JP2023119770A - Distribution control device and distribution control method - Google Patents

Abstract

[Problem] To understand the time required for file distribution and reduce the margin of distribution work time. A distribution control device includes a calculation processing device and a storage device, the storage device holds a communication quality prediction model and a data distribution plan, and the prediction model is a communication quality prediction model. The distribution plan includes a probability density distribution, the distribution plan includes identification information of the data distribution destination, and time information indicating the time when the data distribution will be performed, and the calculation processing device creates the data distribution plan based on the prediction model. The probability that data distribution will be completed from the start time to the end time specified by the time information included in is predicted, and the predicted probability that data distribution will be completed is output. [Selection diagram] Figure 9

Description

The present invention relates to a method of controlling distribution of data.

Edge devices such as factory equipment, railroad vehicles, and automobiles are becoming increasingly computerized, and software, information content, applications, functions, etc. are being distributed and deployed to edge devices via networks. In particular, if there is follow-up work such as confirmation and reflection after distribution, it is desirable that information distribution can be completed by the specified date and time, but for that reason, grasp the time required for distribution from the predicted communication quality and create an appropriate distribution plan There is a need. In addition, in order to proceed smoothly with subsequent work, it is necessary to grasp the progress of the distribution plan from the communication state during distribution.

Regarding creation of a distribution plan, Patent Literature 1 discloses a system for optimizing the transmission start time by calculating a plurality of transmission start time candidates and the end time when data is transmitted for each candidate. As for communication quality prediction, Patent Document 2 discloses throughput modeling means using a probability density function. As for grasping the communication state, Patent Document 3 discloses a system for calculating the degree of anomaly in consideration of the statistical reliability from the number of connection failures for the measurement result of the index of communication quality.

JP 2014-45267 A WO2013/008387 Japanese Patent Application Laid-Open No. 2006-340050

The technique of Patent Literature 1 is a point estimation of the required delivery time, and does not deal with the prediction results in a probabilistic manner. The technique of Patent Document 2 uses a probabilistic model to predict communication quality, but it is difficult to grasp the time required for distribution and assist in the creation of a distribution plan based on this technique alone. The technique of Patent Document 3 detects anomalies using statistical reliability from the number of connection failures, but cannot cope with cases where communication quality is degraded even though connection has not failed.

Accordingly, it is an object of the present invention to create a distribution plan in a form that can be statistically evaluated and manage the progress of the distribution plan by predicting the required distribution time and judging the communication state using probabilities.

In order to solve at least one of the above problems, the present invention provides a distribution control device having a calculation processing device and a storage device, wherein the storage device stores a communication quality prediction model and a data distribution plan. , wherein the prediction model includes the probability density distribution of the communication quality, and the distribution plan includes identification information of the distribution destination of the data and time information indicating the time at which the data is distributed. based on the prediction model, the calculation processing device predicts the probability that the data distribution will be completed from the start time to the end time specified by the time information included in the data distribution plan, and predicting It is characterized by outputting the probability of completion of distribution of the data.

According to one aspect of the present invention, it is possible to create a distribution plan and manage the progress of the distribution plan in a form that can be statistically evaluated. Specifically, selection of distribution timing, setting of a margin for distribution time, and setting of criteria for abnormality detection determination can be performed based on statistical criteria. Problems, configurations, and effects other than those described above will be clarified by the following description of the embodiments.

It is a block diagram showing an example of the configuration of the system of the embodiment of the present invention. It is a block diagram showing an example of the hardware constitutions of the system of the example of the present invention. It is a flowchart which shows an example of the whole delivery work in the Example of this invention. FIG. 4 is an explanatory diagram showing an example of a provisional distribution plan file created by the worker terminal according to the embodiment of this invention; It is explanatory drawing which shows an example of this delivery plan file produced by the terminal for workers of the Example of this invention. It is a block diagram which shows an example of the functional structure of the delivery assistance part of the Example of this invention. It is a flowchart which shows an example of operation|movement of the delivery completion probability prediction by the delivery control apparatus of the Example of this invention. It is explanatory drawing which shows an example of the prediction result output screen in the Example of this invention. FIG. 4 is an explanatory diagram showing an example of detailed operation of the terminal information database in the embodiment of the present invention; 9 is a flow chart showing an example of details of a distribution completion probability prediction operation in the distribution auxiliary unit of the embodiment of the present invention; 5 is a flow chart showing an example of operations of communication state determination, prediction model update, and abnormality notification in the delivery auxiliary unit of the embodiment of the present invention; FIG. 5 is an explanatory diagram showing an example of an anomaly notification display screen in the embodiment of the present invention; 4 is a flow chart showing an example of details of the operation of the communication state determination function in the embodiment of the present invention; It is a flowchart which shows an example of the detail of operation|movement of the prediction model update function in the Example of this invention.

FIG. 1A is a block diagram showing an example of system configuration according to an embodiment of the present invention.

