KR101971553B1 - Device management system and method based on Internet Of Things - Google Patents

Abstract

[0001] The present invention relates to a system and method for managing Internet based devices, and more particularly, Management system and method. More particularly, the present invention relates to a method and apparatus for receiving sensor information related to a monitoring target device, storing time series sensor information immediately before an error of a device corresponding to a category to which the monitoring target device belongs as reference data, And to perform an error countermeasure operation when an error risk of the monitoring target device is detected based on the calculated similarity.

Description

[0001] The present invention relates to an Internet-based device management system and method,

[0001] The present invention relates to a system and method for managing Internet-based devices, and more particularly, it relates to a system and method for managing Internet-based devices, which can diagnose a device in advance and detect errors or failures before they occur so that damage can be prevented And more particularly,

Background Art [0002] With the recent increase in the use of electronic devices, there is an increasing interest in systems for controlling electronic devices for safer and more efficient use of electronic devices. Such an electronic device may cause various kinds of failures for various reasons. In the case of an electronic device which is continuously operated, if a failure occurs, the operation may be interrupted to cause various damages, and secondary damage There is a high possibility. If the device is diagnosed in advance and it is detected before an error or failure occurs, taking measures beforehand can prevent the damage caused by the failure of the device.

Conventionally, Korean Patent No. 10-1703163 entitled " Apparatus and Method for Predicting Multiple Faults of a Vehicle " discloses a multi-artificial neural network and a regression analysis method, which are data mining techniques, Is predicted.

The above prior art includes a configuration in which a multi-artificial neural network model is learned and modeled according to the characteristics of the vehicle model to predict the failure of the vehicle. Techniques applied to the system by modeling using artificial neural network technology are time- There is a problem that it takes a lot of time for data collection and modeling work.

Therefore, there is a need for a failure prediction technique that can be applied immediately without any time and cost-consuming modeling work, which can diagnose the device in advance and detect it before a failure occurs and take action.

Korean Patent No. 10-1703163

Since the present invention can be applied immediately without modeling operation, it is possible to find general patterns by generating metadata by the sensor data pattern transition, and it is possible to detect errors or failures of devices more quickly by saving cost and time, To prevent the damage caused by

According to the present invention, the first degree of similarity and the second degree of similarity are calculated for each of the reference data and the sensor data, and whether or not an error-correcting operation is performed is determined based on the general similarity in which the first and second similarities are combined. It is aimed to accurately judge the judgment of the failure.

The present invention is characterized in that the first degree of similarity calculation unit determines the sign of the difference value between the specific data in the time series data and the data before two points of the specific data for each of the reference data and the sensor data, Is to calculate the ratio of the case where the number of the devices is the same, thereby detecting errors or failures of the device by simple calculation without modeling work.

The present invention is characterized in that a second degree of similarity calculation unit calculates an average of calculation results by using a difference value between specific data in the time series data and data before one time point of the specific data for each of the reference data and the sensor data, The purpose of the calculation is to detect errors or failures of the device by simple calculation without modeling work.

The present invention collects sensor information related to a plurality of devices, stores sensor information immediately before occurrence of an error in the collected sensor information as preliminary reference data, groups preliminary reference data having a similar pattern among preliminary reference data, And generating the reference data from among the preliminary reference data generated by the preliminary reference data generating unit, so that the reference data having the highest accuracy in the grouped group can be generated to more accurately measure the similarity between the reference data and the sensor data.

In the present invention, two pairs of combinations of preliminary reference data are generated, the data of each combination is extracted, and the similarity degree is calculated. When the calculated similarity degree is equal to or greater than a predetermined reference value, groups of similar preliminary reference data can be grouped .

According to an aspect of the present invention, there is provided a object Internet-based device management system including a sensor data receiver for receiving sensor information related to a monitoring target device, A similarity calculator for calculating the similarity between the sensor data and the reference data, and a comparator for calculating a similarity between the sensor data and the reference data based on the similarity calculated by the similarity calculator, And a control unit for controlling to perform an error countermeasure operation when it is detected.

The similarity calculation unit may calculate a difference between specific data of the time series data and data of the two data before the specific data for each of the reference data and the sensor data, Calculating a difference between the specific data in the time series data and the data in the preceding one time point of the specific data for each of the reference data and the sensor data and using the similarity between the calculated values The apparatus of claim 1, further comprising a second degree of similarity calculation unit for calculating a second degree of similarity, and the control unit may further include determining whether to perform an error corresponding operation based on the general similarity in which the first similarity degree and the second similarity degree are combined have.

The first degree of similarity calculation unit may determine whether the difference between the specific data in the time series data and the data before two points of time of the specific data is +, 0, or - for the reference data and the sensor data, And calculating the ratio of the case where the sign of the difference value between the reference data and the sensor data is the same as the first similarity.

The second similarity calculator calculates the following equation using the difference between the specific data in the time series data and the data in the preceding one of the specific data for each of the reference data and the sensor data, Difference value of data before one point in time of specific data of one-sensor data / difference value of data before one point of specific data of reference data X100), and calculating the average of the calculation results of the above formula by the second similarity degree .

The sensor unit may further include a collecting unit that collects sensor information related to a plurality of devices, a preliminary reference data storage unit that stores, as preliminary reference data, sensor information of a predetermined number of viewpoints immediately before the error of the device, A plurality of preliminary reference data having a high degree of mutual similarity among the plurality of preliminary reference data stored in the preliminary reference data storage unit, grouping the plurality of preliminary reference data having a high degree of mutual similarity, generating reference data related to the grouped plurality of preliminary reference data, And a reference data generator for storing the reference data.

The reference data generation unit may generate a combination of two pairs of the preliminary reference data, extract the data of each combination, calculate the similarity in the similarity calculation unit, and determine whether the calculated similarity is equal to or greater than a predetermined reference value And grouping them into the same group.

Since the present invention can be applied immediately without modeling operation, it is possible to find general patterns by generating metadata by the sensor data pattern transition, and it is possible to detect errors or failures of devices more quickly by saving cost and time, It is possible to prevent the damage caused in advance.

According to the present invention, the first degree of similarity and the second degree of similarity are calculated for each of the reference data and the sensor data, and whether or not an error-correcting operation is performed is determined based on the general similarity in which the first and second similarities are combined. It is possible to accurately judge the judgment of the failure.

The present invention is characterized in that the first degree of similarity calculation unit determines the sign of the difference value between the specific data in the time series data and the data before two points of the specific data for each of the reference data and the sensor data, , It is possible to detect errors or failures of the apparatus by simple calculation without modeling work.

