JP7690752B2 - Peripheral device, monitoring system, monitoring method and program - Google Patents

Description

The present invention relates to peripheral devices, monitoring systems, monitoring methods and programs.

Technology is being developed to detect anomalies in computers for industrial use, etc.

For example, Patent Document 1 describes a technology that generates a prediction model for multiple similar devices and then generates a prediction model dedicated to a newly installed device based on the generated prediction models.

JP 2010-287011 A

The objective of this disclosure is to provide peripheral devices, monitoring systems, monitoring methods, and programs that are more suitable for monitoring the devices to be monitored.

The peripheral device according to one aspect of this embodiment includes a power receiving unit that receives power from a device to be monitored, an environmental sensor that acquires environmental information about the surroundings of the device, a transmitting unit that transmits the environmental information, and a rechargeable battery that is charged by receiving power from the device via the power receiving unit, and the environmental sensor and the transmitting unit operate using discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit.

The monitoring system according to one aspect of the present embodiment includes a peripheral device provided in correspondence with a device to be monitored, and a monitoring server that communicates with the peripheral device. The peripheral device includes a power receiving unit that receives power from the device, an environmental sensor that acquires environmental information about the surroundings of the device, a transmitting unit that transmits the environmental information, and a rechargeable battery that is charged by receiving power from the device via the power receiving unit. When it becomes impossible to receive power via the power receiving unit, the environmental sensor and the transmitting unit operate using discharged power from the rechargeable battery. The monitoring server includes a receiving unit that receives the environmental information from the peripheral device, an operation information acquiring unit that acquires the operation status of the device as operation information, and a status determining unit that uses the operation information and the environmental information to determine the status of the device.

In one aspect of the monitoring method according to this embodiment, a peripheral device executes a charging step of charging a rechargeable battery by receiving power from the device to be monitored, an environmental information acquisition step of acquiring environmental information about the surroundings of the device using discharged power from the rechargeable battery when power cannot be received from the device, and a transmission step of transmitting the environmental information using discharged power from the rechargeable battery.

The program according to one aspect of this embodiment causes a peripheral device to execute, as a monitoring method, a charging step of charging a rechargeable battery by receiving power from a device to be monitored, an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when power cannot be received from the device, and a transmission step of transmitting the environmental information using discharged power from the rechargeable battery.

The present disclosure provides peripheral devices, monitoring systems, monitoring methods, and programs that are more suitable for monitoring the devices to be monitored.

FIG. 2 is a block diagram of a peripheral device according to the first embodiment. 5 is a flowchart illustrating an example of a process executed by the peripheral device according to the first embodiment. 1 is a block diagram of a monitoring system according to a first embodiment. FIG. 4 is a sequence diagram showing an example of a process executed by the monitoring system according to the first embodiment; 1 is a block diagram of a prediction device according to a first embodiment; FIG. 2 is a block diagram of a CNN model unit according to the first embodiment. 1 is a table illustrating an example of teacher data learned by a CNN model unit according to the first embodiment; 1 is a table illustrating an example of teacher data learned by a CNN model unit according to the first embodiment; 1 is a table illustrating an example of data predicted by a CNN model unit according to the first embodiment; 4 is a flowchart illustrating an example of a process executed by the prediction device according to the first embodiment; 1 is a block diagram of a determination device according to a first embodiment; 4 is a flowchart illustrating an example of a process executed by the determination device according to the first embodiment. 1 is a block diagram of a determination device according to a first embodiment; 4 is a flowchart illustrating an example of a process executed by the determination device according to the first embodiment. FIG. 11 is a block diagram of a monitoring system according to a second embodiment. FIG. 11 is a block diagram of a USB sensor according to a second embodiment. FIG. 11 is a block diagram of the monitoring system showing a case where some of the factory computers have abnormally terminated in the second embodiment. FIG. 11 is a block diagram of a USB sensor showing a case where a connected factory computer has abnormally terminated in the second embodiment. FIG. 11 is a block diagram of a monitoring server according to a second embodiment. FIG. 11 is a block diagram of a USB sensor according to a second embodiment. FIG. 11 is a block diagram of a monitoring server according to a second embodiment. FIG. 11 is a diagram showing an example of sound data stored in a database according to the second embodiment; 13 is an example of a graph plotting similarities according to the second embodiment. 13 is a flowchart illustrating an example of a process executed by a monitoring server according to the second embodiment; 13 is a flowchart illustrating an example of a process executed by a monitoring server according to the second embodiment; FIG. 2 is a block diagram showing an example of a hardware configuration of an apparatus according to each embodiment.

First embodiment
(1A)
Hereinafter, a first embodiment of the present disclosure will be described with reference to the drawings. In this embodiment (1A), a peripheral device provided in correspondence with a device to be monitored, such as a computer, will be described.

Figure 1 is a block diagram showing an example of a peripheral device. The peripheral device 10 includes a power receiving unit 11, an environmental sensor 12, a transmitting unit 13, and a rechargeable battery 14. Each part of the peripheral device 10 is controlled by a control unit (controller) (not shown). Each component will be described below.

The power receiving unit 11 is a part that receives power from a device to be monitored (hereinafter, referred to as the target device). The power receiving unit 11 is, for example, a connector that is physically connected to the peripheral device 10, and the peripheral device 10 can be attached and detached from the target device by inserting and removing the power receiving unit 11 from the connector of one target device (not shown). Examples of connectors include terminals such as a USB port and a LAN (Local Area Network) port. When the connector is a USB port, the peripheral device is an external device connected to the target device by a peripheral bus for connecting peripheral devices, such as a USB (Universal Serial Bus) peripheral. In addition, the rechargeable battery 14 described below can be charged by the power supply from the target device via the power receiving unit 11. For example, when the power receiving unit 11 is a LAN port, the target device supplies power to the peripheral device 10 by PoE (Power over Ethernet).

However, the power receiving unit 11 may receive power wirelessly from the target device without physically connecting the peripheral device 10 to the target device. In this case, the power receiving unit 11 may receive power by any method, such as an electromagnetic induction method, a magnetic field resonance method, a radio wave reception method, or an electric field coupling method.

The environmental sensor 12 is a sensor that acquires environmental information around the target device. Examples of environmental information include, but are not limited to, at least one detected value of temperature, humidity, acceleration, vibration (e.g., impact), sound (e.g., sound pressure), air pressure, wind, illuminance, ultraviolet light, and the content of certain substances in the air, such as VOCs (Volatile Organic Compounds). These factors may affect the operation of the target device in at least either the short or long term.

The transmission unit 13 is a transmitter that transmits the environmental information acquired by the environmental sensor 12. The destination of the environmental information may be, for example, a monitoring server that uses the environmental information to determine the state (e.g., the operating state) of the target device. In this example, the transmission unit 13 transmits the environmental information wirelessly R1, but wired transmission may also be used. The timing at which the transmission unit 13 transmits the environmental information may be periodic, or transmission may be performed when the detected environmental information has a characteristic. The environmental sensor 12 and the transmission unit 13 can usually be powered via the power receiving unit 11.

The rechargeable battery 14 is charged by receiving power from the target device via the power receiving unit 11. The rechargeable battery 14 can supply power to at least the environmental sensor 12 and the transmission unit 13. In particular, the environmental sensor 12 and the transmission unit 13 can operate using discharged power from the rechargeable battery 14 when they are no longer able to receive power via the power receiving unit 11. An example of a case in which the environmental sensor 12 and the transmission unit 13 are no longer able to receive power via the power receiving unit 11 is when the target device performs an abnormal operation, such as an abnormal termination.

Figure 2 is a flowchart showing an example of a typical process of the peripheral device 10, and the process of the peripheral device 10 is explained using this flowchart. First, the peripheral device 10 charges the rechargeable battery 14 by receiving power via the power receiving unit 11 with the target device (step S11).

When the target device is no longer able to receive power, the environmental sensor 12 acquires environmental information about the surroundings of the target device using the discharged power from the rechargeable battery 14 (step S12). The transmitter 13 transmits the environmental information using the discharged power from the rechargeable battery 14 (step S13). The peripheral device 10 can operate as described above.

In this way, the peripheral device 10 can acquire and transmit environmental information using the rechargeable battery 14, even when power supply is stopped due to an abnormal termination of the target device (e.g., a system crash). In other words, the peripheral device 10 can accurately acquire information before and after the abnormal termination of the target device, i.e., information for predicting an abnormality in the target device. In addition, the peripheral device 10 can receive power from the target device via the power receiving unit 11. Therefore, even in computers and the like that do not have a status monitoring function, there is no need to incorporate dedicated hardware or device drivers in advance, and this function can be added by retrofitting the peripheral device 10.

In this example, one peripheral device 10 corresponds to one target device, but multiple target devices may be provided and one peripheral device 10 may correspond to the multiple target devices. In this case, the power receiving unit 11 receives power from at least one of the multiple target devices. The environmental sensor 12 acquires environmental information around the multiple target devices, and the transmitting unit 13 transmits the environmental information related to the multiple target devices.

The peripheral device 10 may further include a memory unit (storage) that stores environmental information. When the peripheral device 10 is receiving power through the power receiving unit 11, the transmission unit 13 may transmit the environmental information stored in the memory unit at regular intervals, while when the peripheral device 10 is unable to receive power through the power receiving unit 11, the transmission unit 13 may transmit the environmental information stored in the memory unit regardless of the regular intervals. As another example, the peripheral device 10 may further include a determination unit that determines whether or not the detection value of the environmental information detected by the environmental sensor 12 is equal to or greater than a predetermined threshold. When the determination unit determines that the detection value is less than the predetermined threshold, the transmission unit 13 may transmit the environmental information stored in the memory unit at regular intervals, while when the determination unit determines that the detection value is equal to or greater than the predetermined threshold, the transmission unit 13 may transmit the environmental information stored in the memory unit regardless of the regular intervals.