The system of the first embodiment is composed of a worker terminal 101, a confirmer terminal 102, a distribution control device 103, and a plurality of distribution destination devices 104. FIG.

The distribution control device 103 is a computer equipped with a storage device, a computing device and a communication device. Note that the hardware configuration of a computer that implements the distribution control device 103 and the like will be described later with reference to FIG. 1B.

The distribution control device 103 has, as its functions, a distribution management section 201 and a distribution auxiliary section 202 corresponding to the present invention.

Each part of the distribution control device 103 is realized by executing a program in the storage device in the computation processing device.

The distribution control device 103 stores a distribution file 301 and a distribution command list 302 in a storage device.

The distribution management unit 201 executes the distribution plan to the distribution destination device 104 according to the main distribution plan file 304 created by the operator and the notification from the distribution auxiliary unit 202 . Details of the operation of the distribution management unit 201 will be described later with reference to FIG.

The distribution assistance unit 202 corresponding to the present invention assists the worker in creating a distribution plan, determines and notifies the distribution destination device 104 of the communication state, and updates the communication quality prediction model. A detailed configuration will be described later with reference to FIG.

The delivery destination device 104 is a computer equipped with a storage device, a computing device, and a communication device.

Each destination device 104 is identified by a unique destination ID 901 (see FIG. 3).

The delivery destination device 104 has a communication management unit 203 and a delivery acquisition unit 204 as its functions.

Each part of the delivery destination device 104 is implemented by executing a program in the storage device in the computing processing device.

The communication management unit 203 confirms the connection between the delivery destination device 104 and the delivery control device 103 (step 602) and measures the communication quality (step 604). A detailed operation of the communication management unit 203 will be described later with reference to FIG.

The delivery acquisition unit 204 processes the delivery file 301 from the delivery control device 103, executes the delivered command list 302, and creates work records.

The worker terminal 101 and the checker terminal 102 are computers having a storage device, a calculation processing device, a communication device, and an input/output interface.

The worker terminal 101 and the distribution control device 103 are connected by the communication network 105, the checker terminal 102 and the distribution control device 103 are connected by the communication network 106, and the distribution destination device 104 and the distribution control device 103 are connected by the communication network 107. and are in a state of being able to communicate with each other.

This embodiment is applicable to software distribution to devices. Examples include OS update, execution program distribution, setting file distribution, and display content distribution for equipment such as factory equipment, railroad vehicles, and automobiles, but the present embodiment is not limited to these. In this case, the distribution destination device 104 corresponds to the computer installed in each device, and the distribution control device 103 corresponds to the software management server.

It is assumed that the communication network 107 can be modeled by a probability density function whose communication quality (for example, total throughput) has time information 904 (see FIG. 3) or environmental information 907 (see FIG. 8) as variables. Examples of communication network 107 include, but are not limited to, a wireless network for periodically moving mobile devices, or a communication network in which network traffic volume varies depending on region and time of day. not something.

When this embodiment is applied to software distribution to mobile devices that periodically move, the time information 904 includes the time, day of the week, month and day of the distribution, and the environment information 907 includes the position and speed of each destination device. , altitude and communication device performance, etc., but are not limited thereto.

When this embodiment is applied to software distribution using a network with fluctuations in traffic volume, the time information 904 is the time, day of the week, month and day of the distribution, and the environment information 907 is the location and location of each distribution destination device. The amount of communication of surrounding devices and the like are conceivable, but are not limited to them.

FIG. 1B is a block diagram showing an example of the hardware configuration of the system according to the embodiment of the invention.

The distribution control device 103, the distribution destination device 104, the worker terminal 101, and the confirmer terminal 102 are each realized by a computer 110 as shown in FIG. 1B, for example. The computer 110 has a storage device 111 , a calculation processing device 112 , a communication device 113 and an input/output interface 114 .

The storage device 111 may include a so-called main storage device configured by, for example, a semiconductor memory, and a so-called auxiliary storage device configured by, for example, a large-capacity storage device such as a hard disk drive or solid state drive. The storage device 111 stores programs and data executed by the computation processing device 112 . For example, the distribution file 301 and distribution command list 302 are stored in the storage device 111 of the computer 110 corresponding to the distribution control device 103 . Information managed by a terminal information database 222 (see FIG. 5), which will be described later, is stored in the storage device 111 of the computer 110 corresponding to the distribution control device 103 .

The calculation processing device 112 executes various processes according to programs stored in the storage device 111 . Various functional units are implemented by the computer processing device 112 operating according to the program. For example, the functions of the distribution management unit 201 and the distribution auxiliary unit of the distribution control device 103 are realized by the calculation processing device 112 operating in accordance with the program stored in the storage device 111 in the computer 110 corresponding to the distribution control device 103. be. The functions of the communication management unit 203 and the delivery acquisition unit 204 of the delivery destination device 104 are realized by the calculation processing device 112 operating in accordance with the program stored in the storage device 111 in the computer 110 corresponding to the delivery destination device 104. be done.

