JP6748374B2 - Sensor management system - Google Patents

Description

The present invention relates to a sensor management system, a sensor management method, and a program.

Due to the wide bandwidth of the Internet, the cost reduction, and the miniaturization of wireless communication devices, a sensor management system has been proposed or put into practical use, which obtains sensing data from a large number of geographically dispersed sensors in real time and provides various services. ing.

For example, in Patent Document 1, a server, a gateway connected to the server via a network such as the Internet, various types of sensors connected to the gateway via a network such as ZigBee, and a console connected to the server. A sensor management system having and is described.

JP, 2014-68285, A

By the way, in the sensor management system, there are many cases where the sensor is replaced. For example, the case where the latest sensor is used instead of the sensor that has failed or deteriorated. In such a case, the user will have to replace the sensor, but if the user decides whether or not the sensor after replacement can be used as a substitute for the sensor before replacement, it may cause an unexpected situation. There is. The reason is that if a sensor that cannot be used as a substitute, such as an incompatible sensor, is replaced, the data acquired from the sensor will be changed to a completely different data, and the application processing in the subsequent stage will malfunction. Is.

An object of the present invention is to solve the above-mentioned problem, that is, a sensor that solves the problem that entrusting a user to determine whether a sensor after replacement can be used as a substitute for a sensor before replacement causes an unexpected situation. To provide a management system.

A sensor management system according to an aspect of the present invention is
A sensor management system for managing a sensor,
Detection means for detecting replacement of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, based on the result of the determination, the sensor after replacement is replaced by the sensor before replacement. A replacement coping means for deciding whether or not to use as an alternative to
Have.

A sensor management method according to another aspect of the present invention is
A method for managing a sensor in a sensor management system, comprising:
Detects the replacement of the sensor connected to the sensor management system,
Obtain the sensing data from the replaced sensor,
Determine whether the acquired sensing data of the sensor after replacement is similar to the past sensing data of the sensor before replacement,
Based on the result of the determination, it is determined whether the sensor after replacement is used as a substitute for the sensor before replacement.

A program according to another aspect of the present invention is
Computer,
A detection means for detecting the exchange of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, based on the result of the determination, the sensor after replacement is replaced by the sensor before replacement. A replacement coping means for deciding whether or not to use as an alternative to
And make it work.

Since the present invention has the above-mentioned configuration, it is possible to avoid an unexpected situation due to replacement with an incorrect sensor such as an incompatible sensor.

It is a figure which shows the structure of the sensor management system relevant to this invention. It is a figure which shows the data flow definition used by the sensor management system relevant to this invention. It is a figure which shows the structural example of the sensor management system which concerns on the 1st Embodiment of this invention. It is a figure which shows the example of replacement of the device in the sensor management system which concerns on the 1st Embodiment of this invention. It is a figure which shows the example of the edge GW profile created with the sensor management system which concerns on the 1st Embodiment of this invention. It is a flow chart which shows the flow of influence judgment and corrective action in the sensor management system concerning a 1st embodiment of the present invention. It is a figure which shows the example of a principal part structure of the sensor management system which concerns on the 2nd Embodiment of this invention. It is a figure showing an example of important section composition of a sensor management system concerning a 3rd embodiment of the present invention. It is a figure which shows the structural example of the sensor management system which concerns on the 4th Embodiment of this invention. It is a flowchart which shows an example of operation|movement of the sensor management system which concerns on the 4th Embodiment of this invention. It is an operation|movement flowchart which shows an example of the detection part in the sensor management system which concerns on the 4th Embodiment of this invention. It is an operation|movement flowchart which shows the other example of the detection part in the sensor management system which concerns on the 4th Embodiment of this invention. It is a figure which shows the structural example of the replacement coping part in the sensor management system which concerns on the 4th Embodiment of this invention. It is a figure which shows the other structural example of the exchange coping part in the sensor management system which concerns on the 4th Embodiment of this invention. It is a flowchart which shows an example of operation|movement of the sensor management system which concerns on the 5th Embodiment of this invention. It is a principal part flowchart which shows an example of operation|movement of the sensor management system which concerns on the 6th Embodiment of this invention.

Next, embodiments of the present invention will be described in detail with reference to the drawings.
[First Embodiment]
Hereinafter, the first embodiment of the present invention will be described in detail in the order of problems to be solved, means for solving the problems, configurations, operations, and effects.

<Problems to be solved by this embodiment>
In recent years, expectations for IoT (Internet of Things) technology have increased. When building an IoT system, processing that requires a lot of machine resources is executed by a back-end server such as a cloud environment, and interaction with an actual sensor device (hereinafter simply referred to as device) and immediacy are required. In order to perform processing as close to the device as possible, an edge gateway device (hereinafter referred to as an edge GW device) is often arranged.

The processing that requires immediacy in the edge GW device includes processing that physically affects a person in the real world and is related to the life of the person. Under such circumstances, there is an increasing demand for updating devices and device drivers safely and promptly due to device failures and security problems.

However, up to now, in order to deal with such a problem, the interface specification between the device and the application on the edge GW device is strictly defined, and the device and the device driver can be updated at any time as long as the interface specification is complied with. However, there is a long way to go to standardize all of the technologies around IoT, with many standardization groups scattered around. In addition, due to such circumstances, there are cases in which standards are first set in a closed form in a specific industry to address actual problems, and in such cases, reusability, interoperability, and convenience of the device itself are used. Will be damaged. Therefore, there are the following problems.

On the edge GW device, the device driver reads the data generated from the device, and returns the data (hereinafter referred to as device data) converted into a predetermined format (for example, JSON (Javascript Object Notation) format), and further Combines device data generated from multiple devices, formats the data into normalized data so that the data can be easily processed, executes the application logic using the normalized data, and backs the resulting data. It is assumed that the device is sent to the server and the device is operated by the condition judgment. In such a situation, if you replace the device with another device due to a device failure or if the device driver itself is upgraded, updating the applied device driver will change the returned device data schema. However, the normalization process in the latter stage, application logic, and device operation may malfunction. Hereinafter, this point will be described with reference to the drawings.

FIG. 1 shows the configuration of a sensor management system related to the present invention. In FIG. 1, the edge GW apparatus has a plurality of device drivers to which devices are connected, a normalization processing unit, a user logic processing unit, and a data processing unit including a data flow control unit. The result processed by the data processing unit is transmitted to the backend server by the data transmission unit. The back-end server receives the data by the data receiving unit, processes the data by the data processing unit, and then stores the data in the data storage unit. Further, the edge GW device has a device detection/driver update unit, and when a device change is detected, the applicable device driver is downloaded from the device driver repository of the backend server and applied.

Here, assume a data flow definition flow1 as shown in FIG. With reference to FIG. 2, this data flow definition flow1 has, for example, a temperature sensor s1 and a humidity sensor s2, the device driver for each device returns the data as shown in the drawing, and the normalization definition rule1 causes the data to flow as shown in the drawing. Normalization data stats is generated, a discomfort index is calculated by the user logic L1, and the data flow is transmitted to the back end. At this time, it is assumed that a person in charge at the site where the edge GW device is present has replaced the temperature sensor s1 with another sensor s1' due to, for example, a failure of the temperature sensor s1. At this time, if the sensor s1' is a sensor having completely the same specifications as the original sensor s1 or a compatible sensor, there is almost no influence on the subsequent normalization processing and the user logic. However, if the sensor s1′ that is incompatible with the temperature sensor s1 is replaced by a mistake made by the person in charge, the device data from the sensor s1′ will be different from the device data from the original sensor s1. , The normalization process and user logic in the latter stage will not operate normally.