The above-mentioned "transmitting regardless of the periodic timing" means, for example, that when the peripheral device 10 is no longer able to receive power or when the detection value is equal to or greater than a predetermined threshold, the environmental information is transmitted even if the normal periodic timing for transmission has not arrived. Alternatively, the transmission timing of the transmitting unit may be shortened, i.e., the frequency with which the environmental information is transmitted may be increased. This makes it possible to reliably obtain information for predicting an abnormality in the target device when an abnormality occurs in the target device or when there is a sign of an abnormality.

When the transmission unit 13 transmits the environmental information to the monitoring server and the peripheral device 10 subsequently receives an instruction from the monitoring server to increase the frequency with which the environmental information is transmitted, the transmission unit 13 may be controlled to increase the frequency with which the environmental information is transmitted. This allows the peripheral device 10 to reliably obtain information for predicting anomalies in the target device in response to an instruction from the monitoring server.

(1B)
Next, in (1B), a monitoring system including a peripheral device 10 and a monitoring server will be described.

Figure 3 is a block diagram showing an example of a monitoring system. The monitoring system S1 comprises a peripheral device 10 and a monitoring server 20. The peripheral device 10 has been described in (1A) and will not be described here. The monitoring server comprises a receiving unit (receiver) 21, an operation information acquisition unit 22, and a status determination unit 23. Each of these components will be described below.

The receiving unit 21 receives environmental information from one or more peripheral devices 10 to be monitored. In other words, the monitoring server 20 is the destination to which the transmitting unit 13 of the peripheral device 10 transmits the environmental information, and in this example, is wirelessly connected to the peripheral device 10. However, the monitoring server 20 may also acquire environmental information by performing wired communication with the peripheral device 10.

The operation information acquisition unit 22 acquires the operation status of one or more target devices for which the receiving unit 21 has received environmental information as operation information. The operation status is information that serves as an index showing whether the target device is operating normally, and examples of the operation status include at least one detection value such as the temperature of the target device, the voltage value within the device, and the fan rotation speed of the device, but is not limited to these.

The operation information acquisition unit 22 acquires operation information regarding one or more target devices wirelessly or via a wired connection. The operation information acquisition unit 22 may acquire this operation information, for example, from the target device, or may acquire this operation information from another device such as the peripheral device 10. When acquiring this operation information from the peripheral device 10, the peripheral device 10 is further provided with an operation information unit that acquires the operation information via the power receiving unit 11.

The state determination unit 23 determines the state of one or more target devices using the environmental information received by the receiving unit 21 and the operation information acquired by the operation information acquisition unit 22. For example, the state determination unit 23 may determine whether the target device is currently operating normally (whether an abnormality has occurred) using the operation information and the environmental information. In another example, the state determination unit 23 may predict whether there is a sign of an abnormality (i.e., whether an abnormality is likely to occur in operation in the future) using the operation information and the environmental information. The state determination unit 23 may make this prediction by generating a model using the environmental information and the operation information as teacher data, for example, by AI (Artificial Intelligence) such as machine learning. Details of this prediction will be described later. When determining the state of multiple target devices, the state determination unit 23 can determine the state of each target device individually.

Figure 4 is a sequence diagram showing an example of a typical process of the monitoring system S1, and the process of the monitoring system S1 is explained using this sequence diagram. First, when the environmental sensor 12 of the peripheral device 10 is unable to receive power, it acquires environmental information around the device using the discharged power from the rechargeable battery 14 (step S21). Then, the transmission unit 13 transmits the acquired environmental information to the monitoring server 20 using the discharged power from the rechargeable battery 14 (step S22). Note that before step S21, the rechargeable battery 14 is charged by receiving power from the power receiving unit 11. The details are as described in (1A).

The receiving unit 21 of the monitoring server receives the environmental information transmitted in step S22 (step S23). The operation information acquiring unit 22 acquires the operation status of the target device as operation information (step S24). Note that step S24 may be executed before or after step S23, or may be executed at the same time as step S23. The status determining unit 23 determines the status of the target device using the environmental information and operation information acquired in steps D23 and S24, respectively (step S25).

As shown in (1A), the peripheral device 10 can accurately obtain information before and after the abnormal termination of the target device, i.e., information for predicting an abnormality in the target device. Therefore, the monitoring server 20 can more accurately determine the state of the target device based on the information for predicting an abnormality in the target device.

(1C)
In the following, (1C) to (1E), an example of a device that performs prediction or judgment regarding the state of a target device will be described. This device may be applied as the monitoring server 20 in (1B), or may be applied as other devices.

Figure 5 is a block diagram showing an example of a prediction device. The prediction device 30 includes an acquisition unit 31, a prediction model generation unit 32, and a prediction unit 33. Each of these components will be described below.

The acquisition unit 31 acquires the operating status of one target device as operating information, and also acquires environmental information around the target device. The operating information and environmental information are as explained in (1A). The acquisition unit 31 may acquire the operating status from the target device, or may acquire it from another device (e.g., a peripheral device). The acquisition unit 31 may also acquire the environmental status from a peripheral device provided in correspondence with the target device. The operating information and environmental information acquired by the acquisition unit 31 may be stored in a memory unit within the prediction device 30, or may be stored in a memory unit such as a DB (database) outside the prediction device 30.

The prediction model generation unit 32 sets, as a set of variables, explanatory variables including at least environmental information in a specified time domain and objective variables including at least operation information at a timing immediately after the specified time domain for the data acquired by the acquisition unit 31. The prediction model generation unit 32 then generates a prediction model of the operation status using a set of multiple variables defined by sequentially shifting the time slots in the set of variables as teacher data. The prediction model generation unit 32 can generate this prediction model by extracting the operation information and environmental information stored in the memory unit.

The prediction unit 33 predicts the operating status of the target device using the generated prediction model and the environmental information newly acquired by the acquisition unit 31.

Each unit of the prediction device 30 may execute the above process for each of multiple target devices. Below, we will explain in detail the process examples of the prediction model generation unit 32 and the prediction unit 33.

Figure 6 shows a CNN (Convolutional Neural Network) model unit 321, which is an example of the configuration of a model generated by the prediction model generation unit 32. The CNN model unit 321 outputs output data V2 based on input data V1. As described later, the input data V1, which is an explanatory variable, includes five types of environmental information, and the input size is 5*N. Note that N is the range of time slots and indicates a predetermined time domain (the range of past data used for learning). In addition, the output data V2, which is the objective variable, includes five types of environmental information V2a and one type of operation information V2b, so the output size of the output data V2 is 6. The environmental information V2a is a predicted value of the environmental information at the next time (immediately after) the predetermined time domain, and the operation information V2b is a predicted value of the operation information at the next time (immediately after) the predetermined time domain. In addition, the size of each data is not limited to the case described above.

The CNN model unit 321 has at least a convolutional layer 321a and a fully connected layer 321b for calculations. Each part will be explained below.

The convolutional layer 321a is a layer that extracts features from the input data V1. The convolutional layer 321a has a convolution filter, and may also include a ReLU (Rectified Linear Unit) layer as necessary. The convolutional layer 321a may also include a pooling layer in the subsequent stage (the part connected to the fully connected layer 321b). In this way, the convolutional layer 321a can execute any known process required for convolution processing.

The fully connected layer 321b performs classification processing or regression processing based on the feature amount extracted by the convolution layer 321a. For example, when the fully connected layer 321b predicts the operation information V23b, a threshold value for the operation status may be set, and the operation status may be determined to be normal or abnormal by determining whether the threshold value is greater than the predicted operation information V23b. When the fully connected layer 321b predicts the environmental information V22a, a regression processing may be performed. The fully connected layer 321b may further include at least one or more layers such as a ReLU layer, a dropout layer, a batch normalization layer, and a layer normalization layer, as necessary, in addition to the fully connected layer itself. In addition, one or more output layers may be provided in the intermediate layer of the fully connected layer 321b for the purpose of adding auxiliary loss. The fully connected layer 321b outputs the environmental information V2a and the operation information V2b as output data V2 via an output layer not shown.

In each convolutional layer 321a and fully connected layer 321b, the number of layers of the model, the activation function, the optimization method, the number of epochs, etc. may be hyperparameters that can be freely determined by the user.

Before the CNN model unit 321 actually predicts the operating status of the target device, machine learning must be performed. Below, we will explain the processing performed by the CNN model unit 321 during learning and operation (i.e., when making predictions).

(When studying)
7A and 7B are data given to the CNN model unit 321 as teacher data. This teacher data includes environmental information E1 including temperature, humidity, illuminance, volume, and vibration for each of times T1 to T8, and the operating status of the target device. At each time in FIG. 7A, temperature information is represented by TH1 to TH8, humidity information by HU1 to HU8, illuminance information by IL1 to IL8, volume information by VO1 to VO8, and vibration information by VI1 to VI8. In addition, the operating status is represented by a value between 0 and 1, and the closer the operating status is to 1, the more normal the operation of the target device is, and the closer it is to 0, the more abnormal the operation of the target device is. In FIGS. 7A and 7B, as a simplified example, when the target device is operating normally, 1 is set as the operating status, and when an abnormality occurs in the target device, 0 is set as the operating status.

First, the situation related to FIG. 7A will be described. In this example, the CNN model unit 321 learns a model that predicts the output data V21 including the environmental information V21a and the operation information V21b using the input data V11, which is an explanatory variable. In detail, the CNN model unit 321 learns a model that predicts the detection value of the environmental information V21a and the value of the operation information V21b at the next time T5 from the detection value of the environmental information of the past N (= 4) time slots from time T1 to T4. At this time, the prediction model generation unit 32 extracts data in the range of the input data V11, the environmental information V21a, and the operation information V21b from the data shown in the table of FIG. 7A as a set of variables, and sets them as teacher data for the CNN model unit 321. The CNN model unit 321 uses the sum of squares error as a loss function, for example, for the predicted objective variable and the objective variable in the actual teacher data using this teacher data. The L1 norm or L2 norm may be used as the regularization term for each parameter of the data. For example, the L1 norm is preferable for use in feature selection in a model. However, the loss function and regularization term are not limited to those exemplified above. For example, other types of loss functions such as absolute value error and Huber loss may be used as the loss function.

