JP7604753B2 - Time synchronization method, time synchronization program, and time synchronization system - Google Patents

Description

This invention relates to a time synchronization method, a time synchronization program, and a time synchronization system for time-synchronizing different devices.

Network standards such as Ethernet TSN (Time-Sensitive Networking) prescribe a method for synchronizing time between devices connected to an Ethernet (registered trademark). In this time synchronization, each device connects to a time server that holds standard time and adjusts its time to the "time server."

Individual devices include PCs and servers connected to Ethernet, as well as in-vehicle devices installed in vehicles and connected to a CAN (Controller Area Network), home appliances connected to an ECHONET (Energy Conservation and HOMEcare NETwork) (registered trademark), and factory equipment that operates by executing a control program in a PLC (Programmable Logic Controller), etc.

Conventionally, there are technologies related to time synchronization via a network, such as those described in the following Patent Documents 1 and 2. Patent Document 1 stabilizes clock synchronization by counting a frequency clock when an NTP (Network Time Protocol) server device broadcasts an NTP to a client terminal device and determining whether or not to use a reference time to correct time information. Patent Document 2 synchronizes the time of home appliances and the like from an external communication network using the home appliance time synchronization protocol in accordance with the ECHONET standard.

JP 2018-098711 A JP 2008-152343 A

However, in conventional technology, some devices were unable to synchronize time depending on the connection status of each device on the network, as described below.

Figure 10 is an explanatory diagram of problem 1 of the prior art. Figure 10 shows an example configuration in which devices B to H are connected to time server A via a network such as Ethernet. In the case of time synchronization using time server A, if time server A breaks down as shown in Figure 10, devices B to H will no longer be able to synchronize their time with time server A. Furthermore, devices B to H that are no longer connected to time server A due to a network disconnection or the like will also no longer be able to synchronize their time.

Figure 11 is an explanatory diagram of problem 2 of the prior art. Usually, many networks contain not only Ethernet but also slower networks such as CAN. Figure 11 shows that devices B, C, and D are connected to time server A via a fast connection such as Ethernet, but devices E to G have slower connections. Here, the mixture of communication methods on the network, for example devices connected via Ethernet and devices connected via CAN, use different communication methods, and time synchronization cannot be easily achieved.

In the example shown in Figure 11, device E is connected to device B, and devices E to H are not directly connected to time server A. In this case, only device group GP (devices B to D) directly connected to time server A can synchronize time, and devices E to H that are not directly connected to time server A cannot synchronize time.

In view of the above problems, the present invention aims to enable time synchronization of multiple devices on a network without connecting to a time server.

In order to achieve the above object, the time synchronization method of the present invention is a time synchronization method for synchronizing the times of multiple devices connected to a network, in which a source device transmits time information including the accuracy of its own clock to a destination device connected for communication via the network, the destination device selects the source device which transmitted the time information with the highest accuracy of the clock from the multiple pieces of time information it received, and corrects the time based on the time information transmitted by the selected source device. This process is repeated among devices which can communicate within the network , and the source device checks the accuracy of the clock, the time when the time information was transmitted, and the time from when the time information was transmitted to when it is received by the destination device. and a predicted arrival time at the destination device, the destination device calculates a difference between the time on its own clock when it receives the time information transmitted by the selected source device and the predicted arrival time, corrects the time on its own clock based on the difference, and transmits response time information to the source device in response to the request time information, the response time information including the elapsed time from the time the request time information was received to the time the response time information was transmitted and the transmission time of the response time information, the source device updates and retains the predicted arrival time upon receiving the response time information, and repeatedly transmits the time information to the destination device with which it can communicate .

The source device is further characterized in that it includes the time information in a packet that transmits data to the destination device.

The sending device is also characterized in that it stores the time information in the header of the packet and transmits it.

The network also includes multiple different communication methods, and time is synchronized through data communication between devices using communication methods that allow communication with each other.

The network also includes in-vehicle communication, synchronizing the time between the devices within the vehicle and the devices outside the vehicle, and between the devices within the vehicle and each other.