Further, even in the case of the sensor s1' which can be used as a substitute for the temperature sensor s1, if the specification is different from the original sensor s1, it is necessary to perform the normalization process in the subsequent stage and the correction of the user logic. If such a normalization process in the latter stage and a modification of the user logic are carried out manually, much labor and labor is required. In particular, since the person in charge of the site where the edge GW device is present has no IT knowledge, it is necessary to request the IT person who manages the entire system to use a new device used as a substitute for the original device. The IT staff who received this request manually modified the data processing unit (normalization definition and user logic) so that it could handle the new device, and tested that there was no problem using the new device. You must authorize the use above. Therefore, in the real world, even if a device that is convenient and usable is on the edge GW apparatus side and is in a usable state, it is not possible to promptly use the unapproved device, so that the use of the device is permitted. Takes too long.

On the other hand, there are cases where it is possible to make controlled updates by allowing only devices that comply with a certain standard specification to be connected.However, the standard specifications are scattered and all devices are standardized. If a device is tied to a specific standard specification in a situation where it cannot be expected to comply with the specification, the devices that can be used are limited, which also results in a loss of convenience.

<Outline of means for solving the problem>
In the present embodiment, the edge GW apparatus extracts the characteristic points of the data actually generated from the device by the edge GW apparatus in the application configured by the data flow, the template type normalization definition, and the user logic. It is stored as profile data on the back-end server side via the network. Next, at the timing of replacing the device, with the data processing stopped, characteristic points for judging the tendency of the data related to the replaced device are collected from the actual site, and the characteristic points are collected based on the profile data. Perform compatibility judgment processing according to the extraction method. Then, by this compatibility determination process, the similarity between the data acquired from the device after replacement and the past data of the device before replacement is determined, and whether the device after replacement can be used as a substitute for the device before replacement or not To decide. Then, when it is determined that the replaced device can be used as a replacement of the replaced device, the normalization definition is automatically set to operate the system by using the replaced device as a replacement of the replaced device. After that, fix the problem and confirm that there is no problem in the test, and then restart the data processing. As a result, the impact determination and the corrective action that prevent the adverse effect of the device update on the real world are realized.

As described above, according to the present embodiment, the profile information on the edge GW apparatus that is actually operating is collected, the profile information is held on the back-end server side, and the data is updated at the timing of device replacement. With the processing stopped, characteristic points for determining the tendency of the data related to the device after replacement are collected from the actual site, and based on the above profile data, the data acquired from the device after replacement and the data before replacement The similarity between the device and the past data of the device is determined, and it is determined whether the device after replacement can be used as a substitute for the device before replacement.

In addition, when changing the normalization processing and application logic due to device replacement, the presence or absence of the influence and the range of influence are judged, and those that can be automatically corrected are corrected, and those that cannot be automatically corrected are The edge GW device side may be promptly and safely updated by notifying the user and prompting the user to deal with the influence.

Further, even for an edge GW apparatus in which a large number of different devices are connected, the stored profile information is classified and grouped, and the influence range is specified while extracting the difference between the groups, thereby effectively affecting the influence. After the range is specified, it may be configured so that the user can easily understand the range of influence.

<Configuration of this embodiment>
The configuration of this embodiment will be described. Referring to FIG. 3, the sensor management system 100 according to the present embodiment includes a back-end server 130 such as a cloud, an edge GW device 140 connected to the back-end server 130 via a WAN (Wide Area Network), and its The device 111 includes a device 111 connected to the edge GW apparatus 140, an impact determination and corrective action unit 101 that extends across the back-end server 130 and the edge GW apparatus 140.

The back-end server 130 includes a data receiving unit 121, a data storage unit 122, a data processing unit 123, a normalization definition/data flow definition/user logic repository 125, a device driver repository 124, and an edge GW data processing test unit. 126 and the backend side influence determination and corrective action unit 101.

The edge GW device 140 includes a device driver 112 corresponding to each connected device, an edge GW data processing unit 117, a data transmission unit 116, a device detection/driver update unit 113, a device data collection unit 114, and data. The processing result collection unit 115 and the edge GW apparatus side influence determination and correction processing unit 101 are included.

The influence determination and corrective action unit 101 includes an impact determination and automatic corrective action unit 102, an edge GW profile storage unit 108, a suitability determiner 103, a device update treatment unit 104, an edge GW profile generation unit 105, and features. It is composed of a point extractor 106 and a profile temporary storage unit 107. The impact determination and corrective action unit 101 is also called an exchange coping means.

Each has the following roles.

The device update processing unit 104 receives a device update request from the device detection/driver update unit 113. That is, when the device detection/driver update unit 113 detects that a certain device (for example, the device s1) connected to the edge GW device 140 is replaced by another device (for example, the device z1), the device before and after the replacement is performed. The device update processing unit 104 is notified of the device update request with the information (for example, device name) for identifying the device update request. Upon receiving the device update request, the device update processing unit 104 stops the edge GW data processing unit 117, updates the device driver for a new device, and requests processing of the impact determination and automatic correction processing unit 102, The processed result is reflected in the edge GW data processing unit 117, and the process is restarted.

When requested to be processed by the device update processing means 104, the impact determination and automatic correction processing means 102 uses the conformity determination device 103 to determine whether or not the newly connected device can be used as a substitute for the original device. decide. At this time, if the newly connected device is a multi-sensor having a plurality of sensors (hereinafter referred to as sub-sensors), it is determined which sub-sensor of the newly connected device can be used as a substitute for the original device. .. After that, the impact determination and automatic corrective action means 102 corrects the normalization processing definition as necessary, and uses the edge GW data processing test section 126 based on the profile information stored in the edge GW profile storage section 108. Then, the modified definition is tested and the update on the edge GW is permitted.

Before the device is updated, the edge GW profile generation means 105 collects data on the data processing status in advance during normal operation, temporarily stores it in the profile temporary storage unit 107, and uses the feature point extractor 106 to extract the characteristics. The points are extracted, recorded in the profile, and stored in the edge GW profile storage unit 108.

Now, for example, the device s1 is a temperature sensor, the device s2 is a humidity sensor, and s1. a is temperature, s2. b represents humidity data, the edge GW data processing unit 117-1 normalizes the data as the state of the room, calculates the discomfort index of the room, and then transmits the data to the back-end server 130. It is assumed that the device data is processed by the edge GW data processing unit 117-2.

Here, referring to FIG. 4, for example, it is assumed that the temperature sensor s1 fails and a multi-sensor z1 capable of measuring temperature, humidity, and wind speed is connected as an alternative device. Even if data indicating z1.c is temperature, z1.d is humidity, and z1.e is wind speed, only data generated from the sensor z1 can be used as a substitute for s1.a by observing only the data generated from the sensor z1. I can't judge.