The present invention is characterized in that a second degree of similarity calculation unit calculates an average of calculation results by using a difference value between specific data in the time series data and data before one time point of the specific data for each of the reference data and the sensor data, By calculation, errors or failures of the instrument can be detected with simple calculations without modeling.

The present invention collects sensor information related to a plurality of devices, stores sensor information immediately before occurrence of an error in the collected sensor information as preliminary reference data, groups preliminary reference data having a similar pattern among preliminary reference data, It is possible to more accurately measure the similarity between the reference data and the sensor data by generating the reference data having the highest accuracy in the grouped group by generating the reference data from the preliminary reference data.

In the present invention, two pairs of combinations of preliminary reference data are generated, the data of each combination is extracted, and the similarity degree is calculated. When the calculated similarity degree is equal to or greater than a predetermined reference value, groups of similar preliminary reference data can be grouped have.

FIG. 1 is a block diagram of an object-based Internet-based device management system according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating the overall relationship between the cloud and the object Internet in the object Internet-based device management system according to the embodiment of the present invention.
3 is a diagram showing an example of the similarity between reference data and sensor data according to an embodiment of the present invention.
FIG. 4 is a view showing an example of mutual similarity in a combination of two pairs of preliminary reference data according to an embodiment of the present invention.
FIG. 5 is a flowchart illustrating a flow of a method for managing Internet-based devices according to an embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. Further, in describing the embodiments of the present invention, specific numerical values are merely examples and the scope of the invention is not limited thereby.

The object Internet-based device management system according to the present invention may include a central processing unit (CPU) and a memory (memory), and may be configured in the form of a server connectable to other terminals through a communication network such as the Internet. However, the present invention is not limited by the configuration of the central processing unit and the memory. In addition, the object Internet-based device management system according to the present invention may be physically configured as a single device or may be implemented in a distributed manner in a plurality of devices. Or more.

FIG. 1 is a block diagram of an object-based Internet-based device management system according to an embodiment of the present invention.

The object Internet-based device management system according to the present invention includes a sensor data receiving unit 101, a reference data storage unit 102, a similarity calculation unit 103, a control unit 104, a first similarity calculation unit 105, A calculating unit 106, a collecting unit 107, a preliminary reference data storing unit 108, and a reference data generating unit 109. [

The sensor data receiving unit 101 receives sensor information related to the monitoring target device.

The monitored device may be any electronic or mechanical device commonly found in the vicinity, such as a vehicle, a traffic light, a refrigerator, a washing machine, etc., and may be connected to the Internet. A monitoring target device is a device that the user wants to detect a failure or an error, and is a target to prevent damage due to failure of the device by detecting a failure in advance. Sensors included in or external to the monitoring target device can detect the status information or environment information of the device. The sensor information includes all information that can be measured by the sensor, such as current information, temperature information, humidity information, And the sensor can recognize and record sensor information at regular intervals.

The reference data storage unit 102 stores, as reference data, time series sensor information immediately before occurrence of an error of the device corresponding to the category to which the monitoring target device belongs.

The monitoring target device can be classified into categories according to the features or types of the device, such as the type of device, the product number of the device, and the type of sensor included in the device. The monitoring target device can be classified into a plurality of categories.

The sensor records information about the sensor information at regular intervals, and when an error or failure occurs in the apparatus, it is possible to extract information on a certain time series sensor immediately before an error or a failure occurs. The extracted sensor information may be stored in the reference data storage unit 102 as reference data.

The similarity calculation unit 103 calculates the similarity between the sensor data and the reference data.

The sensor data is data in which sensor information related to a monitoring target device is recorded, and may be data in which sensor information is continuously recorded at regular intervals to continuously monitor the device in case the user generates an error or a failure.

The reference data is the data to be compared with the sensor data, and is data in which the data in the case where an error or a failure has already occurred is recorded and the sensor information is changed in a time-series manner immediately before an error or a failure occurs. For example, when the category of the monitoring target device A is a washing machine and the product number is 'AB-100', the reference data storage unit 102 stores a malfunction of a device whose category is a washing machine and whose product number is 'AB-100' Will be stored for the predetermined time series sensor information just before the occurrence, which will be the reference data. Most equipment failures will not occur suddenly and will show some pattern in the information gathered by the sensor until failure occurs. Therefore, it is possible to see the sensor information immediately before the fault occurrence in a time series for comparison with the previous symptom and the monitoring target device.

If there is a possibility that a fault will occur in a monitoring target device, the sensor of the device will measure it and the sensor data recorded in a time series can be compared with the reference data to measure the similarity. If the monitoring target device shows a pattern similar to the pattern shown in the time series sensor information just before the error of the reference data does not occur but the failure of the monitoring target device is likely to occur, the sensor data and the reference data can be compared and calculated.

The similarity calculation unit 103 is based on a core of a pattern matching analysis system based on a motif algorithm. The motif algorithm method uses an intuitive approach such as a Google search engine to calculate a general pattern of the object Internet data And a method that is used to find the pattern of DNA in a way that can detect the appearance of the pattern in industrial applications is applied as a predictive phenomenon prediction model in object Internet service.

Motif Algorithm Core functionality development can implement spark-based data collection and basic data classification, and Spark is a general-purpose distributed platform that only runs Map & Reduce like Hadoop. It is a general purpose distributed clustering platform that allows not only streaming but also distributed nodes to perform operations on multiple notes. It also adds modules such as Map & Reduce and streaming processing, And a base function for basic big data management can be implemented.

The pattern search function can implement the function of searching the time series sensor data for metadata that is a library of changes in sensor data. For example, CACGTG can be defined and managed as one metadata among various data Lt; / RTI >

The first degree-of-similarity calculation unit 105 calculates the difference between the specific data in the time-series data and the data before two points of the specific data for each of the reference data and the sensor data, and calculates the first degree of similarity using the degree of similarity between the calculated values .

The specific data in the time series data for each of the reference data and the sensor data may be data at one point in the time series data and may be data corresponding to one interval of the data in which the sensor records information of the apparatus in a time interval . For example, if the reference data is 15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, , 5, 20, etc. may be specific data. Also, in general, in general cases where no errors or failures occur, specific data is often maintained in a certain pattern or a certain level of data value. In general, when a fault or failure occurs, Other data will be derived.

