frequency deviation appears on the propagation time. g It is more than [n] time.
[0093] L g [n] also changes from moment to moment and cannot be calculated directly, but if the allowable frequency deviation of the clock 20 is e, the absolute value of the apparent frequency deviation between devices using the same type of clock 20 will not exceed |2e|. g [n]=T N / |2e| is set as the selection criterion for the pair of information and communication. From the information and communication at timing m and timing n selected by this criterion, two equations (10) are obtained, and from these two equations, the frequency ratio r b / a [n,m] can be found. TIFF0007725311000012.tif15156
[0094] From the information communication at timing m and timing n selected by this criterion, two equations are obtained based on equation (10). One equation is for timing n, and is the t b,RThe other equation (10)' is an equation for timing m, and is the equation for t where n in the above equation (10) is m. b,R [m]=t a,r [m]+g b / a [m]+d b / a [m]. Therefore, subtract equation (10)' from equation (10), that is, t b,R [n]-t b,R [m], as shown in equation (12), the frequency ratio r b / a An equation can be created to find [n,m].
[0095] However, the latter term of the numerator in equation (12) is the propagation time fluctuation (d b / a [n]-d b / a [m]) cannot be observed directly and is included as an error in the frequency deviation. b / a [n]-d b / a If [m]) becomes zero and is cancelled out, the ideal frequency deviation can be obtained.
[0096] In order to minimize this error, equation (13) is obtained from the two equations (10) and (10)' that give equation (12). TIFF0007725311000013.tif17154
[0097] As shown in Figure 10, the time kL at which the influence of the frequency deviation does not appear relative to the timing m g Information communication is performed m'+1 times by timing m+m', which is separated by an interval of less than (the above-mentioned predetermined time). In other words, communication is performed in a short burst. Here, k is a positive coefficient that can be set according to the required accuracy of the time synchronization system. Similarly, there is a time kL at which the influence of frequency deviation does not appear with respect to timing n. g Information is communicated n'+1 times by timing n+n', which is separated by a time interval of 0.001. In other words, communication is performed in short bursts.
[0098] From the set of information communications related to timing m, one information communication timing is taken out and called m *Let n be the timing of information communication taken out from the set of information communication at timing n. * When formula (13) is calculated using these information communications, the fluctuation of the time difference (g b / a [n * ]-g b / a [m * ]) can be considered to take the same value. In other words, the latter term (g b / a [n * ]-g b / a [m * ]) is a constant value and can be considered not to fluctuate. Therefore, if the right-hand side of equation (13) can be minimized, the first half of the left-hand side (d b / a [n]-d b / a Here, the right side of equation (13) is the difference between the reception interval time and the transmission interval time.
[0099] Hereinafter, the first half of the left side of equation (13) will be referred to as the d difference (the difference due to fluctuations in propagation time), and the second half will be referred to as the g difference (the difference due to fluctuations in time difference). Here, the d difference and the g difference cannot be observed separately by simply using the transmission and reception timing of information communication. Generally, in an asynchronous state, the g difference exceeds the d difference, and the g difference and the d difference become equal midway through synchronization, and once synchronization is stable, the g difference falls below the d difference. Therefore, we will virtually classify the magnitude relationship between the d difference and the g difference and organize the error in equation (12) due to the d difference.
[0100] If the absolute value of the g difference is greater than or equal to the absolute value of the d difference, for example, if their ratio is 10 or more, then any m * and n * Even if d is selected, Equation (13) takes a constant value with an error of the order of d, and the frequency ratio can be calculated from Equation (12). In other words, the frequency deviation can be roughly adjusted.
[0101] On the other hand, if the absolute value of the d difference is larger than the absolute value of the g difference and is dominant, for example, if the ratio between them is 10 or more, all m * and n *If we find the (m´+1)(n´+1) combinations of equation (13) and select the combination that minimizes the right-hand side of the equation, we get |d b / a [n * ]-d b / a [m * ]| can also be considered to be minimum. That is, from equation (12), the frequency ratio r b / a [n * ,m * ] can be obtained with minimum error. In other words, the frequency ratio can be obtained based on the combination with the minimum error in propagation time.