A communication device 113 is an interface for connection with a communication network. For example, the communication device 113 of the computer 110 corresponding to the delivery control device 103 is connected to the communication networks 105 , 106 and 107 and communicates with the worker terminal 101 , the checker terminal 102 and the delivery destination device 104 .

The input/output interface 114 may include an input device such as a keyboard, mouse, and/or touch panel, and an output device such as an image display device.

FIG. 2 is a flow chart showing an example of overall distribution work in an embodiment of the present invention.

First, the worker operates the worker terminal 101 to input the created provisional distribution plan file 303 to the distribution auxiliary unit 202 (step 402). An example of the provisional distribution plan file 303 will be described later with reference to FIG. The distribution auxiliary unit 202 predicts the distribution completion probability in the range from the specified scheduled start date and time to the scheduled end date and time for the input provisional delivery plan file 303 (step 403), and returns the prediction result 305. (Step 404). Details of the distribution completion probability prediction operation will be described later with reference to FIG. Also, an example of a screen for outputting the prediction result 305 will be described later with reference to FIG.

For the prediction result 305, the worker operates the worker terminal 101 to modify the provisional distribution plan file 303 and create a final distribution plan file as necessary (step 405). After determining the final start date and time and the scheduled end date and time, the worker terminal 101 creates the final delivery plan file 304 by adding the corresponding prediction result 306 to the provisional delivery plan file. An example of this delivery plan file 304 is described later with reference to FIG.

After creating the distribution plan file 304, the worker terminal 101 transmits the distribution plan file 304 to the distribution control device 103 based on the operator's instruction, and registers it in the distribution management section 201 (step 406).

The distribution management unit 201 distributes the distribution file 301 and the distribution command list 302 to the designated distribution destination device 104 according to the registered distribution plan file 304 (step 407).

The distribution management unit 201 notifies the checker terminal 102 of the prediction result 306 of the distribution completion probability described in the main distribution plan file 304 at the start of the distribution work (step 408). This allows the confirmer to grasp the timing at which the subsequent work will occur.

The delivery acquisition unit 204 in the delivery destination device 104 executes the first-stage command 905 (see FIG. 4, for example, decompressing and arranging the delivery file 301, creating a work record, etc.) according to the received delivery command list 302 (step 409).

After completing the execution of the command in the first stage, the distribution acquisition unit 204 transmits a first stage completion notice to the distribution management unit 201 (step 410).

After receiving the first stage completion notices from all of the delivery destination devices 104, the delivery management unit 201 transmits a second stage approval request to the checker terminal 102 (step 411).

After receiving the approval response (step 412) from the checker terminal 102, the distribution management unit 201 sends the second stage command 906 (see FIG. 4, for example, execution/reflection of the distribution file 301, work record , etc.) is permitted (step 413).

The distribution acquisition unit 204 receives execution permission (step 413), executes the command 906 of the second stage (step 414), and then transmits the work record to the distribution management unit 201 (step 415).

The delivery management unit 201 confirms the consistency of the work records sent from each delivery destination device 104 (step 416).

After that, the delivery management unit 201 transmits an approval request for work completion to the confirmer terminal 102 (step 417). After receiving the approval response 418 from the confirmer terminal 102, the distribution management unit 201 transmits a distribution plan completion notice to the worker terminal 101 (step 419). This completes the preparation of the distribution plan and the distribution work based thereon.

During the execution of the distribution plan, the distribution management unit 201 determines the timing of starting distribution of the distribution file 301 and the distribution command list 302 to the distribution acquisition unit 204 (step 408), and the distribution acquisition of all distribution destination devices 104. At the timing of receiving the first-stage command completion notification from the unit 204 (step 410), a distribution start notification 420 and a distribution completion notification 421 are transmitted to the distribution auxiliary unit 202, respectively. The distribution start notification 420 and distribution completion notification 421 include the distribution destination ID 901, distribution file list 902, distribution file size 903 and time information 904 (see FIG. 4) in the distribution plan file 304. FIG.

The distribution auxiliary unit 202 determines the communication state (step 609) from the time when the distribution start notification is received until the time when the distribution completion notification is received. Notification (step 611) is performed.

The distribution management unit 201 that has received the notification of the abnormal state (step 611) notifies the operator terminal 101 and the checker terminal 102 of the abnormality (step 612). As a result, the worker and the confirmer can grasp the progress delay in the work plan and take measures against the abnormality (step 613). Details of the communication state determination and the abnormality notification operation will be described later with reference to FIG.

FIG. 3 is an explanatory diagram showing an example of the provisional distribution plan file 303 created by the worker terminal 101 according to the embodiment of this invention.

The provisional distribution plan 303 includes a distribution destination ID 901 , a distribution file list 902 , a distribution file size 903 , time information 904 regarding the distribution schedule, and a distribution command list 302 in the distribution destination device 104 .