The first degree-of-similarity calculation unit 105 determines whether the difference between the specific data in the time-series data and the data before two points in time of the specific data is +, 0, or - for the reference data and the sensor data, The ratio of the case where the sign of the difference value between the reference data and the sensor data is the same is calculated by the first similarity degree.

The first degree-of-similarity calculation unit 105 calculates the difference between the specific data in the time-series data and the data before two points of the specific data with respect to the reference data and the sensor data, respectively. For example, the reference data is {15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, 15, 15, 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, when two points in front of each specific data are calculated, Except for the previous two points, the reference data is {0, 0, 0, 10, -15, 5, 5, -10, 5, -5, 5, 0, 0, 0} The sensor data is {0, 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0, 0} except for the previous two points of data. However, in the general case in which no error or failure occurs, it can be estimated that both the reference data and the sensor data are the portion of the specific data 15, and both the reference data and the sensor data have errors or failures The part that is assumed not to be done may be omitted. Therefore, the difference between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {10, -15, 5, 5, -10, 5, -5, 5} The difference between the data and the data before two points of the specific data is {7, -12, 13, 0, -15, 4, -5, 8}. 5, 5, 5, 5, 5, 5, 0, 0, 0} and {10, -15, 5, 5, -10 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0 , 0} and {7, -12, 13, 0, -15, 4, -5, 8} can be treated as the same value to exclude specific data judged not to be related to error or failure.

If the difference between the specific data of the time series data of the reference data and the sensor data and the data before the two points of time of the specific data increases, it can be displayed as + or 1, That is, whether or not increase, decrease, or decrease can be indicated by a sign.

The sign of the difference value between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {1, -1, 1, 1, -1, 1, -1, 1} The sign of the difference between the specific data of the data and the data before the two points of time of the specific data is {1, -1, 1, 0, -1, 1, -1, 1}.

The first similarity degree is calculated at a ratio of the case where the sign of the difference value between the reference data and the sensor data is the same, and the code similarity degree of the reference data and the sensor data is {identical, same, same, same, same, same, }, Which is 7 out of the 8 arrays, the first similarity can be calculated as (7/8) * 100 = 87.5%.

The second similarity calculation unit 106 calculates the difference between the specific data of the time series data and the data of the one point before the specific data with respect to each of the reference data and the sensor data and calculates the second similarity by using the similarity between the calculated values .

For example, the reference data is {15,15,15,15,15,25,0,30,50,15,15,15,15,15} and the sensor data is {15,15,15,15,15,15,15} 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, and calculates the difference between the specific data of the time series data and the data of the specific date 0, 0, 0, 10, -25, 30, -25, 15, -10, 5, 0, 0, 0, 0} and the sensor data is {0, 0, 0 , 0, 7, -19, 32, -32, 17, -13, 8, 0, 0, 0, 0}. The reference data is {10, -25, 30, -25, 15, -10, 5} and the sensor data is {7, -19 , 32, -32, 17, -13, 8}.

The second similarity calculation unit 106 calculates the following equation using the difference value between the specific data of the time series data and the data of one time before the specific data for each of the reference data and the sensor data,

The average of the calculation results of the above equations is calculated as the second similarity.

To calculate the second degree of similarity, the result using the formula is {70, 76, 93.3, 72, 86.6, 70, 40}, and the average can be calculated as 72.56, and the second similarity is 72.56%.

The control unit 104 controls the error counter operation to be performed when an error risk of the monitoring target device is detected based on the similarity calculated by the similarity calculation unit 103.

The error countermeasure operation may be an action to prevent an error or a failure from occurring before a failure or malfunction of the device occurs, such as an operation of turning off the power of the monitoring target device or an operation of checking the system, If a danger is detected, control can be performed to perform an error countermeasure operation such as an operation of turning off the monitoring target device or performing a system check.

The control unit 104 determines whether or not an error corresponding operation is performed based on the general similarity combining the first similarity and the second similarity.

The similarity calculation unit 103 may calculate the overall similarity by combining the first similarity and the second similarity. If the control unit 104 becomes equal to or higher than a predetermined reference based on the calculated general similarity, the error risk of the monitoring target device is detected It can be judged. The overall similarity can be calculated by various methods, and the control unit 104 can detect an error risk of the monitoring target device based on the calculated similarity.

The general similarity degree may be calculated by comparing the first similarity degree and the second similarity degree with each other. For example, it is assumed that the first degree of similarity is 87.5%, the second degree of similarity is 72.56%, and the reference value is 70%. The first similarity degree and the second similarity degree are both determined to be equal to or greater than the reference value of 70%, and the controller 104 can determine that the error risk of the monitoring target device is detected. However, assuming that the reference value is 80%, the first similarity degree is higher than the reference value, the second similarity degree is lower than the reference value, and neither the first similarity degree nor the second similarity degree is greater than the reference value. It can be judged.

In addition, the general similarity degree is calculated as an average of the first similarity degree and the second similarity degree, and when the average value is equal to or greater than the reference value, the control section 104 can determine that the monitoring target device detects an error risk. Assuming, for example, that the first degree of similarity is 87.5% and the second degree of similarity is 72.56%, it is assumed that the reference value for determining that the monitoring target device detects an error risk is 80%. The average of the first and second similarities is 80.03%, and the reference value is 80%, so that the controller 104 can determine that the error risk of the monitoring target device is detected.

In addition, the control unit 104 may determine that the monitoring target device detects an error risk when at least one of the first similarity degree and the second similarity degree is equal to or greater than the reference value. For example, it is assumed that the first degree of similarity is 87.5%, the second degree of similarity is 72.56%, and the reference value is 80%. The second similarity is not equal to or greater than the reference value, but the first similarity is determined to be equal to or greater than the reference value, and the controller 104 may determine that the error risk of the monitoring target device is detected. However, if the reference value is 90%, the control unit 104 can determine that the error risk of the monitoring target device is not detected because the reference value is less than the first similarity and the second similarity.

In addition, the general similarity degree can be calculated by weighting each of the first similarity degree and the second similarity degree. The control unit 104 can determine that the error risk of the monitoring target device is detected when the total similarity degree is equal to or greater than the reference value have. For example, it is assumed that the first degree of similarity is 87.5%, the second degree of similarity is 72.56%, and the weight of the first degree of similarity and the second degree of similarity are respectively 70%: 30% The reference value for judging the error risk is assumed to be 80% or more. (87.5X0.7) + (72.56X0.3) = 83.02 as a result of calculating the total similarity degree, the controller 104 can determine that the error risk of the monitoring target device is detected at the reference value of 80% or more.

