JP2013029340A - Time correction device, time correction method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a time correction device, a time correction method, and a program for exactly performing time correction on the basis of upload information from a sensor.SOLUTION: A time correction device has: a receiving part; a clock part; a communication time calculation part; a model holding part; a model update part; and a time correction part. The receiving part receives information including transmission time based on internal time of an external device. The clock part holds reference time. The communication time calculation part calculates communication time to receiving time when information based on the reference time is received from the transmission time. The model holding part holds a communication time distribution model in which distribution of the communication time is predicted. The model update part updates a communication time distribution model on the basis of the calculated communication time. The time correction part corrects deviation of the transmission time on the basis of the updated communication time distribution model.

Description

  The present invention relates to a time correction device, a time correction method, and a program.

  In recent years, various sensor data can be accumulated due to the downsizing and cost reduction of sensors and the development of networks, and there is an increasing demand for using such sensor data for various solutions.

  On the other hand, the number and types of sensors connected to a network and uploading information to a server are increasing. For example, there is a device that functions as a sensor by installing software in another person's device, such as a personal computer (PC) usage status collection tool. In this way, when software is installed in another person's device, the system settings cannot basically be changed. Also, the types of terminals and the size of resources vary from PCs to portable terminals. In such a situation, it is time-consuming to manage the sensors one by one, and there is a desire to keep each sensor as small as possible during operation.

  However, in order for the server to handle data from various sensors in an integrated manner, it is desirable that the internal time difference between the server and the sensor be as small as possible. For this reason, I would like to automatically correct the internal time of each sensor to match the server.

  As an example of such automatic correction of the internal time, there is a technique in which the sensor transmits to the server, and the server returns the internal time at the time of reception, thereby measuring and correcting the difference between the internal time of the sensor and the server. Are known. As another example, a method using a Flooding Time Synchronization Protocol (FTSP) is known.

JP 2008-209995 A

The Flooding Time Synchronization Protocol Proceedings of the 2nd ACM Conference on Embedded Networked Sensor Systems (SenSys’04), Baltimore, Maryland (2004), by M. Maroti, B. Kusy, G. Simon and A. Ledeczi.

  However, in the method in which the server replies to the communication from the sensor, two-way communication is necessary, and the time is corrected on the sensor side, so it is necessary to put a hand on the sensor side. In addition, since the method using FTSP uses a special protocol for wireless communication, the media access control (MAC) layer on the sensor side must be touched to take a time stamp.

  In this way, it is troublesome to put a hand in each sensor and correct it. Therefore, it is desirable that communication between the sensor and the server can correct the time lag only by uploading sensor information (sensor internal time and sensor value) without using a special synchronization protocol. However, since the way in which the internal time of each sensor advances varies, it is difficult to correct the internal time of the sensor with high accuracy in communication only by uploading sensor information.

  In view of the above problems, a time correction apparatus, a time correction method, and a program capable of performing accurate time correction based on upload information from a sensor are provided.

  The time correction apparatus which is one aspect has a receiving part, a time measuring part, a communication time calculation part, a model holding part, a model update part, and a time correction part. The reception unit receives information including a transmission time based on an internal time of an external device. The timekeeping unit holds the reference time. The communication time calculation unit calculates the communication time from the transmission time to the reception time when the information based on the reference time is received. The model holding unit holds a communication time distribution model in which the communication time distribution is predicted. The model updating unit updates the communication time distribution model based on the calculated communication time. The time correction unit corrects the transmission time deviation based on the updated communication time distribution model.

  In the time correction method according to another aspect, information including a transmission time based on an internal time of a device external to the time correction device is received, and the information based on a reference time in the time correction device is received from the transmission time. The communication time until the received time is calculated. The communication time distribution model in which the communication time distribution is predicted is updated based on the communication time, and the transmission time deviation is corrected based on the updated communication time distribution model.

  In addition, even if it is a program for causing a computer to perform the method according to the present invention described above, since the program is executed by the computer, the same operations and effects as the method according to the present invention described above are achieved. The aforementioned problems are solved.

  According to the time correction apparatus, the time correction method, and the program of the above-described aspect, it is possible to correct the time accurately based on the upload information from the sensor.