<Operation of this embodiment>
First, while all the devices including the temperature sensor s1 before the device replacement is normally operating, the edge GW profile generation unit 105 uses the device data collection unit 114 and the data processing result collection unit 115. Then, the data generated from the device driver 112, the data after the normalization process, and the data after the execution of the user logic are collected, and the edge GW profile (FIG. 5) is generated. The edge GW profile generation means 105 creates profile data (FIG. 5) while storing the generated data in the profile temporary storage unit 107 and extracting the characteristic points of the data using the characteristic point extractor 106.

Here, the feature point extractor 106 performs feature point extraction in accordance with the fitness determining unit 103 on the back end side. For example, when the fitness determination is performed using a statistical method, the feature point extractor 106 calculates statistical information as a feature point and records it in the profile. For example, the feature point extractor 106 records the daily average, variance, standard deviation, and slope (rate of change) in hour units as feature points in the profile.

The generated edge GW profile is periodically transmitted and stored in the edge GW profile storage unit 108 on the backend server side.

When the device s1 is replaced with the device z1 and the device z1 is connected to the edge GW apparatus 140, the device detection/driver update unit 113 detects it, and the device determination processing unit 104 of the impact determination and correction processing unit 101 sets the device s1 as the device. Notify the update request to update to z1.

Referring to FIG. 6, the requested impact determination and corrective action unit 101 operates as follows.

First, the device update processing unit 104 stops the edge GW data processing unit 117 that processes the data of the replaced device (F001). When the device s1 is replaced, only the edge GW data processing unit 117-1 that is inputting the data of the device s1 may be stopped. Next, the device update processing unit 104 acquires the device driver for the sensor z1 from the device driver repository 124 and applies it to the edge GW apparatus 140 (F002).

Next, the device update processing unit 104 makes an inquiry to the edge GW profile storage unit 108, and makes an inquiry about a time zone in which the feature point change is most remarkable at the same time as when the request was made and in the past (F003). For example, assuming that the temperature sensor s1 is replaced in May, a time zone in which the difference in temperature change is most remarkable in May is obtained from past data.

As a result, if results such as 8:00-9:00 and 17:00-18:30 are returned, the device data z1.c, z1.d, and z1.e generated from the device z1 are collected during that time period. Then, using the feature point extractor 106 for each of them, the same feature points as at the time of profile generation are extracted. Then, the collected data of the device z1 and the characteristic points of the device z1 are transmitted to the influence determination and automatic corrective action means 102 as information of the device after replacement and the temperature sensor s1 as information of the device before replacement (F003 to F006). ).

The influence determination and automatic corrective action means 102 receives the extracted feature points of the device z1, and uses the conformity determiner 103 to determine which of the device data z1.c, z1.d, and z1.e is the past data. It is determined whether the data has the characteristics most similar to the device data s1.a (f007). This similarity determination is performed by, for example, a feature vector including a predetermined number of feature amounts extracted from sensing data in a time zone in which the data changes drastically in the daily sensing data of the sensor before replacement, and the same time as the above time zone. It can be performed by calculating the inter-vector distance between a feature vector composed of a predetermined number of types of feature values extracted from the sensing data of the sensor after replacement in the band and comparing the inter-vector distance with a threshold value. It can also be realized by such a method.

For example, the impact determination and automatic corrective action means 102 determines whether the device data z1.c, z1.d, or z1.e is the device data z1.c. If only the data that falls within the range of the distribution of the device data s1.a and has the same slope in the time zone when the feature is extracted, the device data z1. The sensor related to c is determined as an alternative to the temperature sensor s1. When there are a plurality of device data that fall within the distribution range of the device data s1 and have the same slope in the time period when the feature is extracted, the impact determination and automatic correction processing means 102, for example, from the distribution, The data with the best degree of falling within the range and the degree of inclination are determined as alternatives. On the other hand, the influence determination and automatic corrective action means 102 determines that none of the device data z1.c, z1.d, and z1.e falls within the range of the distribution of the device data s1, or the time period when the feature is extracted. If it does not have the same inclination as the device data s1, it is determined that none of the device data z1.c, z1.d, z1.e cannot be used as a substitute for the temperature sensor s1. When it is determined that the device cannot be used, the effect determination and automatic corrective action unit 102 displays an alarm to that effect on a display device (not shown) or sends an alarm message to a terminal (not shown). Here, the device data z1. The description will be continued assuming that it is determined that c can be used as an alternative to the temperature sensor s1.

Device data z1. If it is determined that c can be used as an alternative to the temperature sensor s1, then the impact determination and automatic corrective action means 102 follows the normalization definition/data flow definition/user logic repository according to the configuration information described in the profile. The normalization definition, the user logic, and the data flow definition are acquired from 125 (F008). Next, ${s1. a} placeholder for ${z1. c} (F009), the edge GW data processing test unit 126 is used to test the processing according to the updated normalization definition, and it is confirmed that there is no problem. The data used in this test are the device data z1 and the device data s1 sent from the device update processing means 104. When the test is completed without any problem, the impact determination and automatic corrective action means 102 notifies the device update action means 104 of the completion of the action and the update of the normalization definition, and the device update action means 104 acquires the updated normalization definition rule1. Then, after reflecting on the edge GW data processing unit 117-1, the operation of the edge GW data processing unit 117-1 is restarted (F010-F014).

As a result, the new device z1 can be used in place of the device s1 to continue the data processing that has been performed so far.

<Effects of this embodiment>
According to this embodiment, it is possible to avoid an unexpected situation caused by replacement with an erroneous device such as an incompatible device. The reason is that when the replacement of the device s1 with the device z1 is detected, the influence determination and corrective action unit 101 determines the device z1 based on whether the data acquired from the device z1 is similar to the past data of the device s1. This is to determine whether to use the device s1 as a substitute.

Further, according to the present embodiment, even when the replaced device z1 is a multi-sensor, it is possible to avoid an unexpected situation caused by replacement with an erroneous device such as an incompatible device. The reason is that when the replacement of the device s1 with the multi-sensor z1 is detected, the influence determination and corrective action unit 101 determines which data of the plurality of data acquired from the device z1 is similar to the past data of the device s1. This is because it is determined whether or not any sensor of the device z1 is used as a substitute for the device s1.

Further, according to the present embodiment, even when the meaning of the data generated from the newly connected device is unknown, the data to be applied to the existing normalization definition is specified from the characteristics of the data actually generated, You can handle new devices by automatically modifying the normalization definitions.

Further, according to the present embodiment, from the data of the past characteristic points, the time zone in which the characteristic difference between the device data is most prominent is identified, and the actual data of the same time zone of the newly connected device is collected, By comparing with the characteristics, it is possible to determine which data is suitable as a substitute for the device data originally used. Therefore, the amount of data to be collected and transferred can be suppressed, and the time from the connection of a new device to the restart of data processing can be shortened compared to the case of collecting and determining the actual data for a long time. ..

[Second Embodiment]
The present embodiment is different from the first embodiment in the configurations and operations of the feature point extractor and the fitness determining unit, and the other configurations and operations are the same as in the first embodiment. In the present embodiment, the machine learning technology is applied to the feature point extractor and the goodness-of-fit determiner to learn the occurrence state of device data, classify the type of device data, and determine the type of data of the newly connected device. Specify whether to classify into.