Next, the situation related to FIG. 7B will be described. In this example, the CNN model unit 321 learns a model that predicts the output data V22 including the environmental information V22a and the operation information V22b using the input data V12, which is an explanatory variable. The input data V12 and the output data V22 are shifted by one time slot from the input data V11 and the output data V21. In detail, the CNN model unit 321 learns a model that predicts the detection value of the environmental information V22a and the value of the operation information V22b at the next time T6 from the detection value of the environmental information in the past N (= 4) time slots from time T2 to T5. At this time, the prediction model generation unit 32 extracts data within the range of the input data V12, the environmental information V22a, and the operation information V22b from the data shown in the table of FIG. 7B as a set of variables, and sets it as the teacher data of the CNN model unit 321. The CNN model unit 321 uses this training data to apply a loss function to the predicted objective variables and the objective variables in the actual training data. Details of this process are as described in FIG. 7A.

The prediction model generation unit 32 then sets the time slots of the input data V11 and the output data V21 to be shifted by one time in sequence. The CNN model unit 321 performs the same processing as described above, using the input data V11 and the output data V21 obtained by shifting as training data. This processing is performed across the range of all time slots of the training data, so that the prediction model generation unit 32 updates the model weights in the CNN model unit 321 and improves the accuracy of the prediction model.

(during operation)
FIG. 7C shows data given to the CNN model unit 321 during operation. This teacher data includes environmental information E1 including temperature, humidity, illuminance, sound volume, and vibration for each of times T11 to T14, and the operating status of the target device. Note that time T14 is the current time, and time T15 is the future time to be predicted. At each time in FIG. 7C, temperature information is represented by TH11 to TH14, humidity information is represented by HU11 to HU14, illuminance information is represented by IL11 to IL14, sound volume information is represented by VO11 to VO14, and vibration information is represented by VI11 to VI14. Similarly to FIGS. 7A and 7B, the operating status is represented by a numerical value between 0 and 1. The prediction unit 33 extracts input data V13 from the data shown in the table of FIG. 7C, and sets it as input data for the CNN model unit 321. Here, the input data V13 is the detected value of the environmental information in N (=4) time slots from time T11 to time T14.

Figure 8 is a flowchart showing an example of a typical process of the prediction device 30. First, the acquisition unit 31 of the prediction device 30 acquires the operating status of the target device as operating information and acquires environmental information around the target device as teacher data (step S31). Next, the prediction model generation unit 32 generates a prediction model using the teacher data (step S32). For example, the prediction model generation unit 32 generates a prediction model by training the above-mentioned CNN model unit 321. This is the learning step of the prediction model. Then, the prediction unit 33 predicts the future operating status of the target device by inputting environmental information (e.g., including current environmental information) newly acquired by the acquisition unit 31 into the generated prediction model (step S33).

In the above example, input data V13, which is an explanatory variable, is input to the CNN model unit 321, and the CNN model unit 321 outputs output data V23 (prediction target) including environmental information V23a and operation information V23b. The environmental information V23a is the detected value of the environmental information at the next time T15, and the operation information V23b is the value of the vibration status at time T15. In this manner, the prediction device 30 predicts future environmental information and operation information.

The prediction device 30 can execute, for example, the following processes depending on the prediction result. For example, when the operation information V23b is 0, the prediction device 30 may notify the user of a warning that "an abnormality is occurring in the target device" via a notification unit capable of notifying by image display, sound, etc. Also, when the operation information V23b is not 0 but is smaller than a predetermined threshold value X1 (X1 is 0 or more and less than 1), the prediction device 30 may notify the user of a warning that "there is a sign of an abnormality in the target device" via a notification unit capable of notifying by image display, sound, etc.

Furthermore, after predicting the environmental information V23a, the prediction device 30 may compare the value of the predicted environmental information V23a with the value of the actual environmental information V23a at the timing when the detection value of the actual environmental information V23a is acquired by the acquisition unit 31, and calculate the difference between each of them. As an example, if the difference value of at least one of the five types of environmental information is larger than a predetermined threshold value X2, it is assumed that a sudden change has occurred in the environment. In that case, the prediction device 30 may notify the user via the notification unit of a warning that "the target device is showing signs of an abnormality."

As described above, the prediction device 30 sets multiple sets of variables defined by shifting the time slots of the explanatory variables and the objective variable by one in sequence, and generates a prediction model (CNN model unit 321) of the operating status using these multiple sets of variables as training data. Then, the generated prediction model and the environmental information V13 newly acquired by the acquisition unit 31 are used to predict the operating status V23b of the target device. This process enables the prediction device 30 to accurately predict the operating status V23b of the target device.

In order for the prediction device 30 to generate this prediction model of the operating status with high accuracy, it is preferable to obtain information on "the circumstances under which the target device went down" with high accuracy. Therefore, when the prediction device 30 is applied to the monitoring server 20 shown in (1B), the prediction device 30 can obtain environmental information from the peripheral device 10 when the target device went down, and the prediction device 30 can generate a prediction model with high accuracy. Therefore, the prediction device 30 can be operated more effectively.

In particular, if the target device is a computer installed in a factory, it is expected that the target device will repeatedly execute some kind of processing in conjunction with product manufacturing, etc. Accordingly, it is expected that the fluctuations in the value of the environmental information will be periodic. CNN is suitable for detecting such periodic fluctuations in data. The prediction device 30 monitors such periodic processing, and as described above, can notify factory workers of an abnormality when it detects an abnormality in the target device. The periodicity of the data is, for example, in units of tens of seconds to several hours, and the prediction device 30 may detect signs of an abnormality in the target device by using the most recent data containing such periodicity as training data.

The prediction device 30 may also include environmental information at a timing immediately after a specified time region as the objective variable to be predicted. This allows the prediction device 30 to determine the possibility that an abnormality has occurred in the environment, as described above, thereby increasing the possibility of detecting an abnormality in the target device. However, although the prediction device 30 includes operation information at a timing immediately after a specified time region as the objective variable to be predicted, it is not necessary for it to include environmental information.

(1D)
9 is a block diagram showing an example of a determination device 40. The determination device 40 monitors a plurality of target devices, and includes a receiver (transmitter) 41, an operation information acquisition unit 42, a memory unit (storage) 43, a state determination unit 44, and an instruction output unit 45. Each of these components will be described below.

The receiver 41 receives environmental information about the surroundings of the multiple target devices from multiple peripheral devices provided corresponding to each of the multiple target devices. The peripheral devices may be the peripheral device 10 shown in (1A) or may be a known environmental sensor that does not have a rechargeable battery. The peripheral devices may transmit environmental information wirelessly or via a wired connection.

Each peripheral device transmits identification information for identifying itself along with the environmental information. The determination device 40 compares the identification information of each peripheral device stored in advance inside the device to determine which peripheral device transmitted the environmental information (i.e., which target device transmitted the environmental information).

The operation information acquisition unit 42 acquires the operation status of each of the multiple devices as operation information via wireless or wired communication. The operation information acquisition unit 22 may acquire this operation information, for example, from the target device, or may acquire this operation information from another device, such as the peripheral device 10.

The storage unit 43 stores a first device among a plurality of target devices that are in a similar environment in association with a second device that is different from the first device. Note that a "similar environment" may mean, for example, that in the environment around the first device and the second device, the difference in at least one detection value detected by a peripheral device among the temperature, humidity, acceleration, vibration, sound, air pressure, wind, illuminance, ultraviolet light, the content of a substance in the air such as VOCs, is equal to or less than a predetermined threshold value.

As another example of a "similar environment," it may refer to a case where multiple target devices are installed in a certain location, and among the multiple target devices, a first device and a second device are closest to each other. As yet another example, it may refer to a case where multiple target devices are installed in a factory, and the first device and the second device perform the same process on the line. Even in such cases, it is assumed that the surrounding environments of the first device and the second device will be similar.

The state determination unit 44 determines whether or not there is an abnormality in the first device using the operation information acquired by the operation information acquisition unit 42 and the environmental information acquired by the receiving unit 41. The state determination unit 44 may, for example, determine whether or not the target device is currently operating normally using the operation information and the environmental information to determine whether or not there is an abnormality. In another example, the state determination unit 44 may use the operation information and the environmental information to predict whether or not there is a sign of an abnormality (i.e., whether or not an abnormality is likely to occur in operation in the future), and may determine that there is an abnormality if there is a sign of an abnormality. For example, the state determination unit 23 may make this prediction by generating a model that uses the environmental information and operation information as teacher data, as shown in (1C).

When the state determination unit 44 determines that the first device has an abnormality, the instruction output unit 45 outputs at least one of the following instructions. That is, it outputs an instruction to increase the frequency of transmitting environmental information to a peripheral device provided corresponding to the second device, and outputs an instruction to stop processing to the second device. Note that stopping processing may mean that the second device does not execute processing unless instructed by an operator, or that the second device stops processing for a predetermined period and resumes processing after the predetermined period has elapsed. The instruction output unit 45 refers to the storage unit 43 and recognizes that the second device is associated with the first device determined by the state determination unit 44, thereby recognizing that the second device is an object requiring attention other than the first device.

In addition, in parallel with the above processing, the instruction output unit 45 may execute at least one of outputting an instruction to increase the frequency of transmitting environmental information to a peripheral device provided corresponding to the first device, and outputting an instruction to the first device to stop processing. In addition, the destinations of these devices or peripheral devices to which the instructions are output may be stored in the storage unit 43, for example. The instruction output unit 45 can output instructions by referring to this information.

Figure 10 is a flowchart showing an example of a typical process of the determination device 40. First, the receiving unit 41 of the determination device 40 receives environmental information about the surroundings of the multiple devices from the multiple peripheral devices (step S41). The operation information acquiring unit 42 acquires the operation status of each of the multiple devices as operation information (step S42). Note that step S42 may be executed before or after step S41, or may be executed at the same time as step S41.