[0102] However, if the absolute value of the d difference and the absolute value of the g difference are of similar magnitude, the operation of minimizing the right-hand side of equation (13) will bring equation (12) closer to 1 regardless of the actual frequency ratio, and it will be mistakenly assumed that there is no frequency deviation of the clocks 20 between the information communication devices 1.
[0103] Here, m' and n' are assumed to be approximately equal, and information communication is performed around timing m and timing n. In this case, the set {S b / a [m+k]|0≦k≦m'} and the set {S b / a It can be said that there is a correlation between statistical quantities such as the probability density distribution of {[n+k]|0≦k≦m'} as long as the characteristics of the communication path or the network devices on the communication path do not switch discontinuously. Note that the statistical quantities referred to here include, for example, probability density distributions as well as averages, medians, minimums, maximums, and quantities related to cross-correlation. Furthermore, the vicinity of timing m or n referred to here means, for example, within the above-mentioned predetermined time from timing m or timing n, that is, a time less than the minimum time Lg[n].
[0104] TIFF0007725311000014.tif11157That is, as shown in Figure 11, S b / a The probability density distribution of the set of [m+k] and S near timing n b / aAlthough there is a correlation between the probability density distributions of the set [n+k], due to the influence of clock frequency deviations between the information communication devices 1, b / a [n]-g b / a This shift causes a shift of [m]. Even if you find the minimum right-hand side of equation (13), |d b / a [n * ]-d b / a [m * ]| cannot be minimized.
[0105] To cancel this shift, S b / a Statistics f of the set of [*] s Let [*] be the x that solves the optimization problem (15), where U is a parameter depending on the time synchronization system. TIFF0007725311000015.tif32160
[0106] The calculated statistic f s [n], f s [m] difference (f s [n]-f s [m]) is the time difference fluctuation (g b / a [n]-g b / a [m]), if we transform equation (13) as in equation (16) to cancel the fluctuation of the time difference, we can find the n that minimizes the right-hand side. * ,m * By selecting the combination of |d b / a [n * ]-d b / a [m * ]| can also be considered as the minimum. That is, the frequency ratio r b / a [n * ,m * ] can be calculated with minimal error. TIFF0007725311000016.tif17162
[0107] In addition, directly (f s [n]-f s [m]), using equation (17), the frequency deviation e b / a [n,m] may be estimated. TIFF0007725311000017.tif15162
[0108] Furthermore, prior to obtaining these statistics, the continuity of the correlation may be verified. That is, S b / a It may be determined whether the cross-correlation function between the probability density distributions of the set of [*] has changed discontinuously. If it has changed discontinuously, it is considered that the characteristics of the communication path or the network device on the communication path have changed discontinuously, or that a change has occurred in the traffic of the communication path. In this case, the control for determining and correcting the frequency deviation is suppressed.
[0109] The synchronization control unit 80 can feedback control the clock oscillation frequency of slave device CLb or the drive frequency of clock 30 so that the frequency deviation calculated from equation (12) or equation (17) becomes 0. This enables frequency synchronization of slave device CLb with master device CLa, reducing errors in time calculation during time synchronization.
[0110] [Estimation of time difference under propagation time fluctuations] Next, the calculation of the time difference by the time difference calculation unit 79 will be described (step S106). b / a [n * ]-d b / a [m * Since the frequency deviation between the information communication devices 1 can be estimated by minimizing || and equation (12), the time difference between the clocks 20 of the information communication devices 1 can be kept constant by performing frequency synchronization. That is, the time difference vector g in equation (8) can be calculated by minimizing the absolute values of both elements (g b / a =g a / b ), the error in the propagation time component can be kept to a minimum. In the vector diagram of Figure 8, the time difference vector g is almost the same as the state where there is no error as shown in Figure 7.