It is a block diagram which shows the structure of the time correction system by one embodiment. It is a block diagram which shows the function of the server in the time correction system by one Embodiment. It is a figure which shows an example of distribution with which communication time by one embodiment is estimated. It is a figure which shows an example of the estimated shift | offset | difference of the internal time by one Embodiment. It is a figure which shows an example of the relationship between the transmission time by one embodiment, and communication time. It is a flowchart which shows operation | movement of the time correction system by one Embodiment. It is a flowchart which shows operation | movement of the time correction system by one Embodiment. It is a figure which shows an example of the communication time log | history by one embodiment. It is a figure which shows an example of the sensor information by one embodiment. It is a figure which shows an example of the communication time log | history by one embodiment. It is a figure which shows an example of the sensor information after the time correction by one Embodiment. It is a figure which shows an example of the communication time log | history after the NTP generation | occurrence | production estimated time by one Embodiment is reset. It is a figure which shows an example of a structure of a standard computer.

  Hereinafter, a time correction system according to an embodiment will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the time correction system 1. As shown in FIG. 1, the time correction system 1 is a system in which a server 3, at least one sensor 20, and a network time protocol (NTP) server 37 are connected via a network 35. The server 3 is an information processing apparatus that acquires various measurement values from at least one sensor 20 via the network 35. The sensor 20 is a measuring device that acquires various measurement values and transmits them to the server 3 via the network 35. The network 35 is a communication network such as the Internet or a Local Area Network.

  The server 3 includes a central processing unit (CPU) 5, a storage device 9, a clock 11, an input / output device 13, and a communication device 15, which are connected to each other via a bus 17.

  The CPU 5 is an arithmetic processing device that controls the operation of the entire server 3. The storage device 9 includes, for example, a storage device such as a read only memory (ROM), a random access memory (RAM), and a hard disk and a driving device thereof. The storage device 9 is a storage unit for storing various control programs such as a program for controlling the operation of the server 3 in advance or using it as a work area as necessary when executing the program. The CPU 5 can perform various control processes by reading and executing a predetermined control program recorded in the storage device 9.

  The clock 11 is a time measuring device that holds time (hereinafter referred to as reference time) by the NTP server 37. The clock 11 preferably updates the time by accessing the NTP server 37 at predetermined time intervals, for example. The time interval for performing this update is set to be short enough for the timepiece 11 to hold the reference time, so that the timepiece 11 holds the reference time.

  The input / output device 13 is a general term for an input device and an output device. The input device is a device that, when operated by the user of the server 3, acquires input of various information from the user associated with the operation content, and sends the acquired input information to the CPU 7. Keyboard device, mouse device, etc. The output device is a device that outputs a processing result by the server 3, and includes a display device and the like. For example, the display device displays text and images according to display data sent by the CPU 5.

  The communication device 15 is an interface device that manages transmission / reception of various types of data performed externally by wire or wireless. The bus 17 is a communication path for connecting the above devices and the like to exchange data.

  The sensor 20 includes a storage device 21 including, for example, a ROM and a RAM, a CPU 23, a detection unit 25, a clock 27, and a communication device 29, which are connected to each other via a bus 31. The sensor 20 is a measuring device that performs various measurements and transmits a measurement value and a transmission time via the network 35.

  The CPU 23 is an arithmetic processing unit that controls the operation of the entire sensor 20. The storage device 21 is a storage device for storing, for example, various control programs for controlling the operation of the sensor 20 in advance or using them as a work area as necessary when executing the programs. The CPU 23 can perform various control processes by reading and executing a predetermined control program recorded in the storage device 21.

  The detection unit 25 is a sensor that detects various measurement values. The communication device 29 is an interface device that manages transmission / reception of various types of data performed externally by wire or wireless. The bus 31 is a communication path for connecting the above devices and the like to exchange data.

  FIG. 2 is a block diagram illustrating functions of the server 3 in the time correction system 1. As shown in FIG. 2, the CPU 5 of the server 3 has functions of a communication time calculation unit 41, a model update unit 43, an occurrence probability calculation unit 45, an NTP time update unit 47, and a time correction unit 49. The storage device 9 includes a communication time history storage unit 51 and a communication time model storage unit 53.

  The communication time calculation unit 41 calculates the communication time from the transmission time included in the sensor information 61 received from the sensor 20 by the communication device 15 and the reception time when the information is received. Further, the communication time calculation unit 41 stores the calculated result in the communication time history storage unit 51.

  The model update unit 43 updates the communication time model calculated by the communication time calculation unit 41. Further, the model update unit 43 stores the updated model in the communication time model storage unit 53. The model will be described later.

  The occurrence probability calculation unit 45 is calculated by the communication time calculation unit 41 based on the model updated by the model update unit 43 and stored in the communication time model storage unit 53, and stored in the communication time history storage unit 51. Calculate the occurrence probability of time.