The state determination unit 44 determines whether or not there is an abnormality in the first device using the operation information acquired by the operation information acquisition unit 42 and the environmental information acquired by the reception unit 41 (step S43). If there is an abnormality in the first device (Yes in step S43), the instruction output unit 45 outputs the above-described instruction to the second device (step S44). On the other hand, if there is no abnormality in the first device (No in step S43), the instruction output unit 45 does not output an instruction to the second device.

As described above, the first device and the second device are in a similar environment. Therefore, when an abnormality is determined in the first device, it is possible that an abnormality has occurred in the second device in a similar environment, or that an abnormality may occur in the future. Therefore, the instruction output unit 45 can execute a process corresponding to that possibility. For example, the instruction output unit 45 outputs an instruction to a peripheral device provided corresponding to the second device to increase the frequency of transmitting environmental information, so that the peripheral device increases the frequency of transmitting environmental information. This allows the determination device 40 to more reliably obtain data related to an abnormality in the vicinity of the second device. In addition, the instruction output unit 45 outputs an instruction to the second device to stop processing, so that it is possible to suppress malfunctions in the processing executed by the second device when an abnormality has occurred in the second device, and to suppress actual occurrence of an abnormality when a sign of an abnormality has occurred in the second device.

(1E)
11 is a block diagram showing an example of a determination device 50. The determination device 50 monitors a target device, and includes an acquisition unit 51, a power spectrum calculation unit 52, a similarity calculation unit 53, and a determination unit 54. Each of these components will be described below.

The acquisition unit 51 acquires the status of one target device as status information related to at least one of sound and vibration. The target device has components that emit relatively loud noises, such as a hard disk drive (HDD) or a fan, and the determination device 50 can be used to detect abnormalities in the components. The acquisition unit 51 may acquire the operating status from the target device, or may acquire the operating status from another device, such as a peripheral device provided in correspondence with the target device and having a sound pressure sensor, a vibration sensor, or the like. An example of this peripheral device is the peripheral device 10 described in (1A).

The power spectrum calculation unit 52 calculates the power spectrum for the status information acquired by the acquisition unit 51 at three or more different timings. The timings may be any time interval. For example, the timings may be relatively long intervals such as several weeks, one to several months, or several years. This is because, as described below, the determination device 50 may determine not only abnormalities in the target device but also deterioration over time. However, the timings may also be short time intervals such as several tens of seconds to several hours.

The similarity calculation unit 53 calculates at least the similarity between the power spectrum at the first timing and the power spectrum at the second timing, and the similarity between the power spectrum at the first timing and the power spectrum at the third timing. The first timing may be, for example, the closest timing among the multiple timings, or may be any other timing. Furthermore, the similarity calculation unit 53 may calculate the similarity between the power spectrum at the first timing and each of the power spectra at all timings other than the first timing. The similarity calculation unit 53 may set a similarity value between the power spectra at the first timing (for example, set it to 1) and use it to compare each similarity calculated as described above.

The determination unit 54 determines the operating status (particularly the deterioration state) of the target device using the multiple similarities calculated by the similarity calculation unit 53. For example, the determination unit 54 may determine that a malfunction has occurred in the target device when the difference between different similarities (i.e., similarities related to power spectra at different timings) is equal to or greater than a predetermined threshold value. Note that a malfunction includes not only a failure event in which the device is unable to execute a desired process, but also a case in which the device is able to execute a desired process but the sound or vibration in the target device is different from that in a normal case (e.g., an initial state). Other examples of determination will be described later in (2B) of the second embodiment.

Each part of the determination device 50 may perform the above processing for each of multiple target devices.

Figure 11 is a flowchart showing an example of a typical process of the determination device 50. First, the acquisition unit 51 of the determination device 50 acquires state information related to at least one of sound and vibration (step S51). The power spectrum calculation unit 52 calculates the power spectrum of the state information at three or more different times (step S52).

The similarity calculation unit 53 calculates multiple similarities between the power spectra (step S53). The determination unit 54 uses the multiple calculated similarities to determine the operating status of the target device (step S54).

By performing the above process, the determination device 50 can accurately determine the state of the target device (e.g., the state of the parts) based on at least one of the sound and vibration data.

(1A) to (1E) in the first embodiment described above can be used in appropriate combinations. In the second embodiment, we will explain such specific examples.

Embodiment 2
(2A)
Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In (2A), an example in which (1B) to (1D) of the above-mentioned first embodiment are combined will be described.

Figure 13 is a block diagram showing an example of a monitoring system according to (2A). The monitoring system S2 is installed in a factory, and includes multiple (three in this example) peripheral devices, USB sensors 100, and one monitoring server 200. The USB sensors 100A, 100B, and 100C are USB peripherals, and are connected to their respective target devices, factory computers FC1, FC2, and FC3. Hereinafter, the USB sensors 100A to 100C are collectively referred to as the USB sensor 100, and the factory computers FC1 to FC3 are collectively referred to as the computer FC.

In FIG. 13, the solid arrow from the computer FC to the monitoring server 200 indicates that the computer FC is outputting its own operation information to the monitoring server 200. Also, the dashed arrow from the USB sensor 100 to the monitoring server 200 indicates that the USB sensor 100 is wirelessly transmitting environmental information to the monitoring server 200.

Figure 14 is a block diagram showing an example of a USB sensor according to (2A). The USB sensor 100 includes a connection unit 101, a controller 102, a non-volatile storage area 103, an illuminance sensor 104, an impact/vibration sensor 105, a temperature/humidity sensor 106, a wireless communication module 107, and a rechargeable battery 108. Each unit of the USB sensor 100 is controlled by the controller 102. Each component will be described below.

The connection unit 101 corresponds to the power receiving unit 11 in (1A), and when inserted into the USB port of the target device, the computer FC, it connects the computer FC to the peripheral device 10 and receives power from the computer FC.

The controller 102 is composed of a processor, and reads and executes a group of software modules stored in the non-volatile storage area 103, causing each part of the USB sensor 100 to execute the processes described below. The controller 102 also supplies power to the illuminance sensor 104, the shock and vibration sensor 105, the temperature and humidity sensor 106, and the wireless communication module 107.

When the controller 102 is able to receive power from the computer FC via the connection unit 101, it operates by receiving power from the computer FC as shown in FIG. 14. On the other hand, when it is unable to receive power from the computer FC, it operates by receiving power from the rechargeable battery 108.

The non-volatile storage area 103 stores the software modules executed by the controller 102, as well as detection values detected by the illuminance sensor 104, the shock and vibration sensor 105, and the temperature and humidity sensor 106. This detection value data is stored by the controller 102 and can also be read out from the controller 102. The non-volatile storage area 103 may be configured with any type of non-volatile storage area, and an example is an SSD (Solid State Drive).

The illuminance sensor 104 is a sensor that detects the illuminance around the computer FC, which is the target device, and outputs the detected value to the controller 102. The shock/vibration sensor 105 is a sensor that detects vibrations (especially shocks) around the computer FC, and outputs the detected value to the controller 102. The temperature/humidity sensor 106 is a sensor that detects the temperature and humidity around the computer FC, and outputs the detected value to the controller 102. Each of the above sensors corresponds to the environmental sensor 12 in (1A). The controller 102 stores the detected values (environmental information) output by each sensor in the non-volatile storage area 103.

The wireless communication module 107 corresponds to the transmitter 13 in (1A), and acquires environmental information stored in the non-volatile storage area 103 from the controller 102, and transmits the environmental information wirelessly to the monitoring server 200. The wireless communication module 107 can perform communication, for example, using a low power wide area (LPWA) communication method.

The controller 102 periodically (cyclically) transmits the environmental information stored in the non-volatile storage area 103 to the monitoring server 200. After transmitting the environmental information, the controller 102 stores the environmental information acquired by each sensor in the non-volatile storage area 103 until the next transmission of the environmental information, and transmits the stored environmental information when the next environmental information is transmitted. The period for transmitting the environmental information can be any period, such as 1 minute, 10 minutes, etc. When transmitting, the controller 102 also transmits the identification information of the USB sensor 100 together with the environmental information.

The rechargeable battery 108 corresponds to the rechargeable battery 14 in (1A). When the rechargeable battery 108 is powered by the computer FC via the connection unit 101, as shown in FIG. 14, it is charged by the power supplied from the computer FC. On the other hand, when the rechargeable battery 108 is not powered by the computer FC, it supplies the charged power to the controller 102. The controller 102 supplies this power to the sensors 104 to 106 and the wireless communication module 107. In other words, the rechargeable battery 108 supplies power to the controller 102 so that the sensors can acquire and wirelessly transmit environmental information for a predetermined period of time even after the computer FC performs an abnormal operation such as an abnormal termination. This predetermined period is a period sufficient for the monitoring server 200 to generate an accurate prediction model based on the environmental information, such as a period of several minutes.

Figure 15 shows the state when computer FC3 abnormally terminates in monitoring system S2 in Figure 13. Because computer FC3 abnormally terminates, it cannot transmit its own operation information to the monitoring server 200. However, USB sensor 100C can obtain and transmit environmental information around computer FC3 for a specified period of time.

Figure 16 is a block diagram showing the state of the USB sensor 100C in Figure 15. For the reasons described above, the USB sensor 100C cannot receive power from the computer FC3. However, the rechargeable battery 108 in the USB sensor 100C supplies power to the controller 102, making it possible for the USB sensor 100C to acquire and transmit environmental information.

The controller 102 can also function as a failure determination unit. Specifically, the controller 102 may detect that the power supply from the computer FC has been cut off or that the power supply has become unstable, and in that case, determine that an abnormality has occurred in the computer FC. The controller 102 may also detect that the detection value of the environmental information detected by at least one of the illuminance sensor 104 to the temperature and humidity sensor 106 has exceeded a predetermined threshold, and in that case, determine that there is a sign of an abnormality in the computer FC. For example, this applies when the shock and vibration sensor 105 detects an impact that is equal to or greater than a predetermined threshold (i.e., detects an explosion). In such a case, as described in (1A), the controller 102 can transmit the environmental information stored in the non-volatile storage area 103 regardless of the regular timing. For example, the controller 102 can transmit the environmental information at the time when the above-mentioned event is detected.