The distribution destination ID 901 is identification information of the distribution destination device 104 to which the file is to be distributed according to the provisional distribution plan 303 . A delivery file 902 and a delivery file size 903 are the identification information and the size of the file to be delivered (that is, the delivery file 301), respectively. A plan for distributing the same data to a plurality of destination apparatuses 104 may be created as one provisional distribution plan 303 . In that case, the distribution destination ID 901 includes identification information of a plurality of distribution destination devices 104 .

The delivery schedule time information 904 is information indicating the scheduled time for file delivery, and includes, for example, the scheduled start date and time and the scheduled end date and time. The scheduled start date and time and the scheduled end date and time may each be information with a range. In the example of FIG. 3, distribution may start between 10:00 and 11:00 on December 1, 2021, and end between 12:30 and 13:30 on the same day. Planned.

Delivery command list 302 includes one or more commands to be executed by delivery destination device 104 when the file is delivered. In the example of FIG. 3, the distribution command list 302 includes first-stage commands 905 (for example, decompression/placement of the distribution file 301, creation of work records, etc.) and second-stage commands 906 (for example, execution/reflection of the distribution file 301). , creation of work records, etc.).

FIG. 4 is an explanatory diagram showing an example of the final distribution plan file 304 created by the worker terminal 101 according to the embodiment of this invention.

This distribution plan 304 includes a distribution destination ID 901, a distribution file list 902, a distribution file size 903, time information 904 regarding the distribution schedule, an execution command list 302 in the distribution destination device 104, and a distribution completion probability prediction result 306 in this plan. include.

A distribution destination ID 901 , a distribution file list 902 , a distribution file size 903 and a distribution command list 302 are the same as the provisional distribution plan 303 on which the actual distribution plan 304 is based. The time information 904 regarding the distribution schedule of the main distribution plan 304 includes the scheduled start date and time selected from the scheduled start date and time included in the time information 904 regarding the distribution schedule of the provisional distribution plan 303, and the prediction results of the scheduled start date and time and distribution completion probability. 306, and a scheduled end date and time. The prediction result 306 of the delivery completion probability in this plan includes the prediction result of the delivery completion probability at each time when the delivery is started at the selected scheduled start date and time. This will be explained in detail later.

FIG. 5 is a block diagram showing an example of the functional configuration of the delivery assistance unit 202 according to the embodiment of the present invention.

The distribution auxiliary unit 202 has a distribution plan information extraction function 221, a terminal information database 222, a prediction model update function 226, and a distribution completion probability prediction function 223 as functions for realizing distribution completion probability prediction (step 403). An operational flow diagram of these functions will be described with reference to FIG.

Further, the distribution auxiliary unit 202 has a communication quality acquisition function 224, a communication state It has a determination function 225 and a prediction model update function 226 . The operational flow of these functions will be described later with reference to FIG.

FIG. 6 is a flow chart showing an example of operation of distribution completion probability prediction (step 403) by the distribution control device 103 of the embodiment of the present invention.

First, the distribution auxiliary unit 202 of the distribution control device 103 extracts information from the provisional distribution plan file 303 input in step 402 by the distribution plan information extraction function 221 (step 501). The extracted distribution destination ID 901 and distribution time information 904 are transferred to the terminal information database 222 (step 502). The extracted distribution file list 902 and distribution file size 903 are passed to the distribution completion probability prediction function 223 (step 503).

The terminal information database 222 inquires the database based on the extracted distribution destination ID 901 and distribution time information 904, and creates the feature quantity table 307 (step 504). The terminal information database 222 and the feature amount table 307 will be described later with reference to FIG.

The created feature amount table 307 is passed to the distribution completion probability prediction function 223 (step 505). The delivery completion probability prediction function 223 determines that there is a request for prediction when the feature amount table 307 is received, and requests the latest prediction model (throughput probability density function) 308 at that time from the prediction model update function 226 ( Step 506) and obtain (Step 507). Details of the prediction model 308 will be described later with reference to FIG.

The distribution completion probability prediction function 223 predicts the distribution completion probability corresponding to the distribution plan using the distribution file size 903, the feature amount table 307, and the prediction model 308 obtained by the above processing (step 508). A prediction result 305 is output to the terminal 101 (step 404). Details of the operation of delivery completion probability prediction (step 508) will be described later with reference to FIG.

FIG. 7 is an explanatory diagram showing an example of a prediction result output screen in the embodiment of the present invention.

A prediction result output screen 701 is a screen for outputting a prediction result 305 of the distribution completion probability by the distribution control device 103 . For example, the distribution control device 103 transmits the prediction result 305 of the distribution completion probability to the worker terminal 101, and the output device of the worker terminal 101 (for example, the input/output interface 114 of the computer 110 corresponding to the worker terminal 101). The image display device included) may display the prediction result output screen 701 .