  The NTP time update unit 47 determines that the NTP time has been updated in the sensor 20 when the occurrence probability calculated by the occurrence probability calculation unit 45 is equal to or less than a predetermined threshold, and the communication time history storage unit 51 and the communication time model storage. The unit 53 updates the NTP time. Hereinafter, the time estimated that the NTP time is updated is referred to as the NTP generation estimated time.

  The time correction unit 49 determines that the NTP time is not updated when the occurrence probability calculated by the occurrence probability calculation unit 45 is greater than a predetermined threshold, and corrects the internal time. In addition, the time correction unit 49 outputs the corrected time-corrected sensor information 63 and updates the communication time history S.

The communication time history storage unit 51 holds a communication time history S. The communication time history S holds sensor information 61 and communication time uploaded after the previous estimated NTP generation time t _NTP . The sensor information 61 is information including the measurement value detected by the detection unit 25 of the sensor 20 and the transmission time of the sensor information 61 from the sensor 20 to the server 3 (transmission time under the internal time of the sensor 20). .

  Here, a model used for correcting the internal time will be described with reference to FIGS. Consider a set of sensors 20 and a server 3. First, consider the communication time x from the sensor 20 to the server 3 when it is assumed that the internal time of the sensor 20 coincides with the internal time of the server 3. The communication time x is considered to be the sum of the “optimum case communication time” and the “communication time noise” whose length is determined probabilistically.

FIG. 3 is a diagram illustrating an example of a predicted distribution of communication time, and FIG. 4 is a diagram illustrating an example of a predicted deviation in internal time. At this time, the distribution of the communication time x is predicted to follow the probability distribution p (x). In FIG. 3, the horizontal axis represents the communication time x, and the vertical axis represents the probability distribution p (x) of the communication time x. As shown in FIG. 3, the predicted communication time distribution can be expressed by Equation 1 below.
p (x) = c (γ) exp (− (x−δ) / γ) (x ≧ δ) (Expression 1)
However,
δ: a parameter indicating the communication time in the optimum case γ: a parameter indicating the degree of variation in communication time noise c (γ): a coefficient for setting the sum of p (x) to 1. In this model, c (γ) = 1 / γ. At this time, it can be easily calculated that the sum of p (x) is always 1.

Next, consider the deviation y (t) of the internal time of the sensor with respect to the server 3 at time t (here, t ≧ t _TNP ). In FIG. 4, the horizontal axis represents the internal time t of the sensor 20, and the vertical axis represents the deviation y (t) of the internal time t of the sensor 20 with respect to the internal time of the server 3. As shown in FIG. 4, the deviation y (t) is 0 when NTP occurs in the sensor (t = t _NTP ), and is considered to be proportional to the elapsed time from NTP. it can.
y (t) = α (t−t _NTP ) (Formula 2)
t _NTP : Last NTP occurrence time α: Parameter indicating how the time lag of the sensor's internal time spreads

Finally, consider the communication time (difference between the reception time of the server 3 and the transmission time of the sensor 20) z measured by the server 3. Here, the reception time of the server 3 is measured based on the internal time of the server 3, and the transmission time of the sensor is measured based on the internal time of the sensor 20. Therefore, the communication time z is considered to be the sum of “the communication time in the optimum case”, “the communication time noise”, and “the deviation of the internal time of the sensor relative to the server”. Therefore, the communication time z follows the probability distribution q (t, z).
q (t, z) = 1 / γ · exp (− (z−αt−β)) / γ) (z ≧ αt + β) (Equation 3)
However, β = δ−αt _NTP .

The outline of time correction using this model is as follows. That is, the communication time history S of the communication time z measured by the server 3 is applied to the probability distribution q (t, z) of Equation 3, and the occurrence probability is calculated. When the occurrence probability becomes sufficiently small, that is, when the communication time history S does not fit in the probability distribution q (t, z), the time is regarded as the NTP occurrence time t _NTP at which the sensor 20 has generated _NTP . If the NTP generation time t _NTP is known, the deviation of the internal time of the sensor 20 relative to the internal time of the server 3 can be estimated by α (t−t _NTP ), so that the internal time of the sensor 20 can be corrected.

  FIG. 5 is a diagram illustrating an example of the relationship between the transmission time t (which is the time based on the internal time t of the sensor 20 and is represented by the same symbol) and the communication time z. In FIG. 5, the horizontal axis represents transmission time t, and the vertical axis represents communication time z. The line graph 72 connects points (t (i), z (i)) representing the transmission time t (i) (i is an integer representing the data transmission order) and the corresponding communication time z (i). It is a graph.