FIG. 7 is a configuration diagram of main parts of the sensor management system according to the present embodiment. As shown in FIG. 7, in the present embodiment, the feature point extractor 106 functions as an identification model generation device, and the goodness of fit determiner 103 functions as an identification device. The feature point extractor 106 inputs the learning data in which the correct class is associated with the sensing data, which is the time-series signal acquired from each sensor before the sensor replacement, and generates the identification model to the suitability determiner 103. Send. The goodness of fit determiner 103 receives and captures this identification model. When the sensing data of the sensor after the sensor replacement is sent, the class to which the received sensing data belongs is identified according to the identification model.

For example, it is assumed that the temperature sensor s1, the humidity sensor s2, and the wind speed sensor s3 are connected to the edge GW device 140. The feature point extractor 106 of the edge GW device 140 outputs the learning data in which the correct classes (for example, the temperature class, the humidity class, and the wind speed class, respectively) are associated with the sensing data that are time-series signals acquired from these sensors. Capture and generate an identification model. The feature point extractor 106 transmits the generated identification model to the fitness determining unit 103 of the backend server 130. In addition, the feature point extractor 106 determines that the sensing data z1. c, z1. d, z1. e is transmitted to the fitness determining unit 103. The goodness-of-fit determiner 103 takes in the identification model and identifies the class of the sensing data of the sensor z1 that is the identification target data. The goodness-of-fit determiner 103 may detect the sensing data z1. c is a temperature class, sensing data z1. d is the humidity class, sensing data z1. e outputs the identification result of the wind speed class.

According to the present embodiment, not only the data of the type and characteristic of a specific device whose characteristics such as temperature and humidity are relatively easy to understand, but also the device of a type whose characteristic point extraction method cannot be easily specified Past data similar to the data obtained from can be determined. Also, by exchanging the feature point extractor and the fitness determining unit, it is possible to deal with various devices and data.

[Third Embodiment]
The present embodiment is different from the first embodiment in that there are a plurality of edge GW devices 140 and that the configuration and operation of the impact determination/automatic corrective action unit 102 and the edge GW profile storage unit 108 are different. Different from the first embodiment, other configurations and operations are the same as those of the first embodiment. In the present embodiment, information on devices that can be treated/supported by the edge GW apparatus 140 is recorded in the edge GW profile storage unit 108, and when the device is replaced in the same edge GW apparatus 140 or another edge GW apparatus 140, If the edge GW profile storage unit 108 has information on the device that is the same as or similar to the device exchange or that can handle the device exchange, the impact determination and the automatic corrective action can be omitted by utilizing the information.

FIG. 8 is a main part configuration diagram of the sensor management system according to the present embodiment. As shown in FIG. 8, in the present embodiment, the edge GW profile storage unit 108 records information on devices that can be treated and dealt with in the past. The device information includes information on a device before replacement, information on a device after replacement, and a normalized definition that has been automatically processed. After the end of step F002 in FIG. 6, the influence determination and automatic corrective action means 102 stores in the edge GW profile storage unit 108 the information of the device which can be dealt with or corresponded to the device replacement which is the same as or similar to the device replacement this time. Investigate whether or not. Next, the impact determination and automatic corrective action means 102 proceeds to the processing of step F003 and subsequent steps in FIG. 6 if it does not exist, but if it does exist, the processing of steps F003 to F012 is skipped and the processing of step F013 is performed. To execute. Then, in step F013, the automatically processed normalized definition recorded in the edge GW profile storage unit 108 is reflected in the data processing unit of the edge GW apparatus 140, and the edge GW data processing unit 117 is restarted.

According to this embodiment, the following effects can be obtained. That is, in the real world, since the same type of device is used in most cases in a similar environment, it is only necessary to execute the impact determination and corrective action processing of the first embodiment once, and the other edge GW device Can be dealt with, and the period from connecting a new device to performing processing using the new data can be shortened, except for the edge GW apparatus that deals first.

[Fourth Embodiment]
Referring to FIG. 9, a sensor management system 400 according to the fourth exemplary embodiment of the present invention is a system that manages a plurality of sensors 411 to 413, and includes a detection unit 401, an acquisition unit 402, and a replacement handling unit 403. ing.

The sensors 411 to 413 are, for example, temperature sensors, humidity sensors, wind sensors, and the like. The sensors 411 to 413 are communicably connected to the sensor management system 400 by wire or wirelessly.

The detection unit 401 has a function of detecting that the sensors 411 to 413 have been replaced. The detection unit 401 can be realized by dedicated hardware, or can be realized by a computer and a program.

The acquisition unit 402 has a function of acquiring sensing data from a sensor connected to the sensor management system 400. The acquisition unit 402 can be realized by dedicated hardware, or can be realized by a computer and a program.

The replacement handling unit 403 has a function of determining whether the sensing data of the sensor after replacement acquired by the acquisition unit 402 is similar to the past sensing data of the sensor before replacement. Further, the replacement coping unit 403 has a function of determining whether to use the sensor after replacement as a substitute for the sensor before replacement, based on the determination result. The exchange handling unit 403 can be realized by dedicated hardware, or can be realized by a computer and a program.

Next, a sensor management method executed by the sensor management system 400 according to this embodiment will be described with reference to FIG.

The detection unit 401 detects whether or not the sensors 411 to 413 have been replaced with another sensor while the sensor management system 400 is operating (F401). When detecting the sensor replacement, the detection unit 401 notifies the acquisition unit 402 of sensor replacement detection information including information for uniquely identifying the sensor before replacement and information for uniquely identifying the sensor after replacement. The identification information is a number string, a character string, or the like that can uniquely identify the sensor. For example, the USB address or MAC address assigned to the sensor may be used as the identification information, but the identification information is not limited thereto.

When the acquisition unit 402 receives the sensor replacement detection information from the detection unit 401, the acquisition unit 402 communicates with the sensor identified by the identification information of the replaced sensor included in the acquisition information, and acquires the sensing data from the replaced sensor. (F402). The acquisition unit 402 notifies the exchange handling unit 403 of the acquired sensing data and the sensor exchange detection information.

When the exchange handling unit 403 receives the sensing data of the sensor after the exchange and the sensor exchange detection information from the acquisition unit 402, the past sensing of the sensor specified by the identification information of the sensor before the exchange included in the sensor exchange detection information. The data is read from a storage device (not shown), and it is determined whether or not the sensing data of the sensor before and after the replacement are similar to each other (F403). Next, the replacement handling unit 403 determines whether to use the sensor after replacement as a substitute for the sensor before replacement, based on the determination result. Specifically, if the sensing data of the sensors before and after the replacement are similar to each other, the replacement handling unit 403 determines to use the sensor after the replacement as a substitute for the sensor before the replacement (F404). At this time, the replacement coping unit 403 may display a message to the effect that the sensor after replacement is used as a substitute for the sensor before replacement, on a display device (not shown), or notify an external device (not shown) by communication. On the other hand, if the sensing data of the sensor before and after the replacement are not similar to each other, the replacement handling unit 403 determines not to use the sensor after the replacement as a substitute for the sensor before the replacement (F404). At this time, the replacement coping unit 403 may display a message to the effect that the sensor after replacement is not used as a substitute for the sensor before replacement on a display device (not shown) and may notify the external device (not shown) by communication. After the above-described operation by the replacement handling unit 403, the above-described operation by the detection unit 401 is executed again.