The prediction result output screen 701 includes, for example, a distribution destination ID 901 similar to the provisional distribution plan 303, a distribution file list 902, a distribution file size 903, and time information 904, as well as a prediction result 305. FIG. The prediction result 305 shows that for each combination of some start dates and times within the range of the scheduled start dates and times of the time information 904 and some end dates and times within the range of the scheduled end dates and times, the delivery that started at the start dates and times ends. Shows the result of predicting the probability of ending on the date and time.

By predicting the delivery completion probability for each set of start date and time and end date and time in a format like the prediction result 305 in FIG. It is possible to obtain the effect of helping margin setting and subsequent work planning.

FIG. 8 is an explanatory diagram showing an example of detailed operation of the terminal information database 222 in the embodiment of the present invention.

In the terminal information database 222, the environment information 907 linked with the delivery destination ID 901 and the time information 904 is input in advance and stored. Examples of the time information 904 include the time, day of the week, and day of month for distribution, and examples of the environment information 907 include the position, speed, altitude, performance of the communication device, and traffic volume of surrounding devices of each distribution destination device. , but not limited to them.

When inquiry information (extracted delivery destination ID 901 and time information 904 ) is input from the provisional delivery plan file 303 , the terminal information database 222 inquires the corresponding environment information 907 and creates the feature amount table 307 . A feature quantity table 307 is a correspondence table of distribution destination IDs 901, time information 904, and environment information 907, and corresponds to multidimensional variables of a prediction model 308, which will be described later.

FIG. 9 is a flow chart showing an example of details of the distribution completion probability prediction (step 508) in the distribution auxiliary unit 202 of the embodiment of the present invention.

First, the distribution completion probability prediction function 223 of the distribution auxiliary unit 202 sets a table of prediction time intervals 314 (start date and time and end date and time) from the time information 904 (step 581). For example, when the time information 904 includes information specifying a range of start dates and times and information specifying a range of end dates and times, one or more start dates and times within the specified range and A combination of one or more end dates and times is set. In this embodiment, nine combinations of three start dates and three end dates are set as the predicted time interval 314 .

Thereafter, the distribution completion probability prediction function 223 performs variable conversion on the set prediction time interval 314 using the feature amount table 307, and associates the prediction time interval 314 with the variables of the prediction model (throughput probability density) 308. A predictor variable interval 315 is created (step 582).

The delivery completion probability prediction function 223 then marginalizes the throughput probability density function according to the predictor variable interval 315 to obtain the probability density function 316 of total throughput in the prediction time interval (step 583).

The distribution completion probability prediction function 223 calculates the distribution completion probability in the prediction time interval by integrating the probability that the distribution file 301 size or larger is obtained with respect to the probability density function 316 of the total throughput in the prediction time interval ( step 584). That is, based on the probability density function 316 of the total throughput, the probability that the size of the data that can be delivered in the predicted time interval from the start date and time to the end date and time is equal to or larger than the size of the distribution live 301 is the distribution completion probability in the predicted time interval. Calculated.

The prediction result 305 shown in FIG. 7 is obtained by performing the above calculation for each set of start date and time and end date and time set in the prediction time interval table created from the time information 904 . When a plurality of distribution destination IDs 901 are specified, after calculating the distribution completion probability for each distribution destination ID 901, the worst value is used as the prediction result. Finally, the distribution completion probability prediction function 223 outputs the prediction result (step 404).

FIG. 10 is a flow chart showing an example of operations of communication state determination, prediction model update, and abnormality notification in the distribution auxiliary unit 202 according to the embodiment of the present invention.

The distribution auxiliary unit 202 in the distribution control device 103 periodically confirms connection with the communication management unit 203 in each distribution destination device 104 by means of the communication quality acquisition function 224 (steps 601 and 602). When the connection is established, the communication quality acquisition function 224 transmits the throughput measurement data 309 to the communication management section 203 (step 603).

Upon receiving the throughput measurement data 309, the communication manager 203 returns the measurement information 310 (step 604). In addition to measurement throughput 908, measurement information 310 includes measurement distribution destination ID 909, measurement time information 910, and measurement environment information 911 such as position, speed, and altitude at the time of measurement.

The acquired measurement information 310 is input to the communication state determination function 225 (step 605).

When communication quality acquisition function 224 fails to establish connection with communication management unit 203, communication quality acquisition function 224 creates measurement information 310 indicating a disconnection state, and inputs it to communication state determination function 225 (step 606).

The communication state determination function 225 receives the distribution start notification 420 and the distribution completion notification 421 from the distribution management unit 201 in parallel with receiving the measurement information 310 from the communication quality acquisition function 224 .

The communication state determination function 225 that has received the measurement information requests the prediction model 308 from the prediction model update function 226 (step 607).

The prediction model update function 226 that has received the request returns the latest prediction model at that time (step 608).

The communication state determination function 225 uses the input measurement information 310 to determine the communication state (step 609).