The collecting unit 107 collects sensor information related to a plurality of devices.

The collecting unit 107 collects sensor information related to various electronic devices, such as a vehicle, a traffic light, a refrigerator, a washing machine, etc., The sensor information may be information on all the sensors included in the apparatus and may be information recorded in a time interval at a predetermined interval or at a specific interval.

The preliminary reference data storage unit 108 and the preliminary reference data storage unit 108 store the sensor information of the predetermined number of points of the sensor information collected by the collecting unit 107 just before the error of the device as preliminary reference data.

Sensor information of the device when an error occurs in the sensor information may be recorded and sensor information of a certain point of time immediately before the error of the device may be stored as reference data. The reference data is data to be compared in case of occurrence of errors or failures of other devices. As more preliminary reference data exist, more reference data can be extracted because there are more objects to be compared, so that errors or failures can be predicted more quickly have.

The reference data generation unit 109 groups the plurality of preliminary reference data having a high degree of similarity among the plurality of preliminary reference data stored in the preliminary reference data storage units 108 and 102, And stores the generated data in the reference data storage unit 102.

The reference data generation unit 109 generates a combination of two pairs of the preliminary reference data, extracts the data of each combination, calculates the similarity by the similarity calculation unit 103, If they are more than the reference value, they are grouped into the same group.

There are various patterns of preliminary data in which sensor information immediately before occurrence of error of various devices is stored, and similar patterns among various patterns can be grouped. Grouped preliminary reference data The preliminary reference data having the highest sum of similarities can be generated as reference data by comparing the similarity with other preliminary reference data. The similarity calculated through the similarity calculation unit 103 can be grouped into the same group if the similarity is equal to or greater than a predetermined reference value. For example, if the predetermined reference value is 70, all the reference data in the same group are mated If we compare the similarities, we can say that the similarity can not be more than 70.

For example, in group A, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped. The preliminary reference data 1 has a similarity degree of 90 to the preliminary reference data 2, a similarity to the preliminary reference data 3 is 80, a similarity to the preliminary reference data 4 is 70, a total of 240, and the preliminary reference data 2 has 90 , The degree of similarity to the preliminary reference data 3 is 70, the degree of similarity to the preliminary reference data 4 is 75, and the total is 235. The preliminary reference data 3 has a similarity to the preliminary reference data 1 of 80, a similarity to the preliminary reference data 2 of 70, 4 and the similarity degree is 70, the sum of the preliminary reference data 4 is 70, the degree of similarity with the preliminary reference data 2 is 75, the similarity with the preliminary reference data 3 is 70, and the sum is 215, 1 is the highest with respect to the other preliminary reference data, the reference data of the group A is the preliminary reference data 1, and the preliminary reference data 1 is stored in the reference data storage unit 10 2). ≪ / RTI >

However, when a plurality of preliminary reference data are grouped and it is judged that other preliminary reference data can be entered into a group already grouped, the preliminary reference data is compared with the reference data generated as reference data, The preliminary reference data can be grouped.

For example, if a predetermined value capable of entering an already-grouped group is 90, if it is judged that the other spare baseline data 10 can be grouped into the group A while being grouped into the group A, The preliminary reference data 1 and the preliminary reference data 10 can be grouped into the group A if the similarity degree is 90 or more.

In addition, a plurality of preliminary reference data can be integrated to generate virtual reference data having a high degree of similarity to the grouped preliminary reference data. For example, in Group B, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped, and all four preliminary reference data and virtual reference data Lt; / RTI >

FIG. 2 is a diagram illustrating the overall relationship between the cloud and the object Internet in the object Internet-based device management system according to the embodiment of the present invention.

In the cloud, the collecting unit 107 may collect sensor information of devices of various objects on the Internet, such as the Internet 2 and the Internet 3, and the collected sensor information may include sensor information at a point in time immediately before the error of the device Can be stored as reference data. The preliminary reference data having a high degree of similarity among the stored preliminary reference data can be grouped and the reference data can be generated among the grouped preliminary reference data. The generated reference data can be used to determine whether to perform an error handling operation by calculating the similarity between the sensor information of the object Internet 1 including the monitoring target device and the reference data. If the control unit 104 determines that the error risk of the monitoring target device is detected, the control unit 104 may send a message to the user's mobile device or the communication device that the risk has been detected.

3 is a diagram showing an example of the similarity between reference data and sensor data according to an embodiment of the present invention.

The reference data storage unit 102 stores, as reference data, time series sensor information immediately before occurrence of an error of the device corresponding to the category to which the monitoring target device belongs.

The monitoring target device can be classified into categories according to the features or types of the device, such as the type of device, the product number of the device, and the type of sensor included in the device. The monitoring target device can be classified into a plurality of categories.

The sensor records information about the sensor information at regular intervals, and when an error or failure occurs in the apparatus, it is possible to extract information on a certain time series sensor immediately before an error or a failure occurs. The extracted sensor information may be stored in the reference data storage unit 102 as reference data.

The similarity calculation unit 103 calculates the similarity between the sensor data and the reference data.

The sensor data is data in which sensor information related to a monitoring target device is recorded, and may be data in which sensor information is continuously recorded at regular intervals to continuously monitor the device in case the user generates an error or a failure.

The reference data is the data to be compared with the sensor data, and is data in which the data in the case where an error or a failure has already occurred is recorded and the sensor information is changed in a time-series manner immediately before an error or a failure occurs. For example, when the category of the monitoring target device A is a washing machine and the product number is 'AB-100', the reference data storage unit 102 stores a malfunction of a device whose category is a washing machine and whose product number is 'AB-100' Will be stored for the predetermined time series sensor information just before the occurrence, which will be the reference data. Most equipment failures will not occur suddenly and will show some pattern in the information gathered by the sensor until failure occurs. Therefore, it is possible to see the sensor information immediately before the fault occurrence in a time series for comparison with the previous symptom and the monitoring target device.

If there is a possibility that a fault will occur in a monitoring target device, the sensor of the device will measure it and the sensor data recorded in a time series can be compared with the reference data to measure the similarity. If the monitoring target device shows a pattern similar to the pattern shown in the time series sensor information just before the error of the reference data does not occur but the failure of the monitoring target device is likely to occur, the sensor data and the reference data can be compared and calculated.

The first degree-of-similarity calculation unit 105 calculates the difference between the specific data in the time-series data and the data before two points of the specific data for each of the reference data and the sensor data, and calculates the first degree of similarity using the degree of similarity between the calculated values .