  As shown in FIG. 5, the communication time z suddenly fluctuates greatly with respect to the transmission time t, for example, at points P1 and P2, and it is considered that the NTP time has been updated. Further, between the points P1 and P2, an envelope 70 (represented by a broken line) connecting the minimum values of the communication time z indicates a relationship close to a straight line.

The points P1 and P2 are considered to indicate that the internal time is corrected by accessing the NTP server 37 inside the sensor 20 as described above. At this time, the internal time of the sensor 20 and the server 3 are corrected. The internal time is considered to be in agreement. Further, it is considered that the internal time of the sensor 20 changes at a constant rate with respect to the internal time of the server 3 between the points P1 and P2. That is, as described above, the communication time z during this period is considered to be expressed by the following equation.
z = α (t−t _NTP ) + δ (Expression 4)
Therefore, at this time, the difference between the internal time of the sensor 20 and the reference time held in the server 3 is estimated to be α (t−t _NTP ).

  Next, the operation of the time correction system 1 according to the present embodiment will be described with reference to FIGS. 6 and 7 are flowcharts showing the operation of the time correction system 1. In the time correction system 1, it is assumed that sensor information 61 is irregularly uploaded from the sensor 20 to the server 3.

  As shown in FIG. 6, first, the CPU 5 empties the communication time history S in the communication time history storage unit 51. That is, the number of histories n = 0. The communication device 15 receives the sensor information 61 (S102). On the other hand, the CPU 5 repeats the determination until it is determined that the reception has been performed (S103: NO).

  When the CPU 5 determines that the sensor information 61 has been received by the communication device 15 (S103: YES), the communication time calculation unit 41 acquires the sensor information 61. The communication time calculation unit 41 calculates the communication time z based on the transmission time based on the internal time of the sensor 20 included in the sensor information 61 and the reception time in the server 3 (S104). Further, the communication time calculation unit 41 adds the acquired sensor information 61 and the calculation result to the communication time history S.

  Thereafter, the communication time history S holds n pieces of information, and the transmission time of the i-th uploaded information is t (i), and the communication time at that time is z (i). And The communication time calculation unit 41 measures the reception time u (i) under the internal time of the server 3 and calculates the communication time z (i) by z (i) = u (i) −t (i). To do.

FIG. 8 shows an example of the communication time history S when n = 2. As shown in FIG. 8, when n = 1, transmission time t (1) = 2011/4/1 0:00:00, and at this time, communication time z (1) = 200 (milliseconds). Here, NTP generation time t _NTP = t (1). When n = 2, transmission time t (2) = 2011/4/1 0:01:00, and at this time, communication time z (2) = 400 (milliseconds).

  FIG. 9 shows an example of sensor information 61. As illustrated in FIG. 9, the communication device 15 receives sensor information 61 including, for example, sensor value = 10000 and transmission time t (3) = 2011/4/1 0:02:00 at n = 3. To do. At this time, the communication time calculation unit 41 acquires the reception time of the sensor information 61 as reception time u (3) = 2011/4/1 0: 02: 00: 00.

  In this example, since the difference between the reception time and the transmission time is 300 milliseconds, z (3) = 300 (milliseconds), and the communication time calculation unit 41 stores the calculated result in the communication time history S. To do. FIG. 10 shows an example of the communication time history S when n = 3. As shown in FIG. 10, one line of data related to n = 3 is added to the communication time history S (S105).

  Here, the communication time calculation unit 41 determines whether or not n> 1 (S106). If n> 1 is not satisfied (S106: YES), the process returns to S102 and waits for reception of the next sensor information 61. If n> 1 (S106: YES), the CPU 5 advances the process to 107.

  Hereinafter, the model update unit 43 re-estimates the parameters α, β, and γ of the probability distribution q (t, z) of the communication time model. The parameters α, β, and γ can be estimated by the following method. First, in S107, the model update unit 43 uses, for example, points P1 to P2 based on a plurality of transmission times t (i) and the communication times z (i) calculated from the respective transmission times t (i) as in the line graph 72 of FIG. The slope obtained by the least square method is regarded as the parameter α.

  Here, the model update unit 43 determines whether NTP time update has already occurred (S121). Whether or not an NTP time update has occurred is determined by setting a flag indicating that an NTP time update has occurred when it is determined in S127, which will be described later, that the occurrence probability of the communication time history S is less than or equal to a threshold value. It is possible to discriminate them by recording such as.