According to the communication status determination result 311 and the current delivery status 312, the communication status determination function 225 transmits the update data 313 to the prediction model update function 226 (step 610) and notifies the delivery management unit 201 of the abnormality. Notification (step 611) is performed. Details of the operation of the communication state determination function 225 will be described later with reference to FIG.

When receiving the abnormality notification in step 611, the distribution management unit 201 notifies the worker terminal 101 and the checker terminal 102 that the progress of the current distribution plan is in an abnormal state (step 612).

When the operator terminal 101 and the checker terminal 102 receive the abnormality notification in step 612, the operator and the checker can take action against the abnormality (step 613).

FIG. 11 is an explanatory diagram showing an example of an anomaly notification display screen according to the embodiment of the present invention.

The abnormality notification display screen 1101 is a screen for outputting information indicating that the communication state is abnormal as a result of the communication state determination (step 609) by the distribution control device 103. FIG. For example, the distribution control device 103 transmits a notification indicating an abnormality in the communication state to the worker terminal 101 and the checker terminal 102, and the output devices of the worker terminal 101 and the checker terminal 102 (for example, An image display device included in the input/output interface 114 of the computer 110 corresponding to the terminal 101 and the confirmation person terminal 102) may display the abnormality notification display screen 1101. FIG.

On the abnormality notification display screen 1101, in addition to a distribution file list 902, a distribution file size 903, and time information 904 regarding the distribution plan currently being executed, a distribution destination ID 909 where an abnormality was detected, and measurement time information indicating the time when the abnormality was detected. 910, measurement environment information 911 indicating the environment in which the abnormality was detected (for example, position, speed, altitude, etc. at the time of measurement), and determination result 311 of the communication state indicating that the abnormality was detected are displayed.

FIG. 12 is a flow chart showing an example of details of the operation of the communication state determination function 225 in the embodiment of the present invention.

When the measurement information 310 is input, the communication state determination function 225 first requests the current latest prediction model (throughput probability density function) 308 from the prediction model update function 226 (step 607) and acquires it (step 608). After that, the communication state determination function 225 substitutes the value of the measurement environment information 907 for the obtained prediction model 308, and obtains the likelihood of the measured throughput 908 in the measurement environment (step 691). The communication state determination function 225 classifies the likelihood and the measurement information 310 into classes (step 692), thereby determining the communication state (step 609).

The communication state includes a normal state and an abnormal state. For example, if the likelihood is higher than a predetermined standard, the normal state may be determined, and if the likelihood is lower than the predetermined standard, the abnormal state may be determined. Each state may be further subdivided into multiple classes, such as normal (high throughput), normal (medium throughput), normal (low throughput), abnormal (low throughput), and abnormal (disconnected).

Update data 313 including the communication state determination result 311 and the measurement information 310 is input to the prediction model update function 226 . This continuously updates the prediction model 308 to improve and maintain prediction accuracy.

In addition to these processes, the communication state determination function 225 receives a distribution start notification 420 and a distribution completion notification 421 from the distribution management unit 201 in the distribution control device 103 . Upon receiving each notification, the communication status determination function 225 updates the current delivery status 312 (during delivery or non-delivery).

If the communication status determination result 311 is abnormal and the current distribution status 312 is in the process of distribution, the communication status determination function 225 transmits an abnormality notification to the distribution management section 201 (step 611).

FIG. 13 is a flow chart showing an example of details of the operation of the predictive model update function 226 in an embodiment of the invention.

Forecast model update function 226 includes forecast model 308 and forecast model update algorithm 227 therein.

The prediction model 308 is a probability density function having throughput and multi-dimensional feature quantities (for example, feature quantities including distribution destination ID 901, time information 904, and environment information 907) as its variables.

The prediction model update algorithm 227 updates the prediction model 308 based on the update data 313 . Any algorithm such as machine learning may be used as the update algorithm 227 .

Note that the update data 313 includes the measurement information 310 and the communication state determination result 311 as described above. The predictive model update function 226 may use the communication state determination result 311 to update the predictive model 308 . For example, the predictive model update function 226 does not need to use the measurement information 310 whose communication state determination result 311 is abnormal to update the predictive model 308 .

When a prediction model request (step 506 or 607) is made to the prediction model update function 226, the prediction model update function 226 returns the latest prediction model at the time of the request as an output (step 507 or 608).

Moreover, the system of the embodiment of the present invention may be configured as follows.

(1) A distribution control device (for example, distribution control device 103) having a calculation processing device (for example, calculation processing device 112) and a storage device (for example, storage device 111), wherein the storage device is a communication quality prediction model (For example, a prediction model 308) and a data distribution plan (for example, a provisional distribution plan 303) are held, the prediction model includes the probability density distribution of communication quality, and the distribution plan contains identification information of the data distribution destination ( (e.g. delivery destination ID 901) and time information (e.g. time information 904) indicating the time at which data is to be delivered. The probability of completion of data distribution from the specified start time to the end time is predicted (eg, step 403), and the predicted probability of completion of data distribution is output (eg, step 404).