[0111] Note that the transmission of synchronization information from timing m and n is performed within a predetermined time (L gAssuming that the frequency deviation calculation is performed within a short time period (less than [n]), the effect on the frequency deviation will be insignificant, and the error in the time difference will be minimal. Therefore, the above-described frequency deviation estimation and synchronization processing based on the frequency deviation are not essential. In other words, the time difference estimation described below is possible even without the above-described frequency deviation estimation and synchronization processing. Therefore, for example, it is possible to omit the frequency deviation calculation unit 78 of this embodiment. Note that, when the transmission of synchronization information from timings m and n is performed for a predetermined period of time or longer, it is possible to estimate the frequency deviation using other methods.
[0112] From the above, when calculating the synchronization state vector s using equation (8), only the fluctuation in propagation time needs to be considered. Here, factors that cause fluctuations in the propagation time of information communication include the packet communication volume of other communications on the network. Since packet communication time is discrete, the fluctuation in propagation time is also discrete. When changing the set of information communication and repeatedly calculating equation (8), the synchronization state vector s is calculated as g d -d g On a plane, they are arranged in a grid rotated by +π / 4 as shown in Figure 12.
[0113] In other words, assuming that the error of the time difference vector g is minimal, only the synchronization state vector s can be observed in a set of multiple data communications, and these plots are shown in Figure 12. Note that observing the synchronization state vector s corresponds to measuring the four transmission and reception timings and finding the left side of equation (6), i.e., the left side of equation (8), as described above. Bidirectional communications that allow the synchronization state vector s to be found using equation (8) are performed within a short period of time, i.e., within a predetermined period of time during which interruptions by other communications can be eliminated and frequency deviations do not become apparent. This period of time is, for example, less than the minimum time Lg[n]. This is because by performing data communications in bursts over a short period of time, it is possible to create many situations in which synchronization communications can be continued without interruptions by other communications.
[0114] When such information communication for time synchronization is performed continuously in a short period of time, there are times when the synchronization state vector s with a constant propagation time can be obtained, eliminating the waiting for turn due to multiple communications and not being affected by fluctuations in propagation time due to other communications. These times are shown in g in Figure 12. d It is a point on the dotted line perpendicular to the axis of
[0115] Such a set of synchronization state vectors s has the minimum error in the time difference component, and the propagation time can be considered symmetric. d approaches a constant value g, which is the estimated time difference.
[0116] However, to calculate the estimated time difference g, we simply use the element g of the synchronization state vector s. d However, simply finding the average of s does not allow accurate estimation because the arrangement of the synchronization state vector s is biased, as shown in FIG.
[0117] Therefore, for example, the element g of the synchronization state vector s[k] obtained from the set of consecutive information communications in a short period of time is d [k] and g d -d g g on a plane d The distance between the component variable x (g d [k]-x), the evaluation function Σk(g d [k]-x) 4 By solving the optimization problem to find x that minimizes the distance (g), we can find the estimated time difference g. d By summing up the fourth power of [k]-x, and giving a penalty to long distances and reducing the penalty to short distances, it is possible to find the line on which the largest number of synchronization state vectors s are concentrated. In the example of Figure 12, the value corresponding to the vertical dotted line (straight line) is the estimated time difference g. The equation for solving the optimization problem may be in the form of equation (15).
[0118] Synchronization control unit 80 of slave device CLb repeatedly adjusts the clock frequency or time so that the calculated time difference g becomes zero, thereby achieving time synchronization with master device CLa.
[0119] [Estimation of time difference in asymmetric communication paths] As shown in Figure 6 above, depending on the network system to which the master device CLa and slave device CLb are connected, the propagation time may be constantly asymmetric. This is caused by, for example, the difference in the amount of traffic between the upstream (slave → master) and downstream (master → slave) of communication.