For example, when it is determined that the NTP time update has been generated by the above flag or the like (S121: YES), the model update unit 43 determines that all x ( i) = Minimum value of z (i) −α (t (i) −t _NTP ) is represented by δ (Formula 6).

If the occurrence of the NTP time update is not estimated (S121: NO), the model updating unit 43 assumes that δ = 0 (S123), and updates the NTP occurrence estimated time t _{NTP according} to the following Expression 7. (S124, Formula 7).

  Further, the model update unit 43 obtains γ that maximizes the probability of occurrence of the probability distribution q (t, z) in i = 1 to n based on the obtained parameters α and δ. (S125, Formula 8).

However, D (i) = − (z (i) −α (t (i) −tNTP) −δ).
This γ can be easily obtained by the following equation (9).

For example, in the case of the communication time history S in FIG. 10, the following α, δ, and γ are obtained.
α = 0.05
δ = 200 (milliseconds)
γ = 50 (milliseconds)

Reference is now made again to FIG. As described above, z = α (t−t _NTP ) + δ represented by the parameters α, γ, and δ obtained by the model updating unit 43 is a straight line 70 in FIG. The example shown in FIG. 5 is a numerical example different from the example of the communication time history S of FIG.

On the straight line 70, at the point P1, the transmission time t = t _NTP and the communication time z = δ. Further, z = α (t−tNTP) + δ represents the sum of “the internal time difference of the sensor with respect to the server” and “the communication time in the optimum case”, and points (t (i), z (i) ) And the straight line 70 represent “communication time noise”.

  Next, the occurrence probability calculation unit 45 calculates the occurrence probability based on the probability distribution q (t (i), z (i)) in the communication time history S under the reestimated parameters α, β, γ. (S126). This occurrence probability can be obtained by the following calculation formula, for example.

  Here, since the maximum value of the probability distribution q (t, z) is 1 / γ, the maximum value of γq (t, z) is 1, and the probability distribution q (t (i), z (i) ), The probability of occurrence is less likely to be affected by the parameter γ and the number n of the sensor information 61. For example, in the case of the communication time history S of FIG. 10, 0.05 is obtained as the occurrence probability.

Next, the occurrence probability calculation unit 45 determines whether or not the obtained occurrence probability exceeds a predetermined threshold value (S127). When the occurrence probability exceeds the threshold (S127: YES), the time correction unit 49 subtracts α (t (i) −t _NTP ) from the transmission time t (i) included in the sensor information 61. Then, the transmission time t (i) of the sensor information 61 received from the sensor 20 is corrected (S129), and the processing returns to FIG.

For example, in the example of FIG. 10, when the threshold value in the NTP time update unit 47 is 0.01, the occurrence probability = 0.05, so that the NTP occurrence estimated time t _NTP is not reset. In this case, the time correction unit 49 subtracts α (t (i) −t _NTP ) = 100 (milliseconds) from the transmission time t (i), thereby transmitting the transmission time t (i of the sensor information 61 received from the sensor 20. ) And the corrected transmission time t (i) is output.

  FIG. 11 is a diagram illustrating an example of corrected sensor information 63 after time correction. As shown in FIG. 11, the time correction unit 49 corrects the transmission time by 100 (milliseconds). The time correction unit 49 outputs corrected sensor information 63 including transmission time 2011/4/1 0: 01: 59.900 and sensor value = 10000. Further, the CPU 5 updates the communication time history S of the communication time history storage unit 51 with the corrected sensor information 63 obtained by correcting the transmission time.

When the occurrence probability does not exceed the threshold value (S127: NO), the NTP time update unit 47 updates the NTP generation estimated time t _NTP to the transmission time t (i) of the latest sensor information 61 (S130). In this case, the time correction unit 49 does not have to correct the transmission time t (i) of the sensor information 61. At this time, it is presumed that the NTP time update has occurred in the sensor 20, and after deleting all but the last line of the communication time history S, the NTP generation estimated time t _NTP = t (1) is set again (S131).

FIG. 12 is a diagram illustrating an example of the communication time history S after the estimated NTP generation time t _NTP is reset. In the example of FIG. 10, if the threshold value is 0.1, the communication time history S is updated only to data corresponding to i = 1 as shown in FIG. In this example, t (1) = t _NTP = 2011/4/1 0:02:00. At that time, the sensor information 61 received from the sensor 20 becomes the output of the sensor internal time correction system as it is. Then, the CPU 5 returns the process to S102 of FIG.