As described above, according to the present embodiment, it is possible to avoid an unexpected situation caused by replacement with an incorrect sensor such as an incompatible sensor. The reason is that when sensor replacement is detected, sensing data is acquired from the sensor after replacement, it is determined whether it is similar to the past data of the sensor before replacement, and the sensor after replacement is based on the determination result. This is because if it is replaced with a wrong sensor such as an incompatible sensor, it will not be used as a replacement of the sensor before replacement.

Next, the detection unit 401, the acquisition unit 402, and the replacement handling unit 403 according to this embodiment will be described in detail.

<Detection unit 401>
FIG. 11 is an operation flowchart showing an example of the detection unit 401. Referring to FIG. 11, as a part of the initialization processing of the sensor management system 400, the detection unit 401 detects the identification information of the connected sensors 411 to 413 and registers it in a sensor list (not shown) (F411). For example, when the sensors 411 to 413 are connected to a computer constituting the sensor management system 400 through a USB interface, the sensor connected to the USB is searched, and, for example, the address assigned when the sensor is connected to the USB is identified. It is detected as information and registered in the sensor list. Further, when the sensors 411 to 413 are connected to the computer configuring the sensor management system 400 via a wireless interface such as Wi-Fi or Bluetooth (registered trademark), the sensors existing in the vicinity are searched or registered in advance. For example, the MAC address of the sensor is detected as the identification information of the sensor by referring to the list of connected sensors that are connected, and is registered in the sensor list.

After that, when the sensor management system 400 operates, the detection unit 401 constantly detects whether or not a sensor is connected to the sensor management system 400 (F412). Here, the connection of the sensor means, for example, that the sensor is connected and its power is turned on if it is a USB connection. Further, the connection of the sensor means, for example, receiving a connection request or data from the sensor if it is a wireless connection. Then, when detecting the connection of the sensor, the detection unit 401 executes the following processing.

First, the detection unit 401 detects the identification information of the connected sensor (F413), and determines whether the same identification information exists in the sensor list (F414). If the same identification information as the identification information of the connected sensor is present in the sensor list, the detection unit 401 does not detect the connected sensor because it has been reconnected (F415), and the step is performed. The process returns to F412.

On the other hand, when the same identification information as the identification information of the connected sensor does not exist in the sensor list, the detection unit 401 detects the connected sensor as the sensor after replacement because it is the added sensor (F416). Next, the detection unit 401 attempts communication with all the sensors whose identification information is described in the sensor list, and detects a sensor that cannot communicate (F417).

When the number of sensors that cannot communicate is only one (F418), the detection unit 401 detects the sensor that cannot communicate as a sensor before replacement (F419). Next, the detection unit 401 outputs sensor replacement detection information including identification information of the sensor after replacement detected in step F416 and the sensor before replacement detected in step F419 to the replacement handling unit 403 (F420). Then, the detection unit 401 returns to the process of step F412.

On the other hand, when the number of sensors that cannot communicate is not one, the detection unit 401 does not detect the connected sensor as a device after replacement (F415), and returns to the process of step F412. The reason is that if there is no sensor that cannot communicate, the sensor connected this time is considered to be an extension, not a replacement with an existing sensor. If there are two or more sensors that could not communicate, it is unclear which one is the sensor before replacement. Therefore, from the viewpoint of safety, exclude the sensor connected this time from the sensors after replacement. I have to. Here, by performing the sensor replacement work one sensor at a time, there is substantially no case of proceeding from step F418 to step F415.

As described above, the detection unit 401 of the example illustrated in FIG. 11 detects a sensor newly connected to the sensor management system 400 as a sensor after replacement, and is connected to the sensor management system 400 before connection of the sensor after replacement. Among the sensors, the sensor that cannot communicate with the sensor management system 400 after the replacement sensor is connected is detected as the sensor before replacement.

FIG. 12 is an operation flowchart showing another example of the detection unit 401. Referring to FIG. 12, as a part of the initialization process of the sensor management system 400, the detection unit 401 detects the identification information of the connected sensors 411 to 413 and registers it in a sensor list (not shown) (F421). This step F421 is the same as step F411 of FIG.

After that, when the sensor management system 400 operates, the detection unit 401 constantly detects whether or not an instruction to start replacement of the sensor has been input to the sensor management system 400 from an input device such as a keyboard (not shown) (F422). When the sensor replacement instruction is input, the detection unit 401 displays a sensor list on a display device (not shown) and prompts the user to select the sensor before replacement (F423).

Next, when the user selects one sensor on the sensor list (F424), the detection unit 401 detects the selected sensor as a pre-replacement sensor (F425). Next, the detection unit 401 detects whether or not the sensor is connected to the sensor management system 400 within a fixed time (F426). Here, the connection of the sensor means, for example, that the sensor is connected and its power is turned on if it is a USB connection. Further, the connection of the sensor means, for example, receiving a connection request or data from the sensor if it is a wireless connection.

When the sensor is connected within a fixed time, the detection unit 401 detects the identification information of the connected sensor as the identification information of the replaced sensor (F427). Next, the detection unit 401 outputs sensor replacement detection information including identification information of the sensor after replacement detected in step F427 and the sensor before replacement detected in step F425 to the replacement handling unit 403 (F428). Then, the detection unit 401 returns to the process of step F422. On the other hand, when the sensor is not connected within the fixed time, the detection unit 401 skips steps F427 and F428 and returns to the process of step F422.

As described above, the detection unit 401 in the example of FIG. 12 detects the sensor selected by the user from the list of sensors connected to the sensor management system 400 as the sensor before replacement, and then newly detects the sensor in the sensor management system 400. The connected sensor is detected as the sensor after replacement. However, the detection order of the sensors before and after replacement may be reversed. That is, the detection unit 401 detects the sensor newly connected to the sensor management system 400 as the sensor after replacement, and detects the sensor selected from the list of sensors as the sensor before replacement within a fixed time thereafter. May be.

<Acquisition unit 402>
The acquisition unit 402 includes a device driver that acquires sensing data from a sensor connected to the sensor management system 400. The acquisition unit 402 can autonomously acquire the sensing data from the sensor, and can also acquire the sensing data from the sensor requested according to the request from the replacement handling unit 403.

<Exchange handling unit 403>
FIG. 13 is a configuration diagram showing an example of the replacement handling unit 403. The replacement handling unit 403 in this example includes a feature amount extraction unit 431, a past data storage unit 432, a post-replacement sensor data storage unit 433, a collation unit 434, and a driver update unit 440.

The driver update unit 440 has a function of acquiring a device driver corresponding to the replaced sensor detected by the detection unit 401 from a device driver repository (not shown) and updating the device driver of the acquisition unit 402.