Next, the monitoring server 200 will be described in detail. FIG. 17 is a block diagram showing an example of the monitoring server 200. The monitoring server 200 is connected to an externally connected database DB, and is capable of inputting and outputting information to and from the database DB. However, the database DB may be provided inside the monitoring server 200. The monitoring server 200 comprises a receiving unit 201, a learning unit 202, a model saving unit 203, an inference unit 204, and a notification unit 205. Each of these components will be described below.

The receiving unit 201 corresponds to the acquiring unit 31 in (1C) and receives the environmental information transmitted from the USB sensor 100 via wireless R2 and the operation information transmitted from the computer FC. Note that since the USB sensor 100 periodically transmits environmental information, the receiving unit 201 periodically receives new environmental information.

The receiving unit 201 determines which USB sensor 100 transmitted the environmental information (i.e., which computer FC transmitted the environmental information) by comparing it with the identification information of each USB sensor 100 stored in the monitoring server 200. In addition, the operation information transmitted from the computer FC is provided with identification information that allows the computer FC to identify itself. The receiving unit 201 determines which computer FC transmitted the environmental information by comparing it with the identification information of each computer FC stored in the monitoring server 200. The receiving unit 201 associates the environmental information and operation information of a certain computer FC identified in this manner with the identification information of the identified computer FC and stores them in the database DB as learning data.

The learning unit 202 corresponds to the prediction model generation unit 32 in (1C), and for a certain computer FC, reads the environmental information and operation information stored in the database DB using the identification information of that computer FC, and generates teacher data. Then, using this teacher data, a prediction model for each computer FC is generated. Details of this process are as described in (1C). The processing of the learning unit 202 may be performed by instructions from an operator who operates the monitoring server 200, or may be automatically executed by the monitoring server 200. For example, for a certain computer FC, the learning unit 202 may update the prediction model using data related to that computer FC that has been stored in the database DB since the previous time, after a predetermined period has elapsed since the previous prediction model was generated or updated.

The model storage unit 203 stores the prediction model of each computer FC generated by the learning unit 202 in association with the identification information of that computer FC. Note that, when the learning unit 202 generates an updated version of the prediction model, the prediction model stored in the model storage unit 203 is updated to the one generated by the learning unit 202. Also, the model storage unit 203 may be provided externally instead of inside the monitoring server 200.

The inference unit 204 corresponds to the prediction unit 33 in (1C). When the receiving unit 201 receives new environmental information for a certain computer FC, the inference unit 204 acquires a prediction model associated with the identification information of the computer FC from the model storage unit 203. The inference unit 204 inputs the newly acquired environmental information as an explanatory variable into the prediction model to acquire predicted values of future environmental information and operation information, which are the objective variables. Details of this process are as described in (1C). Then, information indicating the presence or absence of an abnormality or a sign of an abnormality regarding the inferred computer FC is output to the notification unit 205. When the inference unit 204 receives new environmental information for each USB sensor 100, it executes this process for each corresponding computer FC. The inference unit 204 may also associate the inference result with data regarding the inferred computer FC stored in advance in the database DB and store it in the database DB.

The notification unit 205 notifies the presence or absence of an abnormality or a sign of an abnormality in the computer FC by at least one of display and sound based on the information output from the inference unit 204. The notification unit 205 has, for example, a display, a speaker, etc. Furthermore, the notification unit 205 may output an instruction to a computer FC for which an abnormality or a sign of an abnormality has been determined, to notify that the computer FC has an abnormality or a sign of an abnormality. Furthermore, if necessary, the notification unit 205 may output an instruction to stop processing of the computer FC.

In addition, when the receiving unit 201 receives environmental information at irregular timing, the inference unit 204 may determine which USB sensor 100 sent what information and notify the operator of the information by the notification unit 205. For example, when the impact/vibration sensor 105 detects a strong impact equal to or greater than a predetermined threshold and the USB sensor 100 sends environmental information, the inference unit 204 determines from the environmental information that a strong impact was detected. In addition, when the temperature/humidity sensor 106 detects a high temperature equal to or greater than a predetermined threshold and the USB sensor 100 sends environmental information, the inference unit 204 determines from the environmental information that a high temperature was detected. The inference unit 204 outputs information related to this determination to the notification unit 205. The notification unit 205 notifies the operator of the information related to this determination by at least one of display and sound based on the information output from the inference unit 204.

As described above, since the USB sensor 100 is inserted into the USB port of the computer FC, it is always powered by the computer FC under normal circumstances. Therefore, since the USB sensor 100 does not need to use power under normal circumstances, it can store power in the rechargeable battery 108 to operate in the event of a system failure or the like of the computer FC. In addition, even if the system failure of the computer FC occurs (especially after the power supply is cut off), it can still acquire and transmit environmental information.

The monitoring server 200 can aggregate and analyze the information sent from each USB sensor 100 to determine whether there is an abnormality or a sign of an abnormality in the computer FC. Furthermore, if an abnormality or a sign of an abnormality is detected, the monitoring server 200 can execute the processing required for maintenance of the computer FC via the notification unit 205.

The model storage unit 203 may store multiple computers FC in similar environments in association with each other. The definition of a similar environment is as described in (1D). Hereinafter, the two computers FC in similar environments are referred to as C1 and C2. Using this information, the inference unit 204 can reuse the prediction model of C1 stored by the model storage unit 203 for C2, which is in a similar environment, thereby reducing the learning time. In other words, the inference unit 204 can input the newly acquired environmental information of C2 into the prediction model of C1, thereby executing a judgment regarding an abnormality of C2. Of course, processing with C1 and C2 reversed is also possible.

Furthermore, the inference unit 204 may derive inference result 1 in which newly acquired environmental information for C1 is input to the prediction model for C1, as well as inference result 2 in which newly acquired environmental information for C1 is input to the prediction model for C2. The inference unit 204 can refer to both and determine whether or not an abnormality or a sign of an abnormality has occurred in C1. For example, even if no abnormality has occurred in inference result 1, if an abnormality has occurred in inference result 2, the inference unit 204 can determine that an abnormality has occurred in C1.

(2B)
In (2B), an example of combining (1B) with (1E) will be described. The overall configuration of the monitoring system according to (2B) is the same as that described in FIG. 13 of (2A), and therefore will not be described.

Figure 18 is a block diagram showing an example of a USB sensor in (2B). Compared to the USB sensor 100, this USB sensor 100 includes a sound sensor 109 instead of the illuminance sensor 104.

The sound sensor 109 is a sensor that detects sound (e.g., sound pressure) around the target device, computer FC, and outputs the detection value to the controller 102. The controller 102 stores the detection value of the sound sensor 109 as environmental information in the non-volatile storage area 103, similar to the detection values output by other sensors. The controller 102 periodically (periodically) transmits this stored environmental information to the monitoring server 200 using the wireless communication module 107. Note that the configuration of the USB sensor 100 in (2B) other than the sound sensor 109 is the same as in (2A), so a description will be omitted.

Next, the monitoring server 200 in (2B) will be described in detail. FIG. 19 is a block diagram showing an example of the monitoring server 200. The monitoring server 200 is connected to an externally connected database DB, and is capable of inputting and outputting information to and from the database DB. However, the database DB may be provided inside the monitoring server 200. The monitoring server 200 comprises a receiving unit 211, a power spectrum calculation unit 212, a similarity calculation unit 213, a determination unit 214, and a notification unit 215. Each of these components will be described below.

By executing the same process as the receiving unit 201, the receiving unit 211 identifies which computer FC the received environmental information is environmental information for, and stores the environmental information in the database DB as learning data. The details of this process are as described in (2A). Although not described in detail here, the receiving unit 211 may also obtain operation information from the computer FC and store this data in the database DB as learning data.

Figure 20 is a diagram showing an example of sound data for computer FC1 stored in the DB. The DB stores current sound data D0, sound data D1 from one month ago, sound data D2 from two months ago, ... sound data DN from N months ago for computer FC1. This sound data D is received from USB sensor 100A. The monitoring server 200 uses this data to monitor computer FC1 at monthly intervals.

The power spectrum calculation unit 212 corresponds to the power spectrum calculation unit 52 in (1E), and performs a Fourier transform on each piece of sound data D stored in the database DB to calculate the power spectrum (frequency characteristics) related to each frequency.

The similarity calculation unit 213 corresponds to the similarity calculation unit 53 in (1E). Specifically, the similarity calculation unit 213 selects the power spectrum (the power spectrum at the most recent timing) P0 of the current sound data D0 and the power spectrum P1 of the sound data D1 from one month ago, and calculates the similarity R1 between the two. Similarly, the similarity calculation unit 213 selects the power spectrum P0 and the power spectrum P2 of the sound data D2 from two months ago, and calculates the similarity R2 between the two. In this way, the similarity calculation unit 213 calculates the similarity R between the power spectrum P0 and the power spectrum of any one of the past sound data D for all past sound data. Finally, the similarity calculation unit 213 selects the power spectrum P0 and the power spectrum PN of the sound data DN from N months ago, and calculates the similarity RN between the two. In addition, the similarity calculation unit 213 sets the similarity R0 between the power spectra P0 to 1 for the purpose of comparing the similarities.

At this time, the similarity calculation unit 213 normalizes the distribution of each power spectrum P so that the average is 0 and the variance is 1, and then calculates the cross-correlation between the power spectra P to calculate the similarity. However, the normalization method is not limited to this. The similarity calculation unit 213 outputs the similarity of the power spectra P between different timings to the determination unit 214 as a sequence of size N. The similarity calculation unit 213 also outputs to the determination unit 214 that the similarity R0 between the power spectra P0 is 1.