  As described above, according to the time correction system 1 according to the present embodiment, the communication time of the sensor information 61 from the sensor 20 is “the deviation of the internal time of the sensor with respect to the server”, “the communication time of the optimal case” and A model that is the sum of “communication time noise” is adopted. The communication time calculation unit 41 calculates the communication time each time the server 3 receives the sensor information 61 from the sensor 20 and adds it to the communication time history S. In the communication time history S, sensor information 61 since the last occurrence of NTP is recorded. When the sensor information 61 from the sensor 20 is received, the model update unit 43 refers to the communication time history S and updates each parameter in the communication time distribution model.

  The occurrence probability calculation unit 45 calculates the occurrence probability for the calculated communication time using the updated communication time distribution model. If the occurrence probability is less than or equal to a predetermined threshold, it is assumed that NTP time update has occurred, and the NTP time update unit 47 updates the estimated NTP occurrence time. When the occurrence probability exceeds a predetermined threshold, the time correction unit 49 estimates the internal time shift of the latest sensor information 61 using each parameter of the communication time distribution model and the estimated NTP generation time.

  As described above, according to the time correction system 1, by transmitting only the sensor information 61 including the transmission time based on the internal time of the sensor 20 from the sensor 20 to the server 3, the transmission time of the sensor information 61 is changed to the server 3. It is possible to correct the time based on the internal time of In the time correction system 1, since the time can be corrected only by performing communication from the sensor 20 to the server 3, communication is not increased.

  In addition, it is not necessary to make any changes to the sensor 20 or perform management, and it is possible to estimate the shift time regardless of the number or type of the sensors 20. Further, for example, by using Equation 10, the time correction system 1 is able to perform the time without being affected by the network congestion situation when the sensor information 61 is transmitted from the sensor 20 to the server 3 or the number of data of the sensor information 61. Correction can be performed. As described above, according to the time correction system 1, it is possible to estimate and correct the deviation of the internal time between the sensor 20 and the server 3 by a simple method.

  Since it is possible to estimate the NTP time update of each sensor, there is a variation in how the internal time advances, and even if the time is suddenly advanced or delayed due to NTP, the internal time difference between the sensor 20 and the server 3 Can be estimated. Further, even if sensors that perform NTP time update and sensors that do not perform are mixed, transmission time correction according to each sensor can be performed.

  In the above embodiment, the communication device 15 is an example of a receiving unit, the clock 11 is an example of a time measuring unit, and the occurrence probability calculating unit 45 is an example of a probability calculating unit and a deviation estimating unit. The communication time model storage unit 53 is an example of a model holding unit, and the NTP time update unit 47 is an example of a time update unit.

  Here, an example of a computer that is commonly applied to cause a computer to perform the operation of the time correction method according to the above embodiment will be described. FIG. 13 is a block diagram illustrating an example of a hardware configuration of a standard computer. As illustrated in FIG. 13, a computer 300 includes a central processing unit (CPU) 302, a memory 304, an input device 306, an output device 308, an external storage device 312, a medium driving device 314, a network connection device, and the like via a bus 310. It is connected.

  The CPU 302 is an arithmetic processing unit that controls the operation of the entire computer 300. The memory 304 is a storage unit for storing in advance a program for controlling the operation of the computer 300 or using it as a work area when necessary when executing the program. The memory 304 is, for example, a random access memory (RAM), a read only memory (ROM), or the like. The input device 306 is a device that, when operated by a computer user, acquires various information input from the user associated with the operation content and sends the acquired input information to the CPU 302. Keyboard device, mouse device, etc. The output device 308 is a device that outputs a processing result by the computer 300, and includes a display device and the like. For example, the display device displays text and images according to display data sent by the CPU 302.

  The external storage device 312 is a storage device such as a hard disk, and stores various control programs executed by the CPU 302, acquired data, and the like. The medium driving device 314 is a device for writing to and reading from the portable recording medium 316. The CPU 302 can read out and execute a predetermined control program recorded on the portable recording medium 316 via the recording medium driving device 314 to perform various control processes. The portable recording medium 316 is, for example, a Compact Disc (CD) -ROM, a Digital Versatile Disc (DVD), a Universal Serial Bus (USB) memory, or the like. The network connection device 318 is an interface device that manages transmission / reception of various data performed between the outside by wired or wireless. A bus 310 is a communication path for connecting the above devices and the like to exchange data.