The specific data in the time series data for each of the reference data and the sensor data may be data at one point in the time series data and may be data corresponding to one interval of the data in which the sensor records information of the apparatus in a time interval . For example, if the reference data is 15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, , 5, 20, etc. may be specific data. Also, in general, in general cases where no errors or failures occur, specific data is often maintained in a certain pattern or a certain level of data value. In general, when a fault or failure occurs, Other data will be derived.

The first degree-of-similarity calculation unit 105 determines whether the difference between the specific data in the time-series data and the data before two points in time of the specific data is +, 0, or - for the reference data and the sensor data, The ratio of the case where the sign of the difference value between the reference data and the sensor data is the same is calculated by the first similarity degree.

The first degree-of-similarity calculation unit 105 calculates the difference between the specific data in the time-series data and the data before two points of the specific data with respect to the reference data and the sensor data, respectively. For example, the reference data is {15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, 15, 15, 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, when two points in front of each specific data are calculated, Except for the previous two points, the reference data is {0, 0, 0, 10, -15, 5, 5, -10, 5, -5, 5, 0, 0, 0} The sensor data is {0, 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0, 0} except for the previous two points of data. However, in the general case in which no error or failure occurs, it can be estimated that both the reference data and the sensor data are the portion of the specific data 15, and both the reference data and the sensor data have errors or failures The part that is assumed not to be done may be omitted. Therefore, the difference between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {10, -15, 5, 5, -10, 5, -5, 5} The difference between the data and the data before two points of the specific data is {7, -12, 13, 0, -15, 4, -5, 8}. 5, 5, 5, 5, 5, 5, 0, 0, 0} and {10, -15, 5, 5, -10 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0 , 0} and {7, -12, 13, 0, -15, 4, -5, 8} can be treated as the same value to exclude specific data judged not to be related to error or failure.

If the difference between the specific data of the time series data of the reference data and the sensor data and the data before the two points of time of the specific data increases, it can be displayed as + or 1, That is, whether or not increase, decrease, or decrease can be indicated by a sign.

The sign of the difference value between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {1, -1, 1, 1, -1, 1, -1, 1} The sign of the difference between the specific data of the data and the data before the two points of time of the specific data is {1, -1, 1, 0, -1, 1, -1, 1}.

The first similarity degree is calculated at a ratio of the case where the sign of the difference value between the reference data and the sensor data is the same, and the code similarity degree of the reference data and the sensor data is {identical, same, same, same, same, same, }, Which is 7 out of the 8 arrays, the first similarity can be calculated as (7/8) * 100 = 87.5%.

The second similarity calculation unit 106 calculates the difference between the specific data of the time series data and the data of the one point before the specific data with respect to each of the reference data and the sensor data and calculates the second similarity by using the similarity between the calculated values .

For example, the reference data is {15,15,15,15,15,25,0,30,50,15,15,15,15,15} and the sensor data is {15,15,15,15,15,15,15} 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, and calculates the difference between the specific data of the time series data and the data of the specific date 0, 0, 0, 10, -25, 30, -25, 15, -10, 5, 0, 0, 0, 0} and the sensor data is {0, 0, 0 , 0, 7, -19, 32, -32, 17, -13, 8, 0, 0, 0, 0}. The reference data is {10, -25, 30, -25, 15, -10, 5} and the sensor data is {7, -19 , 32, -32, 17, -13, 8}.

The second similarity calculation unit 106 calculates the following equation using the difference value between the specific data of the time series data and the data of one time before the specific data for each of the reference data and the sensor data,

The average of the calculation results of the above equations is calculated as the second similarity.

To calculate the second degree of similarity, the result using the formula is {70, 76, 93.3, 72, 86.6, 70, 40}, and the average can be calculated as 72.56, and the second similarity is 72.56%.

FIG. 4 is a view showing an example of mutual similarity in a combination of two pairs of preliminary reference data according to an embodiment of the present invention.

The reference data generation unit 109 groups the plurality of preliminary reference data having a high degree of similarity among the plurality of preliminary reference data stored in the preliminary reference data storage units 108 and 102, And stores the generated data in the reference data storage unit 102.

The reference data generation unit 109 generates a combination of two pairs of the preliminary reference data, extracts the data of each combination, calculates the similarity by the similarity calculation unit 103, If they are more than the reference value, they are grouped into the same group.

There are various patterns of preliminary data in which sensor information immediately before occurrence of error of various devices is stored, and similar patterns among various patterns can be grouped. Grouped preliminary reference data The preliminary reference data having the highest sum of similarities can be generated as reference data by comparing the similarity with other preliminary reference data. The similarity calculated through the similarity calculation unit 103 can be grouped into the same group if the similarity is equal to or greater than a predetermined reference value. For example, if the predetermined reference value is 70, all the reference data in the same group are mated If we compare the similarities, we can say that the similarity can not be more than 70.

For example, in group A, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped. The preliminary reference data 1 has a similarity degree of 90 to the preliminary reference data 2, a similarity to the preliminary reference data 3 is 80, a similarity to the preliminary reference data 4 is 70, a total of 240, and the preliminary reference data 2 has 90 , The degree of similarity to the preliminary reference data 3 is 70, the degree of similarity to the preliminary reference data 4 is 75, and the total is 235. The preliminary reference data 3 has a similarity to the preliminary reference data 1 of 80, a similarity to the preliminary reference data 2 of 70, 4 and the similarity degree is 70, the sum of the preliminary reference data 4 is 70, the degree of similarity with the preliminary reference data 2 is 75, the similarity with the preliminary reference data 3 is 70, and the sum is 215, 1 is the highest with respect to the other preliminary reference data, the reference data of the group A is the preliminary reference data 1, and the preliminary reference data 1 is stored in the reference data storage unit 10 2). ≪ / RTI >

However, when a plurality of preliminary reference data are grouped and it is judged that other preliminary reference data can be entered into a group already grouped, the preliminary reference data is compared with the reference data generated as reference data, The preliminary reference data can be grouped.

For example, if a predetermined value capable of entering an already-grouped group is 90, if it is judged that the other spare baseline data 10 can be grouped into the group A while being grouped into the group A, The preliminary reference data 1 and the preliminary reference data 10 can be grouped into the group A if the similarity degree is 90 or more. This is advantageous in that even when new preliminary reference data is collected, it is not compared with all the preliminary reference data in the group, so that it can be processed faster and save time and money.