The determination unit 214 corresponds to the determination unit 54 in (1E), and uses the similarity calculated by the similarity calculation unit 213 to determine the operating status of the computer FC1. Specifically, the determination unit 214 plots the calculated multiple similarities R from R0 to RN in chronological order, performs linear fitting on the plotted graph, and determines whether the graph approximates a straight line. Whether the graph approximates a straight line can be determined, for example, by whether the difference in similarity over a certain month is greater than or equal to a predetermined threshold value compared to the slope of the graph over another period. If the graph does not approximate a straight line, the determination unit 214 identifies the location of the curved point in the graph, and can estimate the time when a malfunction occurred in the computer FC1 based on the location of the curved point. In this way, the determination unit 214 determines that a malfunction has occurred in the computer FC1.

(i) For example, the determination unit 214 determines whether the difference DI between similarity R0 and similarity R1 (i.e., the slope of the graph for the most recent month) is greater than the slope of the graph for other periods by a predetermined threshold or more. If DI is greater than the slope of the graph for other periods by a predetermined threshold or more, it is determined that the slope of the graph (the rate of change of similarity) has changed rapidly over the most recent month, and therefore the plotted graph does not approximate a straight line. Therefore, the determination unit 214 identifies that there is a bend in the graph during the most recent month, and estimates that a malfunction occurred in computer FC1 during the most recent month. Note that the determination unit 214 can estimate that a malfunction occurred not only during the most recent month, but also for other periods.

When the graph approximates a straight line, the determination unit 214 determines whether the slope of the entire graph is equal to or greater than a predetermined threshold. (ii) When the slope of the graph is equal to or greater than a predetermined threshold, the determination unit 214 estimates that deterioration has occurred in the computer FC1. This is because it can be interpreted that the sound data is changing at a substantially constant rate over time. (iii) When the slope of the graph is less than the predetermined threshold, the determination unit 214 estimates that no abnormality has occurred in the computer FC1. This is because it can be interpreted that the sound data is barely changing between the present and the past.

Figure 21 is an example of a graph plotting similarities R0 to RN. Graphs (i) to (iii) in Figure 21 are examples of graph shapes corresponding to (i) to (iii) above. In this way, the judgment unit 214 can accurately judge the state of computer FC1.

The determination unit 214 outputs the above determination results to the notification unit 215. If the determination unit 214 determines that at least one of a malfunction has occurred in computer FC1 and that computer FC1 has deteriorated over time, the determination unit 214 may generate notification content urging the operator to inspect or replace computer FC1, and output the notification content together with the determination result to the notification unit 215.

The notification unit 215 has a configuration similar to that of the notification unit 205 in (2A), and notifies the operator of the judgment result for the computer FC by at least one of display and sound. For example, if the judgment unit 214 identifies the time when a malfunction occurred in the computer FC1 in the past month, the notification unit 215 displays the content "A sudden change has been observed in the operating sound of the computer FC1 in the past month. As a malfunction may have occurred, we recommend that you promptly inspect or replace the computer FC1." Also, if the judgment unit 214 judges that the computer FC1 has deteriorated due to aging, the notification unit 215 displays the content "No malfunction has been observed in the computer FC1, but it may have deteriorated due to aging. We recommend that you inspect the computer FC1." If the judgment unit 214 judges that no abnormality has occurred in the computer FC1, the notification unit 215 displays the content "No abnormality has been observed in the computer FC1."

The above processing is performed for each of the multiple computers FC. Also, while sound data has been used as an example, similar processing can be performed for vibration (e.g., impact) data.

Figure 22A is a flowchart showing an example of a typical process of the monitoring server 200. First, the receiver 211 of the monitoring server 200 receives sound and vibration data, which is environmental information, from each USB sensor 100 (step S61). The received sound and vibration data is stored in the database DB.

The power spectrum calculation unit 212 performs a Fourier transform on each piece of sound data D stored in the database DB to calculate a power spectrum (step S62). The similarity calculation unit 213 uses each power spectrum to calculate the similarity R between the power spectra (step S63).

The determination unit 214 plots the calculated similarities R in chronological order and determines whether the plotted graph is similar to a straight line, i.e., whether the graph is linear (step S64). If the graph is not linear (No in step S64), the determination unit 214 identifies the location of the inflection point in the graph to determine that a malfunction has occurred in the computer FC1 and when the malfunction occurred (step S65).

If the determination in step S64 shows that the graph is linear (Yes in step S64), the determination unit 214 determines whether the slope of the entire graph is equal to or greater than a predetermined threshold (step S66). If the slope of the graph is equal to or greater than a predetermined threshold (Yes in step S66), the determination unit 214 determines that deterioration has occurred in the computer FC1 (step S67). On the other hand, if the slope of the graph is less than the predetermined threshold (No in step S66), the determination unit 214 determines that no abnormality has occurred in the computer FC1 (step S68). Furthermore, in each of steps S65, S67, and S68, the determination unit 214 outputs notification content based on the determination result to the notification unit 215.

As described above, the monitoring server 200 can detect changes in the status of the HDD and fans by examining fluctuations in sound and vibration values at relatively long intervals.

In addition, (2B) may be further combined with the monitoring server processing shown in (1C). This allows the monitoring server to determine not only long-term changes in the state of the computer FC, but also the presence or absence of short-term abnormalities in the computer FC.

Computer monitoring systems include those that work in conjunction with an OS (Operation System) and those that incorporate a separate power supply or OS. In order to detect an abnormality in a computer to be monitored, it is preferable to detect various information before and after the abnormality occurs in the computer. However, the former monitoring system requires that the computer to be monitored is operating normally to a certain extent, and there is a possibility that information cannot be obtained at the moment the computer goes down. The latter monitoring system uses a separate power supply or system to investigate fault information and save data after the computer goes down. Therefore, like the former, there is a possibility that information cannot be obtained at the moment the computer system goes down. In addition, the latter requires the preparation of dedicated hardware or a separate power supply system, so it is basically not a system that can be retrofitted to a computer. Therefore, it is not easy to add a status monitoring function to a computer that does not have the status monitoring function.

In contrast, the computer monitoring systems exemplified in (2A) and (2B) acquire computer information in systems that require high reliability, such as factory equipment. By utilizing the acquired information, the monitoring system can grasp the situation when a computer failure occurs and detect signs of computer failure, thereby improving computer availability. For example, the monitoring system can accurately acquire information at the moment a computer system goes down, so it can grasp detailed information at the time of failure and determine signs of computer failure with high accuracy. It can also achieve thorough monitoring of computers that operate 24 hours a day in factories.

Note that the present disclosure is not limited to the above-described embodiment, and can be modified as appropriate without departing from the spirit of the present disclosure. For example, the timing for calculating the power spectrum in (1E) and (2B) does not have to be at the same interval.

In (2B), when the graph approximates a straight line, the determination unit 214 may determine whether the slope of the entire graph is greater than or equal to two or more different thresholds. For example, when the slope of the entire graph of computer FC1 is equal to or greater than a large threshold Th1, the determination unit 214 determines that deterioration over time is progressing rapidly. The determination unit 214 generates a notification to the operator indicating the points that seem to require prompt inspection or replacement of computer FC1, and outputs the notification to the notification unit 215.

On the other hand, if the slope of the entire graph for computer FC1 is less than the large threshold Th1 but equal to or greater than the small threshold Th2 (Th1>Th2), the judgment unit 214 judges that deterioration is progressing slowly. The judgment unit 214 generates a notification recommending that the operator have regular inspections of computer FC1, and outputs this to the notification unit 215. Also, if the slope of the entire graph for computer FC1 is less than the small threshold Th2, the judgment unit 214 judges that no deterioration is occurring. This is as described in step S68 of FIG. 22B. In this way, the judgment unit 214 may judge different states of computer FC1 based on the magnitude relationship with two or more different thresholds, and change the content of the notification to the operator by the notification unit 215 accordingly.

In (1D) or (2A), the information of multiple devices stored as being in a similar environment among multiple target devices (computers) can be updated by an operator and the determination device 40 (or the monitoring server 200). For example, the determination device 40 in (1D) may determine target devices in a similar environment using data on the environmental information and operation information of the target devices.

In detail, the state determination unit 44 of the determination device 40 calculates the similarity of time series data of environmental information (including at least one of the various parameters mentioned above, such as temperature and humidity) for a predetermined period between the target devices. Then, if the similarity is equal to or greater than a predetermined threshold, the state determination unit 44 determines that the two target devices are in similar environments, and if the similarity is less than the predetermined threshold, it determines that the two target devices are not in similar environments.

The state determination unit 44 may further calculate the similarity of the time series data of the operation information for a predetermined period and reflect the similarity in the determination. For example, when the comparison of the similarity regarding the environmental information described above determines that the two target devices are in similar environments, if the similarity of the time series data of the operation information of the two devices is less than a predetermined value, the state determination unit 44 may determine that the two target devices are not in similar environments. If the similarity of the time series data of the operation information of the two devices is equal to or greater than a predetermined value, the state determination unit 44 determines that the two target devices are not in similar environments.

The state determination unit 44 then stores information about target devices determined to be in a similar environment in the storage unit 43, and does not store information about target devices determined not to be in a similar environment in the storage unit 43 (or deletes the information from the storage unit 43). Note that the state determination unit 44 calculates the similarity of the above-mentioned environmental information and operation information by applying known techniques such as distance calculations such as Euclidean distance and clustering (k-means method, hierarchical type, etc.).

The controller 102 of the USB sensor 100 shown in (2A) or (2B) may charge the rechargeable battery 108 at a first charging rate up to a certain charge level (e.g., about 80% of full charge), and then change the charging rate to a second charging rate that is slower than the previous charging rate, or may stop charging. This allows the controller 102 to extend the life of the rechargeable battery 108.