  A program that causes a computer to execute the time correction method according to the above embodiment is stored in, for example, the external storage device 312. The CPU 302 reads the program from the external storage device 312 and causes the computer 300 to perform a time correction operation. At this time, first, a control program for causing the CPU 302 to perform time correction processing is created and stored in the external storage device 312. Then, a predetermined instruction is given from the input device 306 to the CPU 302 so that the control program is read from the external storage device 312 and executed. The program may be stored in the portable recording medium 316.

  The present invention is not limited to the embodiments described above, and various configurations or embodiments can be adopted without departing from the gist of the present invention. For example, the communication time distribution model is not limited to the above as long as it is based on “deviation of the internal time of the sensor relative to the server”, “optimum case communication time”, and “communication time noise”, and may take other forms. . For example, the communication time probability distribution model is not necessarily limited to the distribution shown in the formula time correction system 1, and the communication time x is distributed in a range of a predetermined value δ or more, and as the communication time x increases. As long as the distribution decreases, a distribution represented by another formula may be used.

  The configurations of the server 3 and the sensor 20 are within the scope of the present invention as long as they have similar functions. The sensor 20 may be a sensor that does not perform NTP time update. In this case, it is always determined as “NO” in S121, and the time is corrected under the condition of δ = 0. In the above embodiment, the sensor has been described as an example. However, if the transmission time based on the internal time can be transmitted, the time correction of other devices other than the sensor, such as a general-purpose computer, is possible. Is also applicable.

Regarding the above embodiment, the following additional notes are disclosed.
(Appendix 1)
A receiving unit for receiving information including a transmission time based on an internal time of an external device;
A timekeeping section that holds a reference time;
A communication time calculation unit for calculating a communication time from the transmission time to the reception time when the information based on the reference time is received;
A model holding unit for holding a communication time distribution model representing a distribution of communication time with the external device;
A model updating unit for updating the communication time distribution model based on the calculated communication time;
Based on the updated communication time distribution model, a time correction unit that corrects the deviation of the transmission time;
A time correction apparatus comprising:
(Appendix 2)
Based on the communication time distribution model updated by the model update unit, a probability calculation unit that calculates the occurrence probability of the communication time calculated,
A time update unit that updates a generation time of a time update process that matches the internal time with the reference time in the external device when the occurrence probability is a predetermined value or less;
The time correction apparatus according to appendix 1, wherein:
(Appendix 3)
The probability calculation unit estimates that the time update process has been performed when the occurrence probability is equal to or less than a predetermined value,
The time update unit updates the time when the time update process was performed,
The probability calculation unit estimates that the time update process is not performed when the occurrence probability exceeds a predetermined value,
The time correction apparatus according to claim 2, wherein the time correction unit corrects the transmission time based on a time at which the previous time update process has occurred.
(Appendix 4)
The communication time distribution model is based on a first model that represents the internal time lag in the external device and a second model that represents the distribution of the communication time variation that does not include the internal time lag. The time correction apparatus according to Supplementary Note 3, characterized by:
(Appendix 5)
The communication time in the communication time distribution model does not include the internal time lag according to the first model, the communication time distribution according to the second model, the internal time lag and the communication time difference. The time correction apparatus according to appendix 4, wherein the time correction apparatus is a sum of the case and the true communication time.
(Appendix 6)
6. The time correction apparatus according to appendix 4 or appendix 5, wherein the communication time distribution according to the second model is due to noise whose length is stochastically determined and follows the probability distribution.
(Appendix 7)
The time correction apparatus according to any one of appendix 4 to appendix 6, wherein the shift of the internal time due to the first model is proportional to an elapsed time from the time when the time update process occurs.
(Appendix 8)
Receive information including the transmission time based on the internal time of the device outside the time correction device,
Calculating a communication time from the transmission time to the reception time when the information is received based on the reference time in the time correction device;
Update the communication time distribution model predicting the distribution of the communication time based on the communication time,
A time correction method for correcting a shift in the transmission time based on the updated communication time distribution model.
(Appendix 9)
Based on the communication time distribution model updated by the model update unit, to calculate the occurrence probability of the communication time calculated,
When the occurrence probability is less than or equal to a predetermined value, update the occurrence time of the time update process for matching the internal time with the reference time in the external device,
The time correction method according to Supplementary Note 8, wherein
(Appendix 10)
If the occurrence probability is less than or equal to a predetermined value, it is estimated that the time update process has been performed,
Update the time when the time update process was performed,
When the occurrence probability exceeds a predetermined value, it is estimated that the time update process is not performed,
The time correction method according to appendix 9, wherein the transmission time is corrected based on the time at which the immediately preceding time update process occurred.
(Appendix 11)
The communication time distribution model is based on a first model that represents the internal time lag in the external device and a second model that represents the distribution of the communication time variation that does not include the internal time lag. The time correction method according to appendix 10, characterized by:
(Appendix 12)
The communication time in the communication time distribution model does not include the internal time lag according to the first model, the communication time distribution according to the second model, the internal time lag and the communication time difference. The time correction method according to appendix 9, wherein the time correction method is a sum of the case and the true communication time.
(Appendix 13)
The time correction method according to appendix 11 or appendix 12, wherein the communication time distribution according to the second model is due to noise whose length is stochastically determined and follows the probability distribution.
(Appendix 14)
14. The time correction method according to any one of appendix 11 to appendix 13, wherein the shift of the internal time due to the first model is proportional to an elapsed time from the time when the time update process occurs.
(Appendix 15)
Receive information including the transmission time based on the internal time of the device outside the time correction device,
Calculating a communication time from the transmission time to the reception time when the information is received based on the reference time in the time correction device;
Updating a communication time distribution model representing a predicted distribution of the communication time with respect to the internal time based on the communication time;
A program for causing a computer to execute a process of correcting a shift in the transmission time based on the updated communication time distribution model.