The feature amount extraction unit 431 has a function of extracting m types (m is an integer of 1 or more) of feature amounts from the sensing data of the sensor acquired using the acquisition unit 402. For example, the feature amount extraction unit 431 extracts the average value, the variance, the standard deviation, and the slope (rate of change) in hour units of the sensing data per day. An m-dimensional feature vector is formed by arranging m types of feature amounts in a predetermined order with one type of feature amount as one dimension.

The past data storage unit 432 has a function of storing the feature vector acquired by the acquisition unit 402 and extracted by the feature amount extraction unit 431 and generated before the replacement of the sensor in association with the identification information of the sensor. For example, the past data storage unit 432 stores the feature vector (F _1.1 , F _1.2 ,..., F _1.m ) corresponding to the sensor identification information s1. Here, F _1.i is the feature quantity of the i-th dimension acquired from the sensing data of the sensor identified by the sensor identification information s1. Further, the replaced sensor data storage unit 433 has a function of storing the feature vector acquired from the replaced sensor by the acquisition unit 402 and extracted by the feature amount extraction unit 431 in association with the identification information of the replaced sensor. Have. For example, the replaced sensor data storage unit 433 stores feature vectors (F _Z.1 , F _Z.2 ,..., F _Zm ) corresponding to the sensor identification information z1.

The collation unit 434 reads out the feature vector of the sensor after exchange from the sensor data storage unit 433 after exchange, and reads the feature vector of the sensor before exchange detected by the detection unit 401 from the past data storage unit 432. Then, the matching unit 434 calculates the inter-vector distance between the two feature vectors, compares the inter-vector distance with a threshold value, and determines whether the two are similar. The vector-to-vector distance is, for example, assigned a bit value of 0 if the feature values of the same dimension are the same or the absolute value of the difference between them is less than or equal to a predetermined value. Let the number be the distance between vectors. According to this inter-vector distance, if the feature quantities of both are the same or the absolute value of the difference between them is less than or equal to a predetermined value, the inter-vector distance becomes 0, and if they are the same or the absolute value of the difference is not less than or equal to the predetermined value, The value gradually increases from 1 as the feature amount increases. Therefore, the inter-vector distance is compared with a threshold value (for example, a value of 0), and if the inter-vector distance is less than or equal to the threshold value, it is determined to be similar, and otherwise it is determined to be dissimilar. If similar, the collation unit 434 determines to use the sensor after replacement as a substitute for the sensor before replacement, and if not similar, do not use the sensor after replacement as a substitute for the sensor before replacement. To decide. The collation unit 434 displays this determination result on a display device (not shown) or transmits it to an external device by communication.

As described above, the exchange handling unit 403 of FIG. 13 uses the feature vector including the feature amount of the predetermined type extracted from the sensing data of the sensor after the exchange and the feature amount of the predetermined type extracted from the sensing data of the sensor before the exchange. Based on whether or not the inter-vector distance to the feature vector is less than or equal to a threshold value, and based on the result of the determination, the sensor after replacement is used as a substitute for the sensor before replacement. Decide whether or not to do it.

FIG. 14 is a configuration diagram showing another example of the exchange handling unit 403. The exchange handling unit 403 in this example includes an identification model generation unit 435, a sensor identification information/class correspondence table 436, an identification model storage unit 437, an identification unit 438, a comparison unit 439, and a driver update unit 440. Among these, the driver updating unit 440 has the same function as the driver updating unit 440 of FIG.

The sensor identification information/class correspondence table 436 records the correspondence between the sensor identification information and the classes. For example, the sensor identification information/class correspondence table 436 records the temperature class in association with the identification information of the temperature sensor, records the humidity class in association with the identification information of the humidity sensor, and associates it with the identification information of the wind sensor. The wind class is recorded.

The identification model generation unit 435 inputs the sensing data that is a time-series signal acquired from each sensor by the acquisition unit 402 before the sensor replacement, and classifies the class corresponding to the identification information of the acquisition source sensor of the sensing data as the sensor identification information. A discriminant model is generated using learning data obtained from the class correspondence table 436 and in which the obtained sensing data is associated with the obtained class.

The identification unit 438 inputs the sensing data of the replaced sensor acquired by the acquisition unit 402, and uses the identification model stored in the identification model storage unit 437 to identify the class to which the input sensing data belongs. , And outputs the identified class to the comparison unit 439. The comparing unit 439 acquires the class corresponding to the identification information of the sensor before replacement detected by the detecting unit 401 from the sensor identification information/class correspondence table 436, and the acquired class and the class of the identification result of the identifying unit 438. To compare. Then, if the comparison result is a match, the comparing unit 439 determines that the sensing data of the sensor before and after the replacement are similar to each other, and determines to use the sensor after the replacement as a substitute for the sensor before the replacement. .. If the comparison result does not match, the comparing unit 439 determines that the sensing data of the sensor before and after the replacement are dissimilar to each other, and determines not to use the sensor after the replacement as a substitute for the sensor before the replacement. ..

As described above, the exchange coping unit 403 in FIG. 14 generates an identification model using the learning data in which the correct answer class is associated with the sensing data acquired from the sensor before the sensor is exchanged. The sensor to be replaced is identified using the sensing data obtained from the subsequent sensor and the identification model, and based on the result of comparison between the identified class and the class to which the sensor before replacement belongs, it is similar. It is determined whether or not the sensor after replacement is used as a substitute for the sensor before replacement based on the determination result.

[Fifth Embodiment]
The sensor management system 500 according to the fifth embodiment of the present invention is different from the sensor management system 400 according to the fourth embodiment in the function of the replacement coping unit 403, and other than that is the fourth embodiment. Is the same as

When the replacement coping unit 403 in the present embodiment determines to use the replaced sensor as a substitute for the sensor before replacement, it has a function of operating the system by using the sensor after replacement instead of the sensor before replacement. Have.

Next, a sensor management method executed by the sensor management system 500 according to the present embodiment will be described with reference to FIG.

15, steps F501 to F505 are the same as steps F401 to F405 in FIG. In the present embodiment, the replacement coping unit 403 operates the system using the replaced sensor as a replacement of the sensor before replacement after determining to use the replaced sensor as a replacement of the sensor before replacement. (F506). In step F506, the normalization definition is rewritten and tested as described in the first embodiment before the operation is started, and after confirming that there is no problem, the operation is started.

[Sixth Embodiment]
The sensor management system 600 according to the sixth embodiment of the present invention is different from the sensor management systems 400 and 500 according to the fourth or fifth embodiments in the function of the detection unit 401, and other than that is the first. It is the same as the fourth or fifth embodiment.

FIG. 16 is a main part flowchart showing an example of processing of the detection unit 401 in the present embodiment. When it is determined in step F418 of the flowchart shown in FIG. 11 that there is not one sensor that cannot perform communication, the detection unit 401 according to the present embodiment does not immediately proceed to step F415, but instead, step F601 to step F601 shown in FIG. It is configured to perform F605.