In addition, a plurality of preliminary reference data can be integrated to generate virtual reference data having a high degree of similarity to the grouped preliminary reference data. For example, in Group B, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped, and all four preliminary reference data and virtual reference data Lt; / RTI >

FIG. 5 is a flowchart illustrating a flow of a method for managing Internet-based devices according to an embodiment of the present invention.

Step S501 receives the sensor information related to the monitoring target device.

The monitoring target device may be any electronic device such as a vehicle, a traffic light, a refrigerator, a washing machine, etc., which is commonly found in the vicinity, and may be connected to the Internet. A monitoring target device is a device that the user wants to detect a failure or an error, and is a target to prevent damage due to failure of the device by detecting a failure in advance. The sensor included in the monitoring target device can detect all the conditions of the device. The sensor information can be all information that can be measured by the sensor such as current information, temperature information, humidity information, leak information, It is possible to record about the sensor information at intervals.

In step S502, the time series sensor information immediately before occurrence of the error of the device corresponding to the category to which the monitoring target device belongs is stored as reference data.

The monitoring target device can be classified into categories according to the features or types of the device, such as the type of device, the product number of the device, and the type of sensor included in the device. The monitoring target device can be classified into a plurality of categories.

The sensor records information about the sensor information at regular intervals, and when an error or failure occurs in the apparatus, it is possible to extract information on a certain time series sensor immediately before an error or a failure occurs. The extracted sensor information may be stored in step S502 as reference data.

Step S503 calculates the similarity between the sensor data and the reference data.

The sensor data is data in which sensor information related to a monitoring target device is recorded, and may be data in which sensor information is continuously recorded at regular intervals to continuously monitor the device in case the user generates an error or a failure.

The reference data is the data to be compared with the sensor data, and is data in which the data in the case where an error or a failure has already occurred is recorded and the sensor information is changed in a time-series manner immediately before an error or a failure occurs. For example, if the category of the monitoring target device A is a washing machine and the product number is' AB-100 ', then step S502 is a time series immediately before the occurrence of a failure of a device whose category is a washing machine and whose product number is' AB- Sensor information, which will be the reference data. Most equipment failures will not occur suddenly and will show some pattern in the information gathered by the sensor until failure occurs. Therefore, it is possible to see the sensor information immediately before the fault occurrence in a time series for comparison with the previous symptom and the monitoring target device.

If there is a possibility that a fault will occur in a monitoring target device, the sensor of the device will measure it and the sensor data recorded in a time series can be compared with the reference data to measure the similarity. If the monitoring target device shows a pattern similar to the pattern shown in the time series sensor information just before the error of the reference data does not occur but the failure of the monitoring target device is likely to occur, the sensor data and the reference data can be compared and calculated.

The first degree of similarity calculation step calculates the difference between the specific data of the time series data and the data of two points before the specific data for each of the reference data and the sensor data and calculates the first similarity degree by using the similarity between the calculated values .

The specific data in the time series data for each of the reference data and the sensor data may be data at one point in the time series data and may be data corresponding to one interval of the data in which the sensor records information of the apparatus in a time interval . For example, if the reference data is 15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, , 5, 20, etc. may be specific data. Also, in general, in general cases where no errors or failures occur, specific data is often maintained in a certain pattern or a certain level of data value. In general, when a fault or failure occurs, Other data will be derived.

The first similarity calculation step determines whether the difference between the specific data in the time series data and the data before two points of time of the specific data is +, 0, or - for the reference data and the sensor data, And the ratio of the case where the sign of the difference value of each of the sensor data is the same is calculated as the first similarity degree.

The first degree of similarity calculating step calculates the difference between the specific data of the time series data and the data of two points before the specific data with respect to each of the reference data and the sensor data. For example, the reference data is {15, 15, 15, 15, 15, 25, 0, 30, 5, 20, 10, 15, 15, 15, 15, 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, when two points in front of each specific data are calculated, Except for the previous two points, the reference data is {0, 0, 0, 10, -15, 5, 5, -10, 5, -5, 5, 0, 0, 0} The sensor data is {0, 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0, 0} except for the previous two points of data. However, in the general case in which no error or failure occurs, it can be estimated that both the reference data and the sensor data are the portion of the specific data 15, and both the reference data and the sensor data have errors or failures The part that is assumed not to be done may be omitted. Therefore, the difference between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {10, -15, 5, 5, -10, 5, -5, 5} The difference between the data and the data before two points of the specific data is {7, -12, 13, 0, -15, 4, -5, 8}. 5, 5, 5, 5, 5, 5, 0, 0, 0} and {10, -15, 5, 5, -10 0, 0, 7, -12, 13, 0, -15, 4, -5, 8, 0, 0 , 0} and {7, -12, 13, 0, -15, 4, -5, 8} can be treated as the same value to exclude specific data judged not to be related to error or failure.

If the difference between the specific data of the time series data of the reference data and the sensor data and the data before the two points of time of the specific data increases, it can be displayed as + or 1, That is, whether or not increase, decrease, or decrease can be indicated by a sign.

The sign of the difference value between the specific data in the time series data of the reference data and the data before the two points in time of the specific data is {1, -1, 1, 1, -1, 1, -1, 1} The sign of the difference between the specific data of the data and the data before the two points of time of the specific data is {1, -1, 1, 0, -1, 1, -1, 1}.

The first similarity degree is calculated at a ratio of the case where the sign of the difference value between the reference data and the sensor data is the same, and the code similarity degree of the reference data and the sensor data is {identical, same, same, same, same, same, }, Which is 7 out of the 8 arrays, the first similarity can be calculated as (7/8) * 100 = 87.5%.

The second degree of similarity calculation step calculates the difference between the specific data in the time series data and the data of one point before the specific data for each of the reference data and the sensor data and calculates the second similarity degree by using the similarity between the calculated values .

For example, the reference data is {15,15,15,15,15,25,0,30,50,15,15,15,15,15} and the sensor data is {15,15,15,15,15,15,15} 15, 15, 15, 22, 3, 35, 3, 20, 7, 15, 15, 15, 15 and 15}, and calculates the difference between the specific data of the time series data and the data of the specific date 0, 0, 0, 10, -25, 30, -25, 15, -10, 5, 0, 0, 0, 0} and the sensor data is {0, 0, 0 , 0, 7, -19, 32, -32, 17, -13, 8, 0, 0, 0, 0}. The reference data is {10, -25, 30, -25, 15, -10, 5} and the sensor data is {7, -19 , 32, -32, 17, -13, 8}.