A time synchronization program of the present invention is a time synchronization program for synchronizing the times of a plurality of devices connected to a network, the program causing a source device to transmit time information including the accuracy of its own clock to a destination device connected for communication via the network, causing the destination device to select the source device which transmitted the time information with the highest accuracy of the clock from among the plurality of pieces of time information received, and correcting the time based on the time information transmitted by the selected source device. The program causes the source device to repeatedly perform a process among devices which can communicate within the network, and causes the source device to obtain the accuracy of the clock, the time when the time information was transmitted, and the time from when the time information was transmitted to when it is received by the destination device. the destination device calculates a difference between the time on its own clock when it receives the time information transmitted by the selected source device and the estimated arrival time, corrects the time on its own clock based on the difference, and causes the source device to transmit response time information to the request time information, the response time information including the elapsed time from the time the request time information was received to the time the response time information was transmitted and the transmission time of the response time information, the source device updates and retains the estimated arrival time upon receiving the response time information, and repeatedly transmits the time information to the destination device with which communication is possible .

The time synchronization system of the present invention is a time synchronization system for synchronizing the times of a plurality of devices connected to a network, in which a source device has a control unit that transmits time information including the accuracy of its own clock to a destination device connected for communication via the network, and the destination device has a control unit that selects the source device that transmitted the time information with the highest accuracy of the clock from among the plurality of pieces of time information received, and corrects the time based on the time information transmitted by the selected source device, and repeats this process between devices that can communicate within the network , and the control unit of the source device controls the accuracy of the clock, the time when the time information was transmitted, and the time when the destination device receives the time information after transmitting it. a control unit of the destination device calculates a difference between the time of the clock of the destination device when it receives the time information transmitted by the selected source device and the estimated arrival time, corrects the time of the clock of the destination device based on the difference, and transmits response time information to the source device in response to the request time information, the response time information including an elapsed time from the time of receiving the request time information to the time of transmission of the response time information and a transmission time of the response time information, the control unit of the source device updates and holds the estimated arrival time upon receiving the response time information, and repeatedly transmits the time information to the destination device with which it can communicate .

According to the above configuration, it is possible to synchronize the time between devices that are connected to each other for communication based on the time information. Furthermore, even if the network includes various different communication methods, it is possible to synchronize the time of devices in the entire network by synchronizing the time between devices that can communicate. It is also possible to synchronize the time of devices in the network without disposing a clock server.

The present invention has the effect of enabling multiple devices on a network to synchronize their time without connecting to a time server.

FIG. 1 is a diagram illustrating an example of a configuration of a time synchronization system according to an embodiment. FIG. 2 is a diagram illustrating an example of node connections in the time synchronization system according to the embodiment. FIG. 3 is a block diagram of an example of functions of the time synchronization system according to the embodiment. FIG. 4 is a diagram illustrating an example of a hardware configuration of a node according to the embodiment. FIG. 5 is a flowchart illustrating an example of a control process of the time synchronization system according to the embodiment. FIG. 6 is a diagram showing the exchange of time information used in the embodiment. FIG. 7 is a diagram showing variations caused by existing time adjustment techniques. FIG. 8 is a diagram illustrating a convergence state of time variations due to time synchronization according to the embodiment. FIG. 9A is a diagram showing an example of control between existing devices. FIG. 9B is a diagram showing an example of control between devices based on existing time synchronization. FIG. 9C illustrates an example of control between devices based on time synchronization according to the embodiment. FIG. 10 is an explanatory diagram of problem 1 of the prior art. FIG. 11 is an explanatory diagram of problem 2 of the prior art.

(Embodiment)
Preferred embodiments of a time synchronization method, a time synchronization program, and a time synchronization system according to the present invention will be described in detail below with reference to the accompanying drawings.

Figure 1 is a diagram showing an example of the configuration of a time synchronization system according to an embodiment. As shown in Figure 1, multiple devices 100 that perform time synchronization are communicatively connected to each other via various networks.

For example, in the device 100, the smartphone 100a is communicatively connected to the PC 100b via the Internet 111. The PC 100b is communicatively connected to the control board 100c via Ethernet 112a. The PC 100b is communicatively connected to the control board 100d via Ethernet 112b. The control board 100d is communicatively connected to the control board 100c via Ethernet 112c.

Control board 100c is communicatively connected to control board 100e via CAN 113. Control board 100e is communicatively connected to control board 100f via serial communication 114.

Of the devices 100 shown in Figure 1, PC 100b and control boards 100c and 100d perform time synchronization over Ethernet based on the IEEE 802.1AS communication standard, for example. As shown in Figure 1, in the embodiment, there is no need to place a time server, as described in the prior art, on the system. Note that in the embodiment, a specific device, such as PC 100b, that can communicate with a time server may be placed on the system.