[0120] The steady asymmetry of the propagation time in the two-way communication path constantly adds an error of the time difference component to the propagation time vector d in the above equation (8), resulting in an offset (deviation due to error) of the time difference component of the obtained synchronization state vector s. In other words, the asymmetry causes a steady deviation in the synchronization phase.
[0121] Regarding the asymmetry of the time difference, the minimum time L shown in the above equation (11) is within a predetermined time in which the frequency deviation of the clock 20 between the master device CLa and the slave device CLb does not appear. g If the synchronization state vector s is found in a time less than [n], the time difference between the upstream and downstream communications will be the same and can be considered symmetric. Furthermore, by performing a series of operations to solve the optimization problem shown in equation (15) above, such as synchronizing the frequencies of the clocks 30 between the devices, it is possible to control the denominator of equation (11) to be very small, thereby increasing the period during which the time difference is symmetric.
[0122] In order to accommodate such asymmetric communication paths, time difference calculation unit 79 calculates the time difference, and as shown in FIG. 13, information is transmitted in bursts from both master device CLa and slave device CLb. To ensure symmetry of the time difference, the information burst transmission is performed within a predetermined time, i.e., a time less than Lg[n]. While FIG. 13 shows burst transmissions being performed sequentially in each direction, burst transmissions can also be performed simultaneously from both directions (for example, within a common predetermined time). While FIG. 13 shows the number of burst transmissions in each direction as n'+1 from timing n to timing n+n', the number does not have to be the same for each direction. In other words, the number of transmissions may differ for master device CLa and slave device CLb.
[0123] Here, similarly to equation (14), the difference in transmission and reception timing in the uplink direction is expressed as equation (18). At this time, S obtained in the downlink burst transmission is b / a Let the set of [k] be S b / a (n,n´)={S b / a [n+k]|0≦k≦n´}, and S obtained by the upstream burst transmission a / b Let the set of [k] be S a / b (n,n´)={S a / b [n+k]|0≦k≦n´}.
[0124] TIFF0007725311000018.tif11158
[0125] By calculating the variance of these sets and selecting the direction of the time difference depending on the size of the variance, the time difference can be estimated with high accuracy. b / a Variance of (n,n´) VAR[S b / a (n,n')] is T b / a Below, and S a / b Variance of (n,n´) VAR[S a / b (n,n')] is T a / b If the time is less than the above, the time difference estimation method 1 will be selected as follows, otherwise the time difference estimation method 2 will be selected as described below. b / a and T a / bis a value determined according to the traffic of the communication path and the quality of time synchronization. b / a (n,n')] is T b / a Exceeding and variance VAR[S a / b (n,n')] is T a / b If it exceeds this, it can be said that the degree of propagation time asymmetry exceeds a predetermined standard.
[0126] In an environment where time difference estimation method 1 is selected, S b / a (n,n´) and S a / b The variation in (n,n') is small, and the traffic on the communication path is stable. In such an environment, we can select the synchronization state vector s with the smallest propagation time component and estimate the time difference component as the time difference.
[0127] That is, S b / a (n,n´) and S a / b Extract elements one by one from (n,n') and perform the operation of finding the synchronization state vector from equation (8) for all combinations of elements to find a set S of synchronization state vectors s in equation (19) below. Then, select the synchronization state vector s with the smallest propagation time component from set S.
[0128] TIFF0007725311000019.tif18159
[0129] In an environment where time difference estimation method 2 is selected, S b / a (n,n´) and S a / b The variation of (n,n') is large. In other words, the traffic on the communication path is unstable, and the propagation time may become asymmetric. In such an environment, it is necessary to estimate the asymmetry of the propagation time and remove the offset of the time difference component from the synchronization state vector s.
[0130] To do this, first, as shown in Figure 14, a set S of synchronization state vectors is calculated using equation (19) in the same way as in time difference estimation method 1, and its center of gravity c is calculated. Here, a=(g, 0) T When the center of gravity c and point a are d= a + t(ca) is called the propagation time baseline. d The component g is a variable.