When the monitoring server 200 determines that there is an abnormality or a sign of an abnormality in a certain computer FC as described in the first and second embodiments, the monitoring server 200 may transmit an instruction to the USB sensor 100 connected to the computer FC to change the charging method of the rechargeable battery 108. For example, even if the rechargeable battery 108 has been charged to a certain charge value, the monitoring server 200 may allow the rechargeable battery 108 to quickly achieve 100% charge by keeping the charging speed unchanged and setting it to a first charging speed (or a charging speed faster than the second charging speed). The monitoring server 200 can execute such processing in the following cases, for example.
・When it is determined in step S33 of (1C) that there is an abnormality (or a sign of an abnormality)・When it is determined in step S43 of (1D) that there is an abnormality・When it is determined in (2A) that an abnormality has occurred in the target device C1 due to the use of a prediction model of the target device C2 similar to the target device C1・When it is determined in step S65 of (2B) that a malfunction has occurred・When it is determined in step S67 of (2B) that aging deterioration has occurred, and the slope of the entire graph is greater than the above-mentioned threshold value Th1 for determining aging deterioration, and it is determined that aging deterioration is progressing rapidlyIn these cases, it is considered that there is a relatively high possibility that an abnormal termination (e.g., a system down) will occur in the target device in the near future. Therefore, it is considered that it is better to charge the rechargeable battery 108 close to 100% so that environmental information can be reliably obtained when an abnormal termination occurs.

In the above embodiment, this disclosure has been described as a hardware configuration, but this disclosure is not limited to this. This disclosure can also be realized by having a processor in a computer execute a computer program to execute the processing (steps) of the device (any of a peripheral device, a monitoring server, a determination device, or a prediction device) described in the above embodiment.

Fig. 23 is a block diagram showing an example of the hardware configuration of an information processing device (signal processing device) on which the processing of each of the above-described embodiments is executed. Referring to Fig. 23, this information processing device 90 includes a signal processing circuit 91, a processor 92, and a memory 93.

The signal processing circuit 91 is a circuit for processing signals according to the control of the processor 92. The signal processing circuit 91 may also include a communication circuit for receiving signals from a transmitting device.

The processor 92 reads and executes software (computer programs) from the memory 93 to perform the processing of the device described in the above embodiment. As an example of the processor 92, one of a CPU (Central Processing Unit), an MPU (Micro Processing Unit), an FPGA (Field-Programmable Gate Array), a DSP (Demand-Side Platform), and an ASIC (Application Specific Integrated Circuit) may be used, or multiple of these may be used in parallel.

The memory 93 is composed of a volatile memory, a non-volatile memory, or a combination of both. The number of memories 93 is not limited to one, and multiple memories may be provided. The volatile memory may be, for example, a RAM (Random Access Memory) such as a DRAM (Dynamic Random Access Memory) or an SRAM (Static Random Access Memory). The non-volatile memory may be, for example, a ROM (Random Only Memory) such as a PROM (Programmable Random Only Memory) or an EPROM (Erasable Programmable Read Only Memory), or an SSD (Solid State Drive).

The memory 93 is used to store one or more instructions. Here, the one or more instructions are stored in the memory 93 as a group of software modules. The processor 92 can perform the processing described in the above embodiment by reading and executing these groups of software modules from the memory 93.

Note that the memory 93 may include memory built into the processor 92 in addition to memory provided outside the processor 92. The memory 93 may also include storage located away from the processors constituting the processor 92. In this case, the processor 92 can access the memory 93 via an I/O (Input/Output) interface.

As described above, one or more processors in each device in the above-mentioned embodiments execute one or more programs including a set of instructions for causing a computer to execute the algorithm described using the drawings. This process realizes the signal processing method described in each embodiment.

The program can be stored and supplied to the computer using various types of non-transitory computer readable media. Non-transitory computer readable media include various types of tangible storage media. Examples of non-transitory computer readable media include magnetic recording media (e.g., flexible disks, magnetic tapes, hard disk drives), magneto-optical recording media (e.g., magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R/Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (Random Access Memory)). The program may also be supplied to the computer by various types of transitory computer readable media. Examples of transitory computer readable media include electrical signals, optical signals, and electromagnetic waves. The transitory computer readable media can supply the program to the computer via wired communication paths such as electric wires and optical fibers, or wireless communication paths.

A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes.
(Appendix 1)
a power receiving unit that receives power from a device to be monitored;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit,
When it becomes impossible to receive power via the power receiving unit, the environmental sensor and the transmitting unit are operated by discharged power from the rechargeable battery.
Peripheral devices.
(Appendix 2)
A storage unit that stores the environmental information,
When the transmission unit is receiving power via the power receiving unit, the transmission unit transmits the environmental information stored in the storage unit at regular intervals, and when it becomes impossible to receive power via the power receiving unit, the transmission unit transmits the environmental information stored in the storage unit regardless of the regular intervals.
2. The peripheral device of claim 1.
(Appendix 3)
A storage unit that stores the environmental information;
a determination unit that determines whether or not a detection value of the environmental information detected by the environmental sensor is equal to or greater than a predetermined threshold,
When the determination unit determines that the detection value is less than a predetermined threshold, the transmission unit transmits the environmental information stored in the storage unit at regular intervals, and when the determination unit determines that the detection value is equal to or greater than a predetermined threshold, the transmission unit transmits the environmental information stored in the storage unit regardless of the regular intervals.
2. The peripheral device of claim 1.
(Appendix 4)
The transmission unit transmits the environmental information to a monitoring server,
When the peripheral device receives an instruction from the monitoring server to increase a frequency of transmitting the environmental information, the peripheral device controls the transmission unit to increase a frequency of transmitting the environmental information.
A peripheral device according to any one of claims 1 to 3.
(Appendix 5)
The peripheral device is a USB peripheral.
5. A peripheral device according to any one of claims 1 to 4.
(Appendix 6)
A peripheral device provided in correspondence with the device to be monitored;
a monitoring server that communicates with the peripheral device;
The peripheral device is
A power receiving unit that receives power from the device;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit,
the environmental sensor and the transmission unit are operated by discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit;
The monitoring server includes:
a receiving unit that receives the environmental information from the peripheral device;
an operation information acquisition unit that acquires an operation status of the device as operation information;
a status determination unit that determines a status of the device using the operation information and the environmental information.
(Appendix 7)
The state determination unit is
a prediction model generation unit that sets explanatory variables including at least the environmental information in a predetermined time domain and objective variables including at least the operation information at a timing immediately after the predetermined time domain as a set of variables, and generates a prediction model of the operation status using a set of multiple variables defined by sequentially shifting time slots in the set of variables as training data; and
A prediction unit that predicts the operating status by using the prediction model and the newly acquired environmental information.
(Appendix 8)
the monitoring system includes a plurality of the peripheral devices corresponding to the plurality of the monitored devices,
The monitoring server includes:
a storage unit that stores a first device and a second device that are in a similar environment among the plurality of devices in association with each other;
and an instruction output unit that, when the status determination unit determines an abnormality in the first device, performs at least one of outputting an instruction to the peripheral device corresponding to the second device to increase the frequency at which the transmission unit transmits the environmental information, and outputting an instruction to the second device to stop processing.
(Appendix 9)
a charging step of charging a rechargeable battery by receiving power from the device to be monitored;
an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when power supply from the device becomes impossible;
a transmission step of transmitting the environmental information using discharge power from the rechargeable battery;
The peripherals perform the monitoring method.
(Appendix 10)
a charging step of charging a rechargeable battery by receiving power from the device to be monitored;
an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when power supply from the device becomes impossible;
a transmission step of transmitting the environmental information using discharge power from the rechargeable battery;
A program that causes a peripheral device to execute the above as a monitoring method.
(Appendix 11)
The response variable in the set of variables further includes the environmental information at a timing immediately after the predetermined time region,
the prediction model generation unit generates a prediction model of the operation status and the environmental information using the set of the plurality of variables as training data;
The prediction unit predicts the operating status of the device by using the prediction model and the newly acquired environmental information.
8. The monitoring system of claim 7.
(Appendix 12)
the environmental information includes at least one of operation data of sound and vibration indicating an operation status of the device,
The state determination unit is
a power spectrum calculation unit that calculates a power spectrum of the operation data at three or more different timings;
a similarity calculation unit that calculates at least a similarity between the power spectrum at a first timing and the power spectrum at a second timing, and a similarity between the power spectrum at the first timing and the power spectrum at a third timing;
and a degradation state determination unit that determines a degradation state of the device using the multiple similarities calculated by the similarity calculation unit.
(Appendix 13)
A peripheral device provided in correspondence with the device to be monitored,
a charging step of charging a rechargeable battery by receiving power from the device;
an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when the device is unable to receive the power supply;
A transmission step of transmitting the environmental information to a monitoring server using discharged power from the rechargeable battery;
The monitoring server,
a receiving step of receiving the environmental information from the peripheral device;
an operation information acquisition step of acquiring an operation status of the device as operation information;
a status determination step of determining a status of the device using the operation information and the environmental information.
(Appendix 14)
an acquisition unit that acquires operation information of an operation status of a device to be monitored and acquires environmental information around the device;
a prediction model generation unit that sets explanatory variables including at least the environmental information in a predetermined time domain and objective variables including at least the operation information at a timing immediately after the predetermined time domain as a set of variables, and generates a prediction model of the operation status using a set of multiple variables defined by sequentially shifting time slots in the set of variables as training data; and
a prediction unit that predicts the operating status of the device by using the prediction model and the newly acquired environmental information.
(Appendix 15)
The response variable in the set of variables further includes the environmental information at a timing immediately after the predetermined time region,
the prediction model generation unit generates a prediction model of the operation status and the environmental information using the set of the plurality of variables as training data;
The prediction unit predicts an operating status of the device by using the prediction model and the newly acquired environmental information.
15. The prediction device of claim 14.
(Appendix 16)
a receiving unit that receives environmental information around a plurality of devices from a plurality of peripheral devices provided corresponding to each of the plurality of devices to be monitored;
an operation information acquisition unit that acquires operation statuses of the plurality of devices as operation information;
a storage unit that stores a first device and a second device that are in a similar environment among the plurality of devices in association with each other;
a state determination unit that determines whether or not the first device is abnormal by using the operation information and the environmental information;
and an instruction output unit that, when the state determination unit determines that the first device has an abnormality, executes at least one of outputting an instruction to the peripheral device provided corresponding to the second device to increase a frequency of transmitting the environmental information and outputting an instruction to the second device to stop processing.
Judgment device.
(Appendix 17)
an acquisition unit that acquires an operating status of a device to be monitored as operating information related to at least one of sound and vibration;
a power spectrum calculation unit that calculates a power spectrum of the operation information at three or more different timings;
a similarity calculation unit that calculates at least a similarity between the power spectrum at a first timing and the power spectrum at a second timing, and a similarity between the power spectrum at the first timing and the power spectrum at a third timing;
a determination unit that determines an operating status of the device using the plurality of similarities calculated by the similarity calculation unit.
(Appendix 18)
the determination unit determines whether a graph in which the multiple similarities are plotted in chronological order approximates a straight line, and if the graph does not approximate a straight line, identifies a point of curvature in the graph, and estimates a time when a malfunction occurred in the device based on the point of curvature.
18. The determination device of claim 17.
(Appendix 19)
When the graph is approximately a straight line, the determination unit determines whether or not a slope of the entire graph is equal to or greater than a predetermined threshold, and when the slope of the graph is equal to or greater than a predetermined threshold, estimates that deterioration over time is occurring in the device.
19. The determination device of claim 18.
(Appendix 20)
The determination unit determines that no abnormality has occurred in the device when the slope of the graph is less than a predetermined threshold value.
20. The determination device of claim 19.
(Appendix 21)
an acquisition step of acquiring operation information of an operation status of a device to be monitored and acquiring environmental information around the device;
a prediction model generation step of setting explanatory variables including at least the environmental information in a predetermined time domain and a response variable including at least the operation information at a timing immediately after the predetermined time domain as a set of variables, and generating a prediction model of the operation status using a set of multiple variables defined by sequentially shifting time slots in the set of variables as training data;
a prediction step of predicting the operating status of the device by using the prediction model and the newly acquired environmental information;
A monitoring method performed by a prediction device.
(Appendix 22)
a receiving step of receiving environmental information around a plurality of devices from a plurality of peripheral devices provided corresponding to each of the plurality of devices to be monitored;
an operation information acquisition step of acquiring operation statuses of the plurality of devices as operation information;
a state determination step of determining whether or not the first device among the plurality of devices has an abnormality by using the operation information and the environmental information;
an instruction output step of executing at least one of: when it is determined that the first device has an abnormality, outputting an instruction to increase the frequency of transmitting the environmental information to the peripheral device provided corresponding to the second device stored in association with the second device being in a similar environment to the first device, and outputting an instruction to the second device to stop processing;
The monitoring method is performed by a determination device.
(Appendix 23)
an acquisition step of acquiring operation information related to at least one of sound and vibration, the operation status of the device to be monitored;
a power spectrum calculation step of calculating a power spectrum of the operation information at three or more different timings;
a similarity calculation step of calculating at least a similarity between the power spectrum at a first timing and the power spectrum at a second timing, and a similarity between the power spectrum at the first timing and the power spectrum at a third timing;
a determining step of determining an operating status of the device using the calculated similarities;
The monitoring method is performed by a determination device.
(Appendix 24)
an acquisition step of acquiring operation information of an operation status of a device to be monitored and acquiring environmental information around the device;
a prediction model generation step of setting explanatory variables including at least the environmental information in a predetermined time domain and a response variable including at least the operation information at a timing immediately after the predetermined time domain as a set of variables, and generating a prediction model of the operation status using a set of multiple variables defined by sequentially shifting time slots in the set of variables as training data;
a prediction step of predicting the operating status of the device by using the prediction model and the newly acquired environmental information;
A program for causing a prediction device to execute the above as a monitoring method.
(Appendix 25)
a receiving step of receiving environmental information around a plurality of devices from a plurality of peripheral devices provided corresponding to each of the plurality of devices to be monitored;
an operation information acquisition step of acquiring operation statuses of the plurality of devices as operation information;
a state determination step of determining whether or not the first device among the plurality of devices has an abnormality by using the operation information and the environmental information;
an instruction output step of executing at least one of: when it is determined that the first device has an abnormality, outputting an instruction to increase the frequency of transmitting the environmental information to the peripheral device provided corresponding to the second device stored in association with the second device being in a similar environment to the first device, and outputting an instruction to the second device to stop processing;
A program for causing a determination device to execute the above as a monitoring method.
(Appendix 26)
an acquisition step of acquiring operation information related to at least one of sound and vibration, the operation status of the device to be monitored;
a power spectrum calculation step of calculating a power spectrum of the operation information at three or more different timings;
a similarity calculation step of calculating at least a similarity between the power spectrum at a first timing and the power spectrum at a second timing, and a similarity between the power spectrum at the first timing and the power spectrum at a third timing;
a determining step of determining an operating status of the device using the calculated similarities;
A program for causing a determination device to execute the above as a monitoring method.