The time synchronization system according to the embodiment has the following features.
1. No time server is required.
2. There is no need to obtain standard time from a time server, etc.
3. Time synchronization (time matching) is possible even when different communication methods are mixed within a system or when devices 100 are not directly connected to each other in a network.
4. Even if a device 100 in the network breaks down and is lost, or if a new device 100 is connected to the network, time synchronization can be maintained.

Figure 2 is a diagram showing an example of node connections in a time synchronization system according to an embodiment. In Figure 2, multiple devices 100 are shown as nodes A to H. The time synchronization system according to an embodiment has the following configurations. In the figure, solid lines indicate fast connections between nodes with a communication speed equal to or higher than a predetermined speed, and dotted lines indicate slow connections with a communication speed below a predetermined speed.

1. Each node A to H (device 100) on the network sends time information to the other node when normally exchanging data with the other node. The time information includes the following information a to c:
a: Your time (node time)
b: The accuracy of your clock (node accuracy)
c: The time (taking into account transmission delays) when the data arrives at the other party (estimated arrival time)

The clock accuracy is, for example, the clock accuracy of a node. In the time synchronization system of the embodiment, there is no need to use time-only packets. When a node sends a packet without data to a partner node, the packet is a time-only packet. This allows each node A to H to obtain time information sent from its partner nodes A to H with which it is connected in communication, without the node having to actively go and get the time.

2. Overview of time synchronization 1) A node that has just been turned on and has not yet determined its time acquires one or more pieces of time information according to the number of packets sent from other nodes within a certain period of time. For example, node E receives packets from nodes B, F, and H that are directly connected to it.

2) Node E uses the time information from node B, which has the most accurate clock among the packets received from nodes B, F, and H, as the basis for its own time.

3) Node E returns time information, including the elapsed time until the time was reached, to node B that adopted the time information.
4) Node B, which has received the time information, calculates (updates) the time information to be sent next to node E based on the time elapsed since the transmission time of the previously sent packet. This calculation of the time information is performed using a method that uses the predicted arrival time.

Here, node E responds with time information to node B, and this responded time information includes the following d and e.
d: transmission time of the time information responded by the destination node E e: the elapsed time from when the source node E received the time information from the source node B until the transmission time

5) Node E, which has synchronized its time, can also send time information to the surrounding nodes B, F, and H with which it is connected in communication.

According to the above, even if time information from some of the nodes (e.g., node B) among the nodes B, F, and H connected to node E is interrupted, node E can receive time information from the other nodes F and H. Then, node E can adjust the time of its own node E (time synchronization) using the time information received from the other nodes F and H, thereby improving the resistance to communication failures that occur on unspecified communication paths on the network.

In addition, node E synchronizes the time with other nearby nodes B, F, and H, making it possible to synchronize the time even if there is no time server on the network, and also enabling time synchronization of nodes A to H throughout the system.

Also, assume that there is one highly accurate node (for example, node A) and all nodes B to H are directly connected to this highly accurate node A. In this case, each node B to H will synchronize its time with node A, resulting in time synchronization accuracy similar to that of Ethernet TSN.

Here, even if a failure occurs in node A, each node B to H will be able to automatically adopt the time information of the node with the next highest accuracy (nodes C to H) to synchronize time, which allows for flexible time synchronization that cannot be achieved with Ethernet TSN.

Figure 3 is a block diagram showing an example of the functions of a time synchronization system according to an embodiment. Figure 3 shows devices 100 for a pair of time-synchronized source node B and destination node E. In the network connection example of Figure 1, nodes B, F, and H that are communicatively connected to destination node E are source nodes.

The source node B includes a timer 301, a transmission time acquisition unit 302, a time information processing unit 303, and a communication unit 304. The timer 301 measures time information. The transmission time acquisition unit 302 acquires the transmission time for transmitting time information to the destination node E based on the measurement information of the timer 301.

The time information processing unit 303 generates time information to be transmitted to the destination node E. The time information includes the above a: node time, b: clock accuracy, and c: predicted arrival time. The predicted arrival time can be calculated as an initial value based on the communication speed of the network between the source node B and the destination node E, but can be updated to a more accurate value that takes into account propagation delays by exchanging time information between the source node B and the destination node E.

The communication unit 304 communicates with the source node B and the destination node E using a communication method corresponding to the network between them.

In addition, when the time information processing unit 303 receives a response of time information from the destination node E, it updates the predicted arrival time previously transmitted to the destination node E based on this time information. The responded time information includes: d: the transmission time of the time information transmitted by the destination node E; and e: the elapsed time from when the source node B received the time information from the source node B until the transmission time.