Calculating the second similarity using the difference value between the specific data of the time series data and the data of one time before the specific data with respect to each of the reference data and the sensor data,

The average of the calculation results of the above equations is calculated as the second similarity.

To calculate the second degree of similarity, the result using the formula is {70, 76, 93.3, 72, 86.6, 70, 40}, and the average can be calculated as 72.56, and the second similarity is 72.56%.

In step S504, if an error risk of the monitoring target device is detected based on the degree of similarity calculated in step S503, control is performed to perform an error countermeasure operation.

The error countermeasure operation may be an action to prevent an error or a failure from occurring before a failure or malfunction of the device occurs, such as an operation of turning off the power of the monitoring target device or an operation of checking the system, If a danger is detected, control can be performed to perform an error countermeasure operation such as an operation of turning off the monitoring target device or performing a system check.

In step S504, it is determined whether or not an error countermeasure operation is performed based on the general similarity combining the first similarity degree and the second similarity degree.

In step S503, the overall similarity degree may be calculated by combining the first similarity degree and the second similarity degree. In step S504, it may be determined that the error risk of the monitoring target device is detected when the calculated degree of similarity becomes equal to or higher than a predetermined reference level. The overall similarity can be calculated by various methods, and in step S504 it is possible to detect the error risk of the monitoring target device.

The general similarity degree may be calculated by comparing the first similarity degree and the second similarity degree with each other. For example, it is assumed that the first degree of similarity is 87.5%, the second degree of similarity is 72.56%, and the reference value is 70%. The first similarity degree and the second similarity degree are both determined to be equal to or greater than the reference value of 70%, and it is determined that the error risk of the monitoring target device is detected in step S504. However, assuming that the reference value is 80%, since the first similarity degree is not less than the reference value, and the second similarity degree is less than the reference value, neither the first similarity degree nor the second similarity degree is equal to or greater than the reference value, and thus step S504 determines that the error risk of the monitoring target device is not detected .

In addition, the general similarity degree is calculated as an average of the first similarity degree and the second similarity degree, and if the average value is equal to or greater than the reference value, the step S504 can determine that the monitoring target device detects an error risk. Assuming, for example, that the first degree of similarity is 87.5% and the second degree of similarity is 72.56%, it is assumed that the reference value for determining that the monitoring target device detects an error risk is 80%. The average of the first similarity and the second similarity is 80.03%, and the reference value is 80%. In step S504, it can be determined that the error risk of the monitoring target device is detected.

In addition, the overall similarity degree may be determined when at least one of the first similarity degree and the second similarity degree is equal to or greater than the reference value, and in step S504, the monitoring target device determines that an error risk is detected. For example, it is assumed that the first degree of similarity is 87.5%, the second degree of similarity is 72.56%, and the reference value is 80%. The second degree of similarity is not equal to or greater than the reference value, but the first degree of similarity is determined to be equal to or greater than the reference value, and step S504 can determine that the error risk of the monitoring target device is detected. However, assuming that the reference value is 90%, since the reference value does not reach both the first similarity degree and the second similarity degree, step S504 can determine that the error risk of the monitoring target device is not detected.

In addition, the general similarity degree can be calculated by weighting each of the first similarity degree and the second similarity degree. The overall similarity degree can be calculated by assigning a high weight to the degree of similarity determined to be higher among the first similarity degree and the second similarity degree. In step S504, it can be determined that the error risk of the monitoring target device is detected when the overall similarity degree is equal to or greater than the reference value. For example, assuming that the first degree of similarity is 87.5% and the second degree of similarity is 72.56%, the weights of the first degree of similarity and the second degree of similarity are respectively 70%: 30% Is assumed to be 80% or more. (87.5X0.7) + (72.56X0.3) = 83.02, which is a reference value of 80% or more, in step S504, it can be determined that the error risk of the monitoring target device is detected.

The collecting step collects sensor information related to a plurality of devices.

The collecting step collects sensor information related to various electronic devices such as vehicles, traffic lights, refrigerators, washing machines, The sensor information may be information on all the sensors included in the apparatus and may be information recorded in a time interval at a predetermined interval or at a specific interval.

The preliminary step S502 stores the sensor information of the predetermined number of points of the sensor information collected in the collecting step as the preliminary reference data immediately before the error of the device occurs.

Sensor information of the device when an error occurs in the sensor information may be recorded and sensor information of a certain point of time immediately before the error of the device may be stored as reference data. The reference data is data to be compared in case of occurrence of errors or failures of other devices. As more preliminary reference data exist, more reference data can be extracted because there are more objects to be compared, so that errors or failures can be predicted more quickly have.

The reference data generating step groups the plurality of preliminary reference data having a high degree of similarity among the plurality of preliminary reference data stored in the preliminary step S502, generates reference data relating to a plurality of grouped preliminary reference data, and stores the generated reference data in step S502.

In the step of generating reference data, a combination of two pairs of preliminary reference data is generated, data of each combination is extracted, and the similarity degree is calculated in step S503. If the calculated degree of similarity is equal to or greater than a predetermined reference value, Grouping.

There are various patterns of preliminary data in which sensor information immediately before occurrence of error of various devices is stored, and similar patterns among various patterns can be grouped. Grouped preliminary reference data The preliminary reference data having the highest sum of similarities can be generated as reference data by comparing the similarity with other preliminary reference data. If the similarity calculated through step S503 is equal to or greater than a predetermined reference value, which is a constant value, the data can be grouped into the same group. For example, if the predetermined reference value is 70, all the redundant reference data in the same group are mated If the similarity degree is 70 or more, the group can belong to it.

For example, in group A, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped. The preliminary reference data 1 has a similarity degree of 90 to the preliminary reference data 2, a similarity to the preliminary reference data 3 is 80, a similarity to the preliminary reference data 4 is 70, a total of 240, and the preliminary reference data 2 has 90 , The degree of similarity to the preliminary reference data 3 is 70, the degree of similarity to the preliminary reference data 4 is 75, and the total is 235. The preliminary reference data 3 has a similarity to the preliminary reference data 1 of 80, a similarity to the preliminary reference data 2 of 70, 4 and the similarity degree is 70, the sum of the preliminary reference data 4 is 70, the degree of similarity with the preliminary reference data 2 is 75, the similarity with the preliminary reference data 3 is 70, and the sum is 215, 1 is the most similar to the other preliminary reference data, the reference data of the group A becomes the preliminary reference data 1 and the preliminary reference data 1 can be stored in step S502 have.