This makes it possible to create a distribution plan and manage the progress of the distribution plan in a form that can be statistically evaluated. Specifically, selection of distribution timing, setting of a margin for distribution time, and setting of criteria for abnormality detection determination can be performed based on statistical criteria. That is, it is possible to grasp the time required for file distribution and reduce the margin of the distribution work time.

(2) In (1) above, the probability density distribution is a multidimensional probability that includes time information and at least one of the location, movement speed, and altitude of the data distribution destination (for example, environment information 907) as feature quantities. density distribution.

This makes it possible to calculate the probability of communication quality according to the time and the environment of the distribution destination corresponding to the time.

(3) In (2) above, the storage device includes delivery destination information (for example, terminal information database 222), and based on the distribution destination information, the calculation processing device calculates the position, moving speed, and altitude of the data distribution destination corresponding to the time information and the identification information of the data distribution destination included in the distribution plan. A feature amount including at least one of them is generated (for example, the process of FIG. 8), and the generated feature amount is applied to a prediction model to predict the probability that data delivery will be completed between the start time and the end time.

This makes it possible to calculate the probability of communication quality according to the time and the environment of the distribution destination corresponding to the time.

(4) In (1) above, the computing device measures the communication quality with the data delivery destination at a predetermined timing (for example, steps 603 and 604), and predicts the communication quality based on the measured communication quality and the prediction model. The communication state at the time of measurement is determined by comparing with the quality (for example, step 609), and the determined communication state is output (for example, steps 611 and 612).

Thereby, abnormality detection determination based on statistical criteria can be performed.

(5) In (4) above, the computing device compares the measured communication quality with the communication quality predicted based on the prediction model to calculate the likelihood of the measured communication quality (eg step 691). If the measured likelihood of communication quality is lower than a predetermined standard and data distribution to the data distribution destination is in progress, information indicating that the communication state is abnormal is output (for example, step 611 , 612).

Thereby, abnormality detection determination based on statistical criteria can be performed.

(6) In (1) above, the computing device measures the communication quality with the data delivery destination at a predetermined timing (for example, steps 603 and 604), and calculates the communication quality probability based on the measured communication quality. Update the prediction model by calculating the density distribution (eg, step 614).

This allows the accuracy of the prediction model to be maintained.

(7) In (6) above, the computing device determines the communication state at the time of measurement by comparing the measured communication quality with the communication quality predicted based on the prediction model (for example, step 691), and performs the measurement. Based on the communication quality and the determined communication state, the prediction model is updated by calculating the probability density distribution of the communication quality (eg, step 614).

This allows the accuracy of the prediction model to be maintained.

(8) In the above (7), the probability density distribution is a multi-dimensional probability density distribution including time information and at least one of the position, movement speed and altitude of the data delivery destination as feature quantities, and calculation processing The device compares the measured communication quality with the communication quality predicted based on the prediction model to calculate the likelihood of the measured communication quality (for example, step 691), and the likelihood of the measured communication quality is a predetermined If it is lower than the standard, the communication state is determined to be abnormal (for example, step 692). , the identification information of the data delivery destination and at least one of the data delivery destination's position, moving speed and altitude at the time of measurement, and updating the prediction model by calculating the probability density distribution of the communication quality. (eg step 614).

This allows the accuracy of the prediction model to be maintained.

(9) In (1) above, the time information included in the data distribution plan includes information designating a range of start times and a range of end times, Delivery of data from the start time to the end time based on the predictive model for each of the one or more start and end time combinations based on the start time, one or more end times within the range of end times Predict the probability that .

This assists in creating a delivery plan in a form that can be statistically evaluated.

In addition, the present invention is not limited to the above-described embodiments, and includes various modifications. For example, the above embodiments have been described in detail for better understanding of the present invention, and are not necessarily limited to those having all the configurations described. Also, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Moreover, it is possible to add, delete, or replace a part of the configuration of each embodiment with another configuration.

Further, each of the above configurations, functions, processing units, processing means, and the like may be realized by hardware, for example, by designing a part or all of them using an integrated circuit. Moreover, each of the above configurations, functions, etc. may be realized by software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, files, etc. that realize each function is stored in storage devices such as non-volatile semiconductor memories, hard disk drives, SSDs (Solid State Drives), or computer-readable non-volatile memory such as IC cards, SD cards, DVDs, etc. It can be stored on a temporary data storage medium.

Further, the control lines and information lines indicate those considered necessary for explanation, and not necessarily all the control lines and information lines are indicated on the product. In fact, it may be considered that almost all configurations are interconnected.