Referring to FIG. 16, when the number of sensors that cannot communicate is not one, the detection unit 401 determines whether the number of sensors that cannot communicate is 0 (F601). Next, the detection unit 401 proceeds to step F415 of FIG. 11 if the number of sensors that cannot communicate is 0. On the other hand, if there are two or more sensors that cannot communicate with each other, the detection unit 401 detects the positions of those sensors and the position of the sensor after replacement (F602). The detection unit 401 can detect the position of the sensor, for example, by extracting the position of the sensor from the sensing data acquired from the sensor. That is, generally, the sensing data of the sensor includes the position data of the sensor detected by the GPS mounted on the sensor in distinction from the data such as the temperature data detected by the sensor. The detection unit 401 detects the position of the sensor by extracting the position data.

Next, the detection unit 401 selects a sensor existing at the same position or in the vicinity of the replaced sensor from the plurality of sensors that cannot communicate (F603). Next, the detection unit 401 determines whether or not the selected sensor is only one (F604). If it is not only one unit but a plurality of units, the detection unit 401 proceeds to step F415 in FIG. If there is only one, the detection unit 401 detects that one as the sensor before replacement (F605). Then, the detection unit 401 proceeds to step F420 of FIG.

As described above, the detection unit 401 according to the present exemplary embodiment, when there are a plurality of sensors that cannot communicate with the sensor management system, detects the position information of the sensor that cannot communicate with the sensor management system and exchanges it with the sensor management system. The sensor before replacement is detected based on the result of comparison with the position information of the sensor.

Although the present invention has been described above with reference to some embodiments, the present invention is not limited to the above embodiments, and various additions and modifications can be made within the scope of the present invention.

INDUSTRIAL APPLICABILITY The present invention can be used as a base for general-purpose IoT data processing regardless of industry, and particularly as a base for managing sensors.

The whole or part of the exemplary embodiments disclosed above can be described as, but not limited to, the following supplementary notes.
[Appendix 1]
A sensor management system for managing a sensor,
Detection means for detecting replacement of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, based on the result of the determination, the sensor after replacement is replaced by the sensor before replacement. Replacement coping means for deciding whether or not to use as a substitute of
A sensor management system having.
[Appendix 2]
When the replaced sensor has a plurality of sub-sensors, the replacement coping means acquires sensing data from each of the plurality of sub-sensors, and is similar to the past sensing data of the sensor before replacement for each of the acquired sensing data. Decide whether to do,
The sensor management system according to attachment 1.
[Appendix 3]
When the replacement coping means determines to use the replaced sensor as a replacement for the sensor before replacement, the replaced sensor is used as a replacement for the sensor before replacement, and the system is operated.
The sensor management system according to attachment 1 or 2.
[Appendix 4]
When the replacement coping means decides to use the replaced sensor as a replacement for the sensor before replacement, before the operation of the system using the replaced sensor as a replacement for the sensor before replacement, Rewrite the normalization definition for normalizing the sensing data of the device to match the replaced device,
4. The sensor management system according to any one of appendices 1 to 3.
[Appendix 5]
The replacement coping means performs a test using the sensing data obtained from the device after replacement using the normalized definition after rewriting before operating the sensor after replacement as an alternative to the sensor before replacement. ,
The sensor management system according to attachment 4.
[Appendix 6]
The replacement coping means stores device information including information on a sensor after replacement that succeeded in the test, information on a sensor before replacement, and the normalization definition, and then stores information about a sensor connected to the sensor management system. When the exchange is detected by the detection means, based on the presence or absence of the device information including the detected sensor information after exchange and the sensor information before exchange, the normalization definition of Determine if there is a need for rewriting and performing the test,
The sensor management system according to attachment 5.
[Appendix 7]
The exchange coping means determines, in the determination of whether or not they are similar, a feature vector including a feature amount of a predetermined number of types extracted from the sensing data of the sensor after the exchange and a predetermined number of types extracted from the sensing data of the sensor before the exchange. Based on whether or not the inter-vector distance between the vector and the feature vector consisting of the feature amount is below the threshold,
7. The sensor management system according to any one of appendices 1 to 6.
[Appendix 8]
The replacement coping means includes a feature vector composed of a predetermined number of types of feature quantities extracted from sensing data in a time zone in which the data changes drastically among the daily sensing data of the sensor before the exchange, and the same time zone as the time zone. Based on whether or not the inter-vector distance between a feature vector composed of a predetermined number of feature quantities extracted from the sensing data of the sensor after replacement in is less than or equal to a threshold value, the similarity determination is performed.
7. The sensor management system according to any one of appendices 1 to 6.
[Appendix 9]
In the determination of the similarity, the exchange coping means generates an identification model using learning data in which the correct answer class is associated with the sensing data acquired from the sensor before the sensor is exchanged, and At the time of replacement, the sensing data obtained from the replaced sensor and the identification model are used to identify the class to which the replaced sensor belongs, and based on the result of comparison between the identified class and the class to which the sensor before replacement belongs. To determine the similarity,
7. The sensor management system according to any one of appendices 1 to 6.
[Appendix 10]
The detection unit detects a sensor newly connected to the sensor management system as a sensor after replacement, and the sensor after replacement out of the sensors connected to the sensor management system before connection of the sensor after replacement. After the connection of, the sensor that cannot communicate with the sensor management system is detected as the sensor before replacement,
The sensor management system according to any one of appendices 1 to 9.
[Appendix 11]
When there are a plurality of sensors that cannot communicate with the sensor management system, the detection unit has position information of the sensor that cannot communicate with the sensor management system and position information of the sensor after replacement. Detects the sensor before replacement based on the result of comparing
The sensor management system according to attachment 10.
[Appendix 12]
The detection means detects a sensor newly connected to the sensor management system as a sensor after replacement, and a sensor selected by the user from a list of sensors connected to the sensor management system as a sensor before replacement. Detect,
The sensor management system according to any one of appendices 1 to 9.
[Appendix 13]
A method for managing a sensor in a sensor management system, comprising:
Detects the replacement of the sensor connected to the sensor management system,
Obtain the sensing data from the replaced sensor,
Determine whether the acquired sensing data of the sensor after replacement is similar to the past sensing data of the sensor before replacement,
Based on the result of the determination, determine whether to use the sensor after replacement as a substitute for the sensor before replacement,
Sensor management method.
[Appendix 14]
In the determination of the similarity, when the sensor after replacement has a plurality of sub-sensors, sensing data is acquired from each of the plurality of sub-sensors, and the sensor data before replacement is acquired for each of the acquired sensing data. Determine whether it is similar to the sensing data,
The sensor management method according to attachment 13.
[Appendix 15]
If you decide to use the replaced sensor as a replacement for the sensor before replacement, operate the system using the replaced sensor as a replacement for the sensor before replacement.
The sensor management method according to appendix 13 or 14.
[Appendix 16]
If you decide to use the replaced sensor as a replacement for the replaced sensor, before you operate the system using the replaced sensor as a replacement for the replaced sensor, normalize the sensing data of the replaced device. Rewrite the normalization definition to make it compatible with the replaced device,
16. The sensor management method according to any one of appendices 13 to 15.
[Appendix 17]
Before operating the sensor after replacement as an alternative to the sensor before replacement, perform a test using the sensing data obtained from the device after replacement using the normalized definition after rewriting,
The sensor management method according to attachment 16.
[Appendix 18]
The device information including the information of the sensor after the replacement that succeeded in the test, the information of the sensor before the replacement, and the normalization definition is stored, and thereafter, when the replacement of the sensor connected to the sensor management system is detected, the detection is performed. Based on the presence or absence of the device information including the information of the sensor after the replacement and the information of the sensor before the replacement, it is determined whether it is necessary to rewrite the normalization definition and perform the test accompanying the current sensor replacement. To do
The sensor management method according to attachment 17.
[Appendix 19]
In the determination of whether or not they are similar to each other, a feature vector composed of a feature quantity of a predetermined type extracted from the sensing data of the sensor after replacement and a feature vector composed of a feature quantity of a predetermined type extracted from the sensing data of the sensor before replacement Based on whether the distance between the vector and the vector is less than or equal to the threshold value, the similarity determination is performed.
19. The sensor management method according to any one of appendices 13 to 18.
[Appendix 20]
In the determination as to whether or not they are similar to each other, a feature vector including a predetermined number of types of feature amounts extracted from sensing data in a time zone in which the data changes drastically among the daily sensing data of the sensor before replacement, and the time zone Based on whether or not the inter-vector distance between the feature vector composed of the feature amount of a predetermined type number extracted from the sensing data of the sensor after replacement in the same time zone is less than or equal to a threshold value, the similar determination is performed.
19. The sensor management method according to any one of appendices 13 to 18.
[Appendix 21]
In the determination of whether or not the similarity, before the replacement of the sensor, an identification model is generated using learning data in which the correct answer class is associated with the sensing data acquired from the sensor, and when the sensor is replaced, after the replacement, The sensing data obtained from the sensor and the identification model are used to identify the class to which the replaced sensor belongs, and based on the comparison result between the identified class and the class to which the sensor before replacement belongs, the similarity determination I do,
19. The sensor management method according to any one of appendices 13 to 18.
[Appendix 22]
In the detection of the replacement of the sensor, the sensor newly connected to the sensor management system is detected as a sensor after replacement, and the replacement of the sensors connected to the sensor management system before the connection of the replaced sensor is performed. Detects the sensor that cannot communicate with the sensor management system after the later sensor is connected as the sensor before replacement,
22. The sensor management method according to any one of appendices 13 to 21.
[Appendix 23]
When there are a plurality of sensors that cannot communicate with the sensor management system, the result of comparing the position information of the sensor that cannot communicate with the sensor management system and the position information of the sensor after replacement. Detect the sensor before replacement,
The sensor management method according to attachment 22.
[Appendix 24]
In the detection of the replacement of the sensor, the sensor newly connected to the sensor management system is detected as the sensor after replacement, and the sensor selected by the user from the list of the sensors connected to the sensor management system is replaced before the replacement. As a sensor of
22. The sensor management method according to any one of appendices 13 to 21.
[Appendix 25]
Computer,
Detection means for detecting replacement of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, and the sensor after replacement is replaced with the sensor before replacement based on the result of the determination. Replacement coping means for deciding whether or not to use as a substitute of
And a program to make it work.