However, when a plurality of preliminary reference data are grouped and it is judged that other preliminary reference data can be entered into a group already grouped, the preliminary reference data is compared with the reference data generated as reference data, The preliminary reference data can be grouped.

For example, if a predetermined value capable of entering an already-grouped group is 90, if it is judged that the other spare baseline data 10 can be grouped into the group A while being grouped into the group A, The preliminary reference data 1 and the preliminary reference data 10 can be grouped into the group A if the similarity degree is 90 or more.

In addition, a plurality of preliminary reference data can be integrated to generate virtual reference data having a high degree of similarity to the grouped preliminary reference data. For example, in Group B, preliminary reference data 1, preliminary reference data 2, preliminary reference data 3, and preliminary reference data 4 having four preliminary reference data are grouped, and all four preliminary reference data and virtual reference data Lt; / RTI >

The object Internet-based device management method according to an embodiment of the present invention may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

While the invention has been shown and described with reference to certain preferred embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. This is possible.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined by the equivalents of the claims, as well as the claims.

101: sensor data reception unit 102: reference data storage unit
103: similarity calculation unit 104:
105: first similarity degree calculation unit 106: second similarity degree calculation unit
107: Collecting unit 108: Preliminary reference data storage unit
109: Reference data generation unit

Claims (13)

A sensor data receiving unit for receiving sensor information related to a monitoring target device;
A reference data storage unit for storing, as reference data, time series sensor information immediately before occurrence of an error of a device corresponding to the category to which the monitoring target device belongs;
A similarity calculation unit for calculating the similarity between the sensor data and the reference data;
A controller for controlling the error counter operation to be performed when an error risk of the monitoring target device is detected based on the similarity calculated by the similarity calculator;
Lt; / RTI >
The similarity calculation unit
Calculating a difference between specific data in the time series data and data before two points in time of the specific data with respect to each of the reference data and the sensor data and calculating a first similarity degree using the similarity between the calculated values, Calculating section; And
Calculating a difference between the specific data in the time series data and the data before the one point in time of the specific data for each of the reference data and the sensor data and calculating the second similarity using the similarity between the calculated values, Calculating section;
Lt; / RTI >
The control unit
Based on the combined degree of similarity obtained by combining the first degree of similarity and the second degree of similarity, whether an error corresponding operation is performed or not. delete The method according to claim 1,
The first degree-of-similarity calculation unit
Determining whether the difference between the specific data in the time series data and the data before the two points of time of the specific data is +, 0, or - for each of the reference data and the sensor data, And calculating the ratio of the case where the sign of the difference value of the first similarity degree is the same
Based object management system. The method according to claim 1,
The second degree of similarity calculation unit
Calculating the following equation using the difference between the specific data of the time series data and the data of the specific data at one point in time for each of the reference data and the sensor data,

Calculating the average of the calculation results of the mathematical expression by the second similarity
Based object management system. The method according to claim 1,
A collection unit for collecting sensor information related to a plurality of devices;
A preliminary reference data storage unit for storing, as preliminary reference data, sensor information of a predetermined number of points of the sensor information collected by the collecting unit immediately before occurrence of an error of the apparatus;
The plurality of preliminary reference data having a high degree of similarity among the plurality of preliminary reference data stored in the preliminary reference data storage unit, grouping the plurality of preliminary reference data having a high degree of mutual similarity, generating reference data relating to the grouped plurality of preliminary reference data, A reference data generating unit;
Further comprising: an Internet-based device management system for managing the Internet. 6. The method of claim 5,
The reference data generation unit
A plurality of pairs of preliminary reference data are generated by combining two pairs of data, extracting data of each combination, calculating the similarity in the similarity calculation unit, and grouping the same into the same group if the calculated similarity is equal to or greater than a predetermined reference value
Features Internet-based device management system with features. A sensor data receiving step of receiving sensor information related to a monitoring target device;
A reference data storage step of storing, as reference data, time series sensor information immediately before occurrence of an error of a device corresponding to the category to which the monitoring target device belongs;
A similarity calculation step of calculating the similarity between the sensor data and the reference data;
When the error risk of the monitoring target device is detected based on the similarity calculated in the similarity calculation step,
Lt; / RTI >
The similarity calculation step
Calculating a difference between specific data in the time series data and data before two points in time of the specific data with respect to each of the reference data and the sensor data and calculating a first similarity degree using the similarity between the calculated values, Calculating step; And
Calculating a difference between the specific data in the time series data and the data before the one point in time of the specific data for each of the reference data and the sensor data and calculating the second similarity using the similarity between the calculated values, Calculating step;
Lt; / RTI >
The control step
Determining whether or not the error countermeasure operation is performed based on the total similarity obtained by combining the first similarity degree and the second similarity degree
Wherein the method comprises the steps of: delete 8. The method of claim 7,
The first degree of similarity calculating step
Determining whether the difference between the specific data in the time series data and the data before the two points of time of the specific data is +, 0, or - for each of the reference data and the sensor data, And calculating the ratio of the case where the sign of the difference value of the first similarity degree is the same
Wherein the method comprises the steps of: 8. The method of claim 7,
The second degree of similarity calculating step
Calculating the following equation using the difference between the specific data of the time series data and the data of the specific data at one point in time for each of the reference data and the sensor data,

Calculating the average of the calculation results of the mathematical expression by the second similarity
Wherein the method comprises the steps of: 8. The method of claim 7,
A collection step of collecting sensor information related to a plurality of devices;
A preliminary reference data storing step of storing, as preliminary reference data, sensor information of a predetermined number of points of the sensor information collected in the collecting step immediately before an error of the apparatus occurs;
Grouping the plurality of preliminary reference data having a high degree of similarity among the plurality of preliminary reference data stored in the preliminary reference data storing step and generating reference data related to the grouped plurality of preliminary reference data and storing the reference data A reference data generation step of storing the reference data;
Further comprising the steps of: (a) receiving a request for the Internet service from the Internet service provider; 12. The method of claim 11,
The reference data generation step
The similarity degree is calculated in the step of calculating the degree of similarity by extracting the data of each combination, and if the calculated degree of similarity is equal to or larger than the predetermined reference value, grouping into the same group that
A feature to manage Internet based device management methods. A recording medium on which a program for executing the method of any one of claims 7 and 9 to 12 is recorded.

Images