[0131] Each element s of set S k and the propagation time baseline l d The propagation time baseline l such that the sum of the Euclidean distances between d To obtain this, the optimization problem of the following equation (20) is solved. TIFF0007725311000020.tif26158Note that Figure 14 is simplified to a simple set S, but S b / a (n,n´) and S a / b It can be said that the probability density distribution of (n, n') will generally be different in an environment where time difference estimation method 2 is selected. For this reason, even if the distribution shape of set S is rectangular, the calculated center of gravity c does not necessarily coincide with the center of the rectangle. Regardless of the distribution shape of set S, the center of gravity can be calculated as c using the calculation shown in the constraints of equation (20).
[0132] The propagation time baseline l obtained by solving the optimization problem of equation (20) d corresponds to a straight line that lies on the propagation time vector d, which is an element that makes up the synchronization state vector s. However, since the time difference vector g is added, it is shifted from the propagation time vector d.
[0133] That is, the propagation time baseline l obtained from equation (20) d g of a which is an element of d The propagation time baseline l d In cases where multiple values are obtained, the propagation time baseline l that minimizes ||a|| is d Just select.
[0134] By feedback controlling the clock oscillation frequency or clock drive frequency of slave device CLb so that the time difference calculated by these time difference estimation methods 1 and 2 becomes zero, it becomes possible to synchronize the time of slave device CLb with master device CLa, and errors due to asymmetry in propagation time can be reduced.
[0135] [effect] (1) This embodiment is an information communication system 100 in which a plurality of information communication devices 1 communicate information as master device CLa or slave device CLb, and includes an interval time calculation unit 76 that calculates a transmission interval time and a reception interval time for a first set of information communications that are a plurality of information communications performed between master device CLa and slave device CLb within a predetermined time with respect to a first timing, and a second set of information communications that are a plurality of information communications performed within a predetermined time with respect to a second timing that is the predetermined time or more from the first timing; a selection unit 77 that selects a set of information communications from the first set of information communications and the second set of information communications that has the smallest difference between the reception interval time and the transmission interval time; a frequency deviation calculation unit 78 that calculates a deviation in clock frequency between master device CLa and slave device CLb based on the set of information communications selected by the selection unit 77; and a synchronization control unit 80 that synchronizes the clock frequencies or times of master device CLa and slave device CLb based on the deviation in clock frequency calculated by the frequency deviation calculation unit 78.
[0136] Furthermore, this embodiment is an information communication device 1 that communicates information with another information communication device 1, and includes an interval time calculation unit 76 that calculates a transmission interval time and a reception interval time for a first set of information communications that are multiple information communications performed within a predetermined time relative to a first timing, and a second set of information communications that are multiple information communications performed within a predetermined time relative to a second timing that is the predetermined time or longer from the first timing; a selection unit 77 that selects a set of information communications from the first set of information communications and the second set of information communications that has the smallest difference between the transmission interval time and the reception interval time; a frequency deviation calculation unit 78 that calculates a deviation in clock frequency between master device CLa and slave device CLb based on the set of information communications selected by the selection unit 77; and a synchronization control unit 80 that synchronizes the clock frequencies or times of master device CLa and slave device CLb based on the clock frequency deviation calculated by the frequency deviation calculation unit 78.
[0137] In this way, by selecting a set of information notifications with the smallest difference based on the transmission interval and reception interval, which can be calculated from the transmission time and reception time observable at different times, it is possible to calculate the frequency deviation with minimized error. Therefore, even in an environment where propagation time fluctuates from moment to moment due to frequency fluctuations, it is possible to absorb the propagation time fluctuations, evaluate the symmetry of propagation time, and achieve synchronization control with reduced error. This enables highly accurate time synchronization without using specific network relay devices that support time synchronization, such as BC and TC in the PTP time synchronization system. In other words, it is possible to use the PTP time synchronization system while eliminating the need for BC and TC, and achieve the desired accuracy required by the system.