100... Sensor management system 101... Impact determination and corrective action section 102... Impact determination and automatic correction processing means 103... Fitness determination device 104... Device update action means 105... Edge GW profile generation means 106... Feature point extractor 107... Profile Temporary storage unit 108... Edge GW profile storage unit 111-1... Device s1
111-2... Device s2
111-N... Device N
112-1... Device driver 1
112-2... Device driver 2
112-N... Device driver N
113... Device detection/driver update unit 114... Device data collection unit 115... Data processing result collection unit 116... Data transmission unit 117-1... Edge GW data processing unit 1
117-2... Edge GW data processing unit 2
121... Data receiving section 122... Data processing section 123... Data storage section 124... Device driver repository 125... Normalization definition/data flow definition/user logic repository 126... Edge GW data processing test section 400... Sensor management system 401... Detection section 402... Acquisition unit 403... Replacement handling unit 411... Sensor 412... Sensor 413... Sensor 431... Feature amount extraction unit 432... Past data storage unit 433... After replacement sensor data storage unit 434... Collation unit 435... Identification model generation unit 436... Sensor identification information/class correspondence table 437...Identification model storage unit 438...Identification unit 439...Comparison unit 440...Driver update unit

Claims (10)

A sensor management system for managing a sensor,
Detection means for detecting replacement of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, and the sensor after replacement is replaced with the sensor before replacement based on the result of the determination. Replacement coping means for deciding whether or not to use as a substitute of
A sensor management system having. When the replaced sensor has a plurality of sub-sensors, the replacement coping means acquires sensing data from each of the plurality of sub-sensors, and is similar to the past sensing data of the sensor before replacement for each of the acquired sensing data. Decide whether to do,
The sensor management system according to claim 1. When the replacement coping means determines to use the replaced sensor as a replacement for the sensor before replacement, the replaced sensor is used as a replacement for the sensor before replacement, and the system is operated.
The sensor management system according to claim 1. When the replacement coping means decides to use the replaced sensor as a replacement for the sensor before replacement, before the operation of the system using the replaced sensor as a replacement for the sensor before replacement, Rewrite the normalization definition for normalizing the sensing data of the device to match the replaced device,
The sensor management system according to claim 1. The replacement coping means performs a test using the sensing data obtained from the device after replacement using the normalized definition after rewriting before operating the sensor after replacement as an alternative to the sensor before replacement. ,
The sensor management system according to claim 4. The replacement coping means stores device information including information on a sensor after replacement that succeeded in the test, information on a sensor before replacement, and the normalization definition, and then stores information about a sensor connected to the sensor management system. When the exchange is detected by the detection means, based on the presence or absence of the device information including the detected sensor information after exchange and the sensor information before exchange, the normalization definition of Determine if there is a need for rewriting and performing the test,
The sensor management system according to claim 5. The replacement coping means includes a feature vector composed of a predetermined number of types of feature quantities extracted from sensing data in a time zone in which the data changes drastically among the daily sensing data of the sensor before the exchange, and the same time zone as the time zone. Based on whether or not the inter-vector distance between a feature vector composed of a predetermined number of feature quantities extracted from the sensing data of the sensor after replacement in is less than or equal to a threshold value, the similarity determination is performed.
The sensor management system according to claim 1. The detection unit detects a sensor newly connected to the sensor management system as a sensor after replacement, and the sensor after replacement out of the sensors connected to the sensor management system before connection of the sensor after replacement. After the connection of, the sensor that cannot communicate with the sensor management system is detected as the sensor before replacement,
The sensor management system according to claim 1. A method for managing a sensor in a sensor management system, comprising:
Detects the replacement of the sensor connected to the sensor management system,
Obtain the sensing data from the replaced sensor,
Determine whether the acquired sensing data of the sensor after replacement is similar to the past sensing data of the sensor before replacement,
Based on the result of the determination, determine whether to use the sensor after replacement as a substitute for the sensor before replacement,
Sensor management method. Computer,
Detection means for detecting replacement of the sensor connected to the sensor management system,
An acquisition unit that acquires sensing data from the sensor,
It is determined whether the sensing data of the sensor after replacement acquired by the acquisition unit is similar to the past sensing data of the sensor before replacement, based on the result of the determination, the sensor after replacement is replaced by the sensor before replacement. A replacement coping means for deciding whether or not to use as an alternative to
And a program to make it work.

Images