The time information processing unit 303 of the source node B updates and retains the predicted arrival time included in the time information for the destination node E. The time information processing unit 303 includes the updated arrival time in the time information when transmitting the time information to the destination node E next time.

The destination node E includes a communication unit 311, a reception time acquisition unit 312, a time information processing unit 313, and a timer 314. The communication unit 311 communicates with the destination node E and the source node B using a communication method corresponding to the network between them.

The reception time acquisition unit 312 acquires the time information received from the source node B. The time information processing unit 313 acquires from the timer 314 the reception time at which the time information received from the source node B was received.

The time information processing unit 313 determines whether or not to adopt the time information received from the source node B based on the "clock accuracy" of the source node B, and when adopting the information, corrects the time of its own node E according to the difference from the predicted arrival time contained in the time information.

Here, destination node E receives time information from each of source nodes B, F, and H with which it is directly connected, and selects the source node (node B) with the highest clock accuracy contained in the time information received from source nodes B, F, and H, and synchronizes its own node E with the time of source node B.

Then, the time information processing unit 313 corrects the advance/delay of the time measured by the timer 314 in the own node E in order to synchronize the own node E with the time of the source node B. The time information processing unit 313 also transmits time information to the source node B. The time information includes the transmission time of this time information and the elapsed time from when the time information is received from the source node B to the transmission time.

In Figure 3, for convenience, the functions of the source node and the destination node are shown separately. In reality, each node (device 100) has the functions of a source node and a destination node. Therefore, a given node (device 100) has the functions of the source node and the destination node shown in Figure 3. Also, the communication units 304, 311 in Figure 3 can be arranged as one function per node (device 100), and the time information processing units 303, 313 can also be arranged as one function per node (device 100).

Figure 4 is a diagram showing an example of the hardware configuration of a node according to an embodiment. The devices 100 constituting the above-mentioned nodes A to H can be configured with the hardware shown in Figure 4. In Figure 4, each of the devices 100 includes a control unit (CPU) 401, a read-only memory (ROM) 402, a random access memory (RAM) 403, an auxiliary storage unit 404 such as a semiconductor memory or a disk drive, a communication interface (I/F) 405, and an input/output I/F 410. These CPU 401 to input/output I/F 410 are each connected by a bus 406.

The device 100 may have a display, keyboard, mouse, scanner, and printer corresponding to the device configuration of a PC, smartphone, various boards, etc., which can be connected via the input/output I/F 410. Note that in the device 100 such as a smartphone or tablet, a touch panel that combines the functions of a display, keyboard, and mouse can also be used.

The CPU 401 is an arithmetic processing device that controls the device 100 and, in the embodiment, is responsible for time synchronization. The ROM 402 is a non-volatile memory that stores the programs of the device 100. The RAM 403 is a volatile memory that is used as a work area when the CPU 401 executes arithmetic processing of the programs.

The communication I/F 405 controls the network NW and the internal interface, and performs input and output of data between the other devices 100. Specifically, the communication I/F 405 performs communication with the other devices 100 via various networks NW such as LAN (Local Area Network), CAN, and ECHONET through a communication line.

In addition to the above, the network NW may include Wi-Fi (registered trademark), Bluetooth (registered trademark), infrared communication (IrDA), etc., and may communicate wirelessly/wired with other devices 100.

Figure 5 is a flowchart showing an example of control processing of a time synchronization system according to an embodiment. Figure 5 shows an example of control processing executed by each control unit (CPU 401) of source node B (100) and destination node E (100) in accordance with the description of Figure 2 as an example of processing of the source node and destination node devices 100.

To explain the control process on the source node B side, the source node B obtains the transmission time when transmitting time information to the destination node E (100) from the timer 301 (step S501), and transmits the time information (step S502). The time information includes the above a: node time, b: clock accuracy, and c: expected arrival time.

After this, the source node B waits for a response of time information from the destination node E, and upon receiving the time information (step S503), it updates the estimated arrival time of node E that it has stored based on d: transmission time and e: elapsed time contained in the returned time information (step S504). This update allows the source node B to obtain an estimated arrival time that takes into account the propagation delay between nodes B and E, and improves the accuracy of the estimated arrival time contained in the time information to be transmitted to node E next time.