[0138] (2) The predetermined time is the minimum interval time at which the influence of frequency deviation appears in the information propagation time between master device CLa and slave device CLb. As a result, the influence of frequency deviation does not appear in the first set of information communications and the second set of information communications, but the influence of frequency deviation appears in the pair of the first and second information communications. Therefore, by selecting the pair of information communications that minimizes the difference between the transmission interval time and the reception interval time, it is possible to determine the frequency deviation that minimizes the error.
[0139] (3) The frequency deviation calculation unit 78 calculates the clock frequency deviation by canceling the fluctuation of the time difference based on the difference between the statistics based on the set of information communications within a predetermined time from the first timing and the statistics based on the set of information communications within a predetermined time from the second timing. This enables accurate synchronization control even when the difference due to the fluctuation of the time difference is as large as the difference due to the fluctuation of the propagation time.
[0140] (4) The frequency deviation calculation unit 78 calculates the deviation of the clock frequency based on the difference between a statistic based on a set of information communications within a predetermined time period relative to the first timing and a statistic based on a set of information communications within a predetermined time period relative to the second timing. Therefore, the frequency deviation can be calculated directly from the difference in the statistic.
[0141] (5) A time difference calculation unit 79 is provided that calculates the time difference corresponding to the largest set of multiple sets of information communications between master device CLa and slave device CLb within a specified time period, among sets with the smallest time difference error. More specifically, the time difference calculation unit 79 calculates the time difference corresponding to the line on which the largest number of synchronization state vectors, each of which has elements of the time difference and propagation time in the set of multiple information communications, are gathered. This makes it possible to calculate a time difference with symmetry in propagation time and the smallest error, even in an environment where propagation time fluctuates from moment to moment, enabling precise synchronization control.
[0142] (6) When the degree of asymmetricity in propagation times in two-way communications exceeds a predetermined standard, the time difference calculation unit 79 can also calculate the time difference corresponding to the line passing through the center of gravity of a set of synchronization state vectors, whose elements are time differences and propagation times, for a set of multiple communications. This makes it possible to calculate the time difference with the smallest error, even over asymmetric communication paths, enabling precise synchronization control.
[0143] [Other embodiments] The present invention is not limited to the above-described embodiment, and various inventions can be realized by modifying the components within the scope of the gist of the present invention. Furthermore, various inventions can be realized by appropriately combining the multiple components disclosed in the above-described embodiment. [Explanation of symbols]
[0144] 1. Information and communication devices 10. Communications Department 11 Transmitter 12 Receiver 13 Transmission timing detection unit 14 Reception timing detection unit 20 Clock 30 Clock 50 Storage section 60 external interfaces 70 Control Unit 71 Main control unit 72 Sending and receiving data I / F 73 Communication control section 74 Scheduler 75 Time Recording Unit 76 Interval time calculation unit 77 Selection section 78 Frequency deviation calculation section 79 Time difference calculation unit 80 Synchronization control section 81 Timing information reference section 100 Information and Communication Systems
Claims
1. An information communication system in which a plurality of information communication devices communicate information as master devices or slave devices, an interval time calculation unit that calculates a transmission interval time and a reception interval time for each pair of information communications from the first set of information communications and information communications from the second set of information communications, the first set of information communications being multiple information communications performed within a predetermined time from a first timing between the master device and the slave device, and a second set of information communications being multiple information communications performed within the predetermined time from a second timing that is equal to or longer than the predetermined time from the first timing; a selection unit that selects, from the first set of information communications and the second set of information communications, a set of information communications that minimizes a difference between the reception interval time and the transmission interval time; a frequency deviation calculation unit that calculates a deviation in clock frequency between the master device and the slave device based on the set of information communications selected by the selection unit; a synchronization control unit that synchronizes the clock frequencies or times of the master device and the slave device based on the clock frequency deviation calculated by the frequency deviation calculation unit; and An information communication system, wherein the predetermined time is the minimum interval time at which the influence of frequency deviation appears in the propagation time of information between the master device and the slave device.