The present disclosure has been described above with reference to the embodiments, but the present disclosure is not limited to the above. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the disclosure.

10 Peripheral device 11 Power receiving unit 12 Environmental sensor 13 Transmission unit 14 Rechargeable battery 20 Monitoring server 21 Reception unit 22 Operation information acquisition unit 23 State determination unit 30 Prediction device 31 Acquisition unit 32 Prediction model generation unit 321 CNN model unit 321a Convolution layer 321b Fully connected layer 33 Prediction unit 40 Determination device 41 Reception unit 42 Operation information acquisition unit 43 Memory unit 44 State determination unit 45 Instruction output unit 50 Determination device 51 Acquisition unit 52 Power spectrum calculation unit 53 Similarity calculation unit 54 Determination unit 100 USB sensor 101 Connection unit 102 Controller 103 Non-volatile storage area 104 Illuminance sensor 105 Impact/vibration sensor 106 Temperature/humidity sensor 107 Wireless communication module 108 Rechargeable battery 109 Sound sensor 200 Monitoring server 201 Reception unit 202 Learning unit 203 Model storage unit 204 Inference unit 205 Notification units S1 and S2 Monitoring system 211 Reception unit 212 Power spectrum calculation unit 213 Similarity calculation unit 214 Determination unit 215 Notification unit

Claims (9)

a power receiving unit that receives power from a device to be monitored;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit;
A storage unit that stores the environmental information ,
the environmental sensor and the transmission unit are operated by discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit;
When the transmission unit is receiving power via the power receiving unit, the transmission unit transmits the environmental information stored in the storage unit at regular intervals, and when it becomes impossible to receive power via the power receiving unit, the transmission unit transmits the environmental information stored in the storage unit regardless of the regular intervals.
Peripheral devices. a power receiving unit that receives power from a device to be monitored;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit;
A storage unit that stores the environmental information;
a determination unit that determines whether or not a detection value of the environmental information detected by the environmental sensor is equal to or greater than a predetermined threshold value ,
the environmental sensor and the transmission unit are operated by discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit;
the transmission unit transmits the environmental information stored in the storage unit at regular intervals when the determination unit determines that the detection value is less than the predetermined threshold, and transmits the environmental information stored in the storage unit regardless of the regular intervals when the determination unit determines that the detection value is equal to or greater than the predetermined threshold.
Peripheral devices. a power receiving unit that receives power from a device to be monitored;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit,
the environmental sensor and the transmission unit are operated by discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit;
The transmission unit transmits the environmental information to a monitoring server, and when an instruction to increase the frequency of transmitting the environmental information is received from the monitoring server, increases the frequency of transmitting the environmental information .
Peripheral devices. The peripheral device is a USB peripheral.
The peripheral device according to any one of claims 1 to 3 . A peripheral device provided in correspondence with the device to be monitored;
a monitoring server that communicates with the peripheral device;
The peripheral device is
A power receiving unit that receives power from the device;
An environmental sensor for acquiring environmental information around the device;
A transmission unit that transmits the environmental information;
a rechargeable battery that is charged by receiving power from the device via the power receiving unit,
the environmental sensor and the transmission unit are operated by discharged power from the rechargeable battery when it becomes impossible to receive power via the power receiving unit;
The monitoring server includes:
a receiving unit that receives the environmental information from the peripheral device;
an operation information acquisition unit that acquires an operation status of the device as operation information;
a status determination unit that determines a status of the device using the operation information and the environmental information. The state determination unit is
a prediction model generation unit that sets explanatory variables including at least the environmental information in a predetermined time domain and objective variables including at least the operation information at a timing immediately after the predetermined time domain as a set of variables, and generates a prediction model of the operation status using a set of multiple variables defined by sequentially shifting time slots in the set of variables as training data; and
The monitoring system according to claim 5 , further comprising: a prediction unit that predicts the operating status by using the prediction model and the newly acquired environmental information. the monitoring system includes a plurality of the peripheral devices corresponding to the plurality of the monitored devices,
The monitoring server includes:
a storage unit that stores a first device and a second device that are in a similar environment among the plurality of devices in association with each other;
The monitoring system of claim 5, further comprising an instruction output unit that, when the status determination unit determines an abnormality in the first device, performs at least one of outputting an instruction to the peripheral device corresponding to the second device to increase the frequency at which the transmission unit transmits the environmental information, and outputting an instruction to the second device to stop processing. a charging step of charging a rechargeable battery by receiving power from the device to be monitored;
an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when power supply from the device becomes impossible;
a storage step of storing the environmental information;
a transmission step of transmitting the stored environmental information at regular intervals when the device is receiving power from the device, and transmitting the stored environmental information using discharged power from the rechargeable battery regardless of the regular intervals when the device is no longer receiving power from the device;
The peripherals perform the monitoring method. a charging step of charging a rechargeable battery by receiving power from the device to be monitored;
an environmental information acquisition step of acquiring environmental information around the device using discharged power from the rechargeable battery when power supply from the device becomes impossible;
a storage step of storing the environmental information;
a determination step of determining whether or not the detected value of the environmental information is equal to or greater than a predetermined threshold;
a transmission step of transmitting the stored environmental information at regular intervals when it is determined that the detection value is less than the predetermined threshold, and transmitting the stored environmental information regardless of the regular intervals when it is determined that the detection value is equal to or greater than the predetermined threshold;
A program that causes a peripheral device to execute the above as a monitoring method.

Images