By exchanging the above series of processes with node E, the source node B can successively update the estimated arrival time, which corresponds to the time synchronization processing time at node E and the fluctuation over time in the transmission characteristics of the network between nodes B and E, thereby improving the accuracy of time synchronization.

Next, the control process on the side of the destination node E will be described. When the destination node E receives time information from the source node B (step S511), it obtains the reception time of the time information from the timer 314 of the node E (step S512).

Next, destination node E determines whether to accept the time information received from source node B (step S513). Destination node E receives time information from nodes B, F, and H, which are directly connected to it. In step S513, destination node E compares the "clock accuracy" contained in the time information of each of nodes B, F, and H, and selects the node with the highest "clock accuracy." In the example of Figure 5, source node B is selected.

Then, the destination node E decides to adopt the time information of the source node B (step S513: Yes) and executes the process of step S514 and onward. On the other hand, the destination node E decides not to adopt the time information of the source nodes F and H (step S513: No) and proceeds to the process of step S517.

In step S514, the destination node E calculates the difference between the reception time acquired in step S512 and the estimated arrival time included in the clock information (step S514). The destination node E then corrects the time (advance/delay) of the clock (timer 314) of the destination node E based on the calculated difference (step S515). For example, the destination node E corrects the time of the clock based on the slew/step correction value of the NTP time synchronization (step S516).

After this, the destination node E responds with the time information to the source node B (step S517). This response time information includes: d: the transmission time of the time information transmitted by the destination node E; and e: the elapsed time from when the destination node E received the time information from the source node B until the transmission time.

In this way, the destination node E can synchronize and correct its own time with the source node B, which has the highest "clock accuracy," based on the time information received from the (adjacent) source nodes B, F, and H with which it is directly connected in communication.

Furthermore, after time synchronization, node E can perform time synchronization on other destination nodes (nodes F and H in the example of Figure 2) that are directly connected as source node E, making it possible to time synchronize all nodes A to H in the network. Here, since directly connected nodes can communicate with each other using a specified communication method, by repeating the above process, it is possible to time synchronize the nodes (devices 100) throughout the entire network, which uses a mixture of various communication methods.

Figure 6 is a diagram showing the exchange of time information used in the embodiment. Figure 6 shows the source node B and destination node E (device 100) mentioned above. When source node B transmits specified data (packet) to destination node E, it includes time information in the packet and sends it. The time information may be stored as part of the packet header, or may use a specified area of the packet data.

The source node B transmits time information (notification data) including the above a: node time (T1), b: clock accuracy, and c: expected arrival time (ExpectTime) to the destination node E at the timing of time T1.

Destination node E receives time information (notification data) from source node B at time T2. Destination node E uses the received time information (T1 + ExpectTime) to correct the time.

After this, at timing T3, destination node E responds to source node B with the above d: the transmission time of the time information transmitted by destination node E, and e: time information (notification data) including the elapsed time from when destination node E received the time information from source node B to the transmission time.

The source node B receives time information (notification data) responded from the destination node E at timing T4.

The estimated arrival time ES is calculated from the timing of the request and the response. For example, the estimated arrival time ES between nodes B and E can be calculated based on the following formula (1).
((T4-T1)-(T3-T2))/2)...(1)

Here, the predicted arrival time is calculated between source node B and destination node E, regardless of whether or not the time is corrected at destination node E. For this reason, destination node E includes in the time information it responds to source node B the difference between the system time at the time of receiving the time information transmitted from source node B (the time measured by timer 314) and the system time at the time of transmitting the time information it responds to (e: elapsed time above).

The source node B transmits (requests) time information to the destination nodes E, F, and H with which it is directly connected in communication, for example at regular periodic intervals, and the destination nodes E, F, and H transmit (respond) time information to the source node B for each request.

Here, when the destination node B transmits a packet of predetermined data to the destination node E, it includes time information in the header of the packet to be transmitted, and the destination node E extracts the time information from the received packet. Also, when the source node B does not transmit a packet of data to the destination node E for a certain period of time, it generates a packet for time information and transmits it to the destination node E.

Destination node E extracts the time information from the received packet. This enables destination node B to transmit time information at regular intervals to nodes E, F, and H that are directly connected to it, enabling time synchronization at regular intervals.

According to the embodiment, the time of the entire network can be synchronized even if the "clock accuracy" of each node A to H in the network is different. When time synchronization begins, there is a large deviation (variation), but after a predetermined time (e.g., about 20 msec has passed), the time of nodes A to H in the entire network can be synchronized to a reference time, for example, node A, which has the highest "clock accuracy" (the standard deviation of the time of each node B to H is minimized).