101 Terminal for operator 102 Terminal for confirmer 103 Distribution control device 104 Distribution destination device 105 Communication network (terminal for worker - distribution control device)
106 Communication network (confirmer terminal-distribution control device)
107 communication network (delivery control device - delivery destination device)
201 distribution management unit 202 distribution auxiliary unit 203 communication management unit 204 distribution acquisition unit 221 distribution plan information extraction function 222 terminal information database 223 distribution completion probability prediction function 224 communication quality acquisition function 225 communication state determination function 226 prediction model update function 227 probability density Function update algorithm 301 Distribution file 302 Distribution command list 303 Temporary distribution plan file 304 Final distribution plan file 305 Distribution completion probability prediction result for temporary distribution plan 306 Distribution completion probability prediction result for main distribution plan 307 Feature amount table 308 Prediction model 309 Throughput measurement Data for use 310 Measurement information 311 Communication state determination result 312 Delivery status 313 Terminal information database update data 314 Predicted time interval 315 Predictor variable interval 316 Probability density function 901 Delivery destination ID
902 Distribution file list 903 Distribution file size 904 Time information 905 Command list (first stage)
906 Command list (second stage)
907 Environment information 908 Measurement throughput 909 Measurement distribution destination ID
910 Measurement time information 911 Measurement environment information

Claims (10)

A distribution control device having a calculation processing device and a storage device,
The storage device holds a prediction model of communication quality and a data distribution plan,
The prediction model includes a probability density distribution of the communication quality,
the distribution plan includes identification information of a distribution destination of the data and time information indicating a time when the data is distributed;
The computing device is
Based on the prediction model, predicting the probability that the data delivery will be completed from the start time to the end time specified by the time information included in the data delivery plan;
A distribution control device that outputs a predicted probability of completion of distribution of the data. The distribution control device according to claim 1,
The distribution control device, wherein the probability density distribution is a multi-dimensional probability density distribution including the time information and at least one of the position, moving speed and altitude of the distribution destination of the data as feature quantities. The distribution control device according to claim 2,
the storage device holds delivery destination information that associates the time information, identification information of the delivery destination of the data, and at least one of the location, moving speed, and altitude of the delivery destination of the data;
Based on the distribution destination information, the calculation processing device determines at least one of the position, moving speed, and altitude of the distribution destination of the data corresponding to the time information and the identification information of the distribution destination of the data included in the distribution plan. generating the feature quantity including
A distribution control device that predicts a probability of completion of distribution of the data from the start time to the end time by applying the generated feature amount to the prediction model. The distribution control device according to claim 1,
The computing device is
measuring the communication quality with the delivery destination of the data at a predetermined timing;
determining the communication state at the time of measurement by comparing the measured communication quality with the communication quality predicted based on the prediction model;
A delivery control device, characterized by outputting the determined communication state. The distribution control device according to claim 4,
The computing device is
calculating the likelihood of the measured communication quality by comparing the measured communication quality with the communication quality predicted based on the prediction model;
outputting information indicating that the communication state is abnormal when the measured likelihood of the communication quality is lower than a predetermined standard and the data is being distributed to the data distribution destination; A distribution control device characterized by: The distribution control device according to claim 1,
The computing device is
measuring the communication quality with the delivery destination of the data at a predetermined timing;
A distribution control device, wherein the prediction model is updated by calculating a probability density distribution of the communication quality based on the measured communication quality. The distribution control device according to claim 6,
The computing device is
determining the communication state at the time of measurement by comparing the measured communication quality with the communication quality predicted based on the prediction model;
A distribution control device, wherein the prediction model is updated by calculating a probability density distribution of the communication quality based on the measured communication quality and the determined communication state. The distribution control device according to claim 7,
The probability density distribution is a multi-dimensional probability density distribution that includes the time information and at least one of the position, movement speed, and altitude of the delivery destination of the data as feature quantities,
The computing device is
calculating the likelihood of the measured communication quality by comparing the measured communication quality with the communication quality predicted based on the prediction model;
determining that the communication state is abnormal when the measured likelihood of communication quality is lower than a predetermined standard;
Among the measured communication qualities, the communication quality for which the communication state was not determined to be abnormal, the time when the communication quality was measured, the identification information of the delivery destination of the data, and the delivery of the data at the time of measurement A distribution control device, wherein the prediction model is updated by calculating the probability density distribution of the communication quality based on at least one of a previous position, moving speed and altitude. The distribution control device according to claim 1,
The time information included in the data distribution plan includes information specifying the start time range and the end time range,
The calculation processing device determines one or more combinations of the start time and the end time based on one or more start times within the range of the start times and one or more end times within the range of the end times. A distribution control device for predicting a probability that distribution of the data will be completed from the start time to the end time based on the prediction model. A distribution control method executed by a computer system having a computation processing device and a storage device,
The storage device holds a prediction model of communication quality and a data distribution plan,
The prediction model includes a probability density distribution of the communication quality,
the distribution plan includes identification information of a distribution destination of the data and time information indicating a time when the data is distributed;
The delivery control method includes:
a step of predicting, based on the prediction model, the probability that the distribution of the data will be completed between the start time and the end time specified by the time information included in the data distribution plan;
A distribution control method, comprising: a step of outputting a predicted probability of completion of distribution of the data by the calculation processing device.

Images