2. 2. The information and communication system according to claim 1, wherein the frequency deviation calculation unit cancels fluctuations in time difference and calculates the deviation of the clock frequency based on the difference between a statistical quantity relating to a set of information communications within the predetermined time period relative to the first timing and a statistical quantity relating to a set of information communications within the predetermined time period relative to the second timing.
3. 2. The information and communication system according to claim 1, wherein the frequency deviation calculation unit calculates the deviation of the clock frequency based on a difference between a statistical quantity relating to a set of information communications within the predetermined time period relative to the first timing and a statistical quantity relating to a set of information communications within the predetermined time period relative to the second timing.
4. 4. An information and communication system according to claim 1, further comprising a time difference calculation unit that calculates the time difference corresponding to the line along which the greatest number of synchronization state vectors, each of which has a time difference and propagation time as elements, are gathered in a set of a plurality of information communications between the master device and the slave device within a predetermined time period, thereby minimizing the time difference error in the set of the plurality of information communications and corresponding to the greatest set.
5. The information and communication system according to claim 4, characterized in that, when the degree of asymmetry in propagation times in two-way communication exceeds a predetermined standard, the time difference calculation unit calculates the time difference corresponding to a straight line passing through the center of gravity of a set of synchronization state vectors, whose elements are the time difference and propagation time, in the set of multiple information communications.
6. An information communication device that communicates information with another information communication device, an interval time calculation unit that calculates a transmission interval time and a reception interval time for each pair of information communications from the first set of information communications and information communications from the second set of information communications, in a first set of information communications that is a plurality of information communications performed within a predetermined time with respect to a first timing, and a second set of information communications that is a plurality of information communications performed within the predetermined time with respect to a second timing that is equal to or longer than the predetermined time from the first timing; a selection unit that selects, from the first set of information communications and the second set of information communications, a set of information communications that minimizes a difference between the transmission interval time and the reception interval time; a frequency deviation calculation unit that calculates a deviation of a clock frequency between the information communication device and another information communication device based on the set of information communication selected by the selection unit; a synchronization control unit that synchronizes a clock frequency or a time based on the clock frequency deviation calculated by the frequency deviation calculation unit; and The predetermined time is the minimum interval time at which the influence of frequency deviation appears in the propagation time of information between the information communication device and another information communication device.
7. 7. The information communication device according to claim 6, wherein the frequency deviation calculation unit cancels fluctuations in time difference and calculates the deviation of the clock frequency based on a difference between a statistical quantity relating to a set of information communications within the predetermined time period relative to a first timing and a statistical quantity relating to a set of information communications within the predetermined time period relative to a second timing.
8. 7. The information communication device according to claim 6, wherein the frequency deviation calculation unit calculates the deviation of the clock frequency based on a difference between a statistical quantity relating to a set of information communications within the predetermined time period relative to a first timing and a statistical quantity relating to a set of information communications within the predetermined time period relative to a second timing.
9. An information communication device according to any one of claims 6 to 8, characterized in that it has a time difference calculation unit that calculates the time difference corresponding to the line along which the largest number of synchronization state vectors, each of which has time difference and propagation time as elements, are gathered in a set of multiple information communications within a predetermined time period, thereby minimizing the time difference error in the set of multiple information communications and corresponding to the largest set.
10. The information communication device according to claim 9, characterized in that, when the degree of asymmetry in propagation time in two-way communication exceeds a predetermined standard, the time difference calculation unit calculates the time difference corresponding to a straight line passing through the center of gravity of a set of synchronization state vectors, whose elements are the time difference and propagation time, in the set of multiple information communications.
Citation Information
Patent Citations
Telecommunication system
JP2009071645A
Time synchronization method and time synchronization device
JP2013030892A
Slave device, master / slave system and time synchronization method
JP2015004649A
Time synchronization device, time synchronization method, and time synchronization program
JP2016225880A
Telecommunication system and information communication device
JP2021005773A