Figure 7 shows the variation caused by existing time correction technology. The horizontal axis is time, and the vertical axis is each node A to D (device 100). Node A has the highest clock accuracy, nodes B and C have slightly poorer clock accuracy, and node D has the lowest clock accuracy.

Regarding time correction for multiple existing devices 100, for example in FA (Factory Automation), there is time synchronization by "periodic processing", where each device A to D performs processing, for example, every 10 msec. This "periodic processing" does not synchronize the time between each of nodes A to D, so smooth control between nodes A to D may not be possible due to the different clock accuracy of each of nodes A to D. In the example of Figure 7, node D, which has the lowest clock accuracy, is delayed by one cycle (10 msec) from other nodes (for example, node C) after a certain period of time has passed.

Figure 8 is a diagram showing the convergence state of time variation due to time synchronization in an embodiment. Figure 8 (a) shows nodes A to D. By exchanging time information between each of directly connected nodes A to D and implementing time synchronization control at regular intervals, the time (time period) of each of nodes A to D can be kept within a certain variation, enabling smooth control without time variation between each of nodes A to D.

Furthermore, as shown in Figure 8 (b) (a configuration corresponding to an embodiment), by performing time synchronization between nodes A to D using node A, which has a "high clock accuracy," as a reference, it is possible to correct the clock accuracy of nodes B to D even more than in Figure 8 (a).

(Example of control by time synchronization)
Next, an example of control between existing devices and control between devices by time synchronization according to the embodiment will be described with reference to Figures 9A to 9C. These figures show an example of application to cooperative control of devices in an autonomous vehicle.

Figure 9A is a diagram showing an example of control between existing devices. Devices 1 to 4 are connected to a central control unit (sensor) via CAN or the like. For example, device 1 is an accelerator, device 2 is a brake, device 3 is a steering wheel, and other in-vehicle devices necessary for autonomous driving.

In existing technology, devices 1 to 4 operate individually at regular intervals and transmit detection information (control plan information) to sensors, but in the entire autonomous driving system, they operate independently. For this reason, existing autonomous driving systems have poor maneuverability and ride comfort.

Figure 9B is a diagram showing an example of control between devices using existing time synchronization. In contrast to the configuration in Figure 9A, this shows an example in which an in-vehicle sensor and devices 1 to 4 are connected via a network with a time synchronization function such as Ethernet TSN. In this case, a precise clock is used with Ethernet TSN, and control plan information is sent from the sending side (sensor) to each of devices 1 to 4 at regular intervals.

For example, the sensor can prevent communication congestion by controlling the timing of data acquisition and the timing of sending control plan information to each device 1 to 4 at different times. This allows for smoother control and improved ride comfort. However, if the network includes different networks (CAN), some devices will only be able to communicate via CAN, and these devices will not be able to synchronize time.

Figure 9C is a diagram showing an example of control between devices using time synchronization according to an embodiment. In time synchronization according to an embodiment, as in Figure 9B, for example, a sensor can prevent communication congestion by controlling the timing of data acquisition and the timing of sending control plan information to each of devices 1 to 4 at different times.

Furthermore, in the time synchronization according to the embodiment, it is possible to synchronize the time of each of the devices 1 to 4. As a result, in the embodiment, it is possible to perform control that synchronizes the control timing of the devices 1 to 4 of the entire autonomous driving system while suppressing congestion in data transmission and reception, and it is possible to perform autonomous driving with the smoothest and most comfortable (smooth) control.

The time synchronization described above can be applied to various systems. For example, it can be applied to the above-mentioned autonomous driving. The device 100 installed in this autonomous driving vehicle uses a large number of system-on-chips (SoCs), microcomputers, etc., which work in conjunction with each other. For example, when autonomous driving is performed using camera recognition, the SoC etc. determines the behavior of the vehicle based on information such as camera images, vehicle speed, accelerator opening, braking state, steering angle, and obstacle detection by sensors such as radar.

By synchronizing the time of multiple devices 100, the control timing of the steering, accelerator, and brakes can be aligned, enabling smooth driving (see FIG. 9C). Time synchronization according to the embodiment can be controlled within the vehicle, and multiple devices 100 can be time-synchronized without the need for a time server.