1 Time correction system 3 Server 5 CPU
9 Storage Device 11 Clock 13 Input / Output Device 15 Communication Device 17 Bus 20 Sensor 21 Storage Device 23 CPU
25 Detection unit 27 Clock 29 Communication device 31 Bus 35 Network 37 NTP server 41 Communication time calculation unit 43 Model update unit 45 Occurrence probability calculation unit 47 NTP time update unit 49 Time correction unit 51 Communication time history storage unit 53 Communication time model storage unit 61 Sensor information 63 Corrected sensor information

Claims (9)

A receiving unit for receiving information including a transmission time based on an internal time of an external device;
A timekeeping section that holds a reference time;
A communication time calculation unit for calculating a communication time from the transmission time to the reception time when the information based on the reference time is received;
A model holding unit for holding a communication time distribution model representing a distribution of communication time with the external device;
A model updating unit for updating the communication time distribution model based on the calculated communication time;
Based on the updated communication time distribution model, a time correction unit that corrects the deviation of the transmission time;
A time correction apparatus comprising: Based on the communication time distribution model updated by the model update unit, a probability calculation unit that calculates the occurrence probability of the communication time calculated,
A time update unit that updates a generation time of a time update process that matches the internal time with the reference time in the external device when the occurrence probability is a predetermined value or less;
The time correction apparatus according to claim 1, further comprising: The probability calculation unit estimates that the time update process has been performed when the occurrence probability is equal to or less than a predetermined value,
The time update unit updates the time when the time update process was performed,
The probability calculation unit estimates that the time update process is not performed when the occurrence probability exceeds a predetermined value,
The time correction apparatus according to claim 2, wherein the time correction unit corrects the transmission time based on a time at which the immediately preceding time update process has occurred.   The communication time distribution model is based on a first model that represents the internal time lag in the external device and a second model that represents the distribution of the communication time variation that does not include the internal time lag. The time correction apparatus according to claim 3.   The communication time in the communication time distribution model does not include the internal time lag according to the first model, the communication time distribution according to the second model, the internal time lag and the communication time difference. The time correction apparatus according to claim 4, wherein the time correction apparatus is a sum of a case and a true communication time.   6. The time correction apparatus according to claim 4, wherein the communication time distribution according to the second model is due to noise whose length is determined stochastically, and follows the probability distribution.   The time correction apparatus according to claim 4, wherein the deviation of the internal time due to the first model is proportional to an elapsed time from the time when the time update process occurs. Receive information including the transmission time based on the internal time of the device outside the time correction device,
Calculating a communication time from the transmission time to the reception time when the information is received based on the reference time in the time correction device;
Update the communication time distribution model predicting the distribution of the communication time based on the communication time,
A time correction method for correcting a shift in the transmission time based on the updated communication time distribution model. Receive information including the transmission time based on the internal time of the device outside the time correction device,
Calculating a communication time from the transmission time to the reception time when the information is received based on the reference time in the time correction device;
Updating a communication time distribution model representing a predicted distribution of the communication time with respect to the internal time based on the communication time;
A program for causing a computer to execute a process of correcting a shift in the transmission time based on the updated communication time distribution model.