In addition, the time synchronization of the embodiment can be applied to FA failure analysis. When a failure occurs in a system on-site such as a factory, the time synchronization is performed on the device 100 that controls the data collection device for failure analysis using the video and control log of the failed device. By using multiple microcontrollers and multiple cameras as the device 100 and performing time synchronization between these devices 100, it is possible to collect data (information) with the same time stamp. The collected data includes video from multiple cameras, operation logs at the time of failure, etc. In this application example, a time server can be eliminated, and even if a situation occurs in which the network is cut off or some of the devices 100 fail, the remaining devices 100 can be synchronized in time, making the system resistant to failures.

According to the embodiment described above, in the time synchronization method for synchronizing the time between multiple devices connected to a network, the source device transmits time information including the accuracy of its own clock to the destination device connected to the network, and the destination device selects the source device that transmitted the most accurate time information from the received multiple pieces of time information, and corrects the time based on the time information transmitted by the selected source device. This process is repeated between devices that can communicate within the network. As a result, time synchronization can be performed with the device with the most accurate clock based on the time information between the devices connected to each other. Furthermore, even if various different communication methods are included in the network, by synchronizing the time between the devices that can communicate with each other, it becomes possible to synchronize the time of the entire network, including devices that are not directly connected to each other. Furthermore, it becomes possible to synchronize the time of the devices within the network without disposing a clock server. Furthermore, since the destination device receives time information from multiple source devices, even if a source device in the network breaks down and is lost, or if a new device (the source device and the destination device) is connected to the network, each device can be time synchronized.

The source device may also include time information in the packet that is sent to the destination device. This makes it easy to transmit time information using packets between devices that can communicate with each other.

The sending device may also store time information in the header of the packet before sending it. This allows time information to be easily exchanged between devices without affecting data communication between them.

The source device may also include time information in a data packet sent to the destination device at the timing of sending the time information, and if no data is sent to the destination device for a certain period of time, generate and send a packet for the time information. This makes it possible to synchronize the time at a specified periodic interval while minimizing congestion of packets exchanged between devices.

The network also includes multiple different communication methods, and time is synchronized through data communication between devices using communication methods that can communicate with each other. In this way, time synchronization is performed between devices using the same communication method, and also between devices using communication methods that can communicate with each other, making it possible to synchronize the time of devices across the entire network using different communication methods.

The network may also include in-vehicle communications, and may synchronize the time between devices inside the vehicle and devices outside the vehicle, and between devices inside the vehicle. For example, devices such as a control unit, accelerator, brake, and steering wheel equipped in an autonomous driving vehicle can be time-synchronized with devices outside the vehicle, enabling time synchronization without accessing a time server.

The source device transmits requested time information including the accuracy of its clock, the time when the time information was transmitted, and the estimated time of arrival from when the time information was transmitted until it was received by the destination device. The destination device calculates the difference between the time on its own clock when it received the time information transmitted by the selected source device and the estimated time of arrival, corrects the time on its own clock based on the difference, and transmits response time information to the source device in response to the requested time information. The response time information includes the elapsed time from when the request time information was received until the response time information was transmitted, and the transmission time of the response time information. Upon receiving the response time information, the source device updates and retains the estimated time of arrival, and repeatedly transmits the time information to the destination device with which it can communicate. This enables time synchronization including the propagation state between the source device and the destination device, and enables accurate time synchronization between devices with different communication methods, communication speeds, processing times, etc. Furthermore, even if there are fluctuations in communication speeds and processing times over time, time synchronization can be achieved in response to these fluctuations.

From these points, according to the embodiment, time information can be exchanged between devices that are directly connected to each other for time synchronization. In addition, even if devices that are not directly connected to each other in the network or devices that communicate using different communication methods are included, time synchronization can be performed for other devices by propagating time synchronization between devices that can be connected to each other, and multiple devices in the network connection can be time synchronized. In addition, devices in the network can be time synchronized without disposing a clock server. In addition, since the destination device receives time information from multiple source devices, it can flexibly respond to device failures and the addition of new devices in the network, and each device can be time synchronized. Note that the present invention can be similarly applied even if a specific device is connected to a time server for communication.

As described above, the present invention can be used in a time synchronization system that synchronizes the time between devices connected via a network, and is particularly useful for time synchronization between devices that are connected to each other and communicate with each other via various networks.

100 Device 301, 314 Timer 302 Transmission time acquisition unit 303, 313 Time information processing unit 304, 311 Communication unit 312 Reception time acquisition unit 314 Timer 401 Control unit (CPU)
402 ROM
403 RAM
404 Auxiliary storage unit 405 Communication interface 406 Bus A to H Nodes

Claims (8)

A time synchronization method for synchronizing the times of a plurality of devices connected to a network, comprising the steps of:
The sending device is
Transmitting time information including the accuracy of the clock of the device to each of the destination devices communicatively connected via the network;
The destination device is
Selecting the source device that has transmitted the time information having the highest clock accuracy from among the plurality of received time information;
correcting the time based on the time information transmitted by the selected source device;
The process is repeated among devices capable of communicating within the network;
The source device is
Transmitting requested time information including the accuracy of the clock, the transmission time of the time information, and an expected arrival time from when the time information is transmitted until when the time information is received by the destination device;
The destination device is
calculating a difference between the time of the clock of the device itself when the time information transmitted from the selected transmission source device is received and the predicted arrival time;
correcting the time on the clock of the own device based on the difference;
Transmitting time information in response to the time information of the request to the source device;
the time information of the response includes an elapsed time from the time the time information of the request was received to the time the time information of the response was transmitted, and a transmission time of the time information of the response;
The source device is
Upon receiving the time information of the response, the expected arrival time is updated and stored;
repeatedly transmitting the time information to the destination device with which communication is possible;
A time synchronization method comprising: The source device is
transmitting the time information to the destination device by including it in a packet for data transmission;
The time synchronization method according to claim 1 . the source device stores the time information in a header of the packet and transmits the packet;
The time synchronization method according to claim 2 . The source device is
transmitting the time information included in a packet of data to the destination device at a timing to transmit the time information;
if no data is transmitted to the destination device for a certain period of time, a packet for the time information is generated and transmitted;
The time synchronization method according to claim 1 . The network includes a plurality of different communication methods, and time is synchronized by data communication between the devices of the communication methods that can communicate with each other.
The time synchronization method according to claim 1 . The network includes in-vehicle communication, and synchronizes time between the devices in the vehicle and the devices outside the vehicle, and between the devices in the vehicle and each other.
The time synchronization method according to claim 1 . A time synchronization program for synchronizing the times of a plurality of devices connected to a network, comprising:
On the sending device,
transmitting time information including the accuracy of the clock of the device to each of the destination devices communicatively connected via the network;
The destination device,
selecting the source device that has transmitted the time information having the highest clock accuracy from among the plurality of pieces of time information received;
correcting the time based on the time information transmitted by the selected source device;
Repeating the process among devices that can communicate within the network;
The source device ,
transmitting requested time information including the accuracy of the clock, the transmission time of the time information, and the expected arrival time from when the time information is transmitted until when it is received by the destination device;
The destination device ,
calculating a difference between the time of the clock of the device itself when the time information transmitted from the selected transmission source device is received and the predicted arrival time;
correcting the time on the clock of the own device based on the difference;
causing the source device to transmit time information in response to the time information of the request ;
the time information of the response includes an elapsed time from the time the time information of the request was received to the time the time information of the response was transmitted, and a transmission time of the time information of the response;
The source device ,
Upon receiving the time information of the response, the predicted arrival time is updated and stored;
causing the communication destination device to repeatedly transmit the time information;
A time synchronization program comprising: A time synchronization system for synchronizing the times of multiple devices connected to a network, comprising:
The sending device is
a control unit for transmitting time information including the accuracy of a clock of the device to each of destination devices connected in communication with the device via the network;
The destination device is
Selecting the source device that has transmitted the time information having the highest clock accuracy from among the plurality of received time information;
correcting the time based on the time information transmitted by the selected source device;
A control unit that repeatedly performs processing among devices that can communicate within the network,
The control unit of the source device
Transmitting requested time information including the accuracy of the clock, the transmission time of the time information, and an expected arrival time from when the time information is transmitted until when the time information is received by the destination device;
The control unit of the destination device
calculating a difference between the time of the clock of the device itself when the time information transmitted from the selected transmission source device is received and the predicted arrival time;
correcting the time on the clock of the own device based on the difference;
Transmitting time information in response to the time information of the request to the source device;
the time information of the response includes an elapsed time from the time the time information of the request was received to the time the time information of the response was transmitted, and a transmission time of the time information of the response;
The control unit of the source device
Upon receiving the time information of the response, the expected arrival time is updated and stored;
repeatedly transmitting the time information to the destination device with which communication is possible;
A time synchronization system comprising:

Images