JP2017191395A - Management apparatus and control method - Google Patents

Abstract

The present invention provides a technique capable of appropriately selecting a sensor for increasing a measurement frequency and a technique capable of appropriately determining a timing for increasing the measurement frequency. A relationship sensor set calculation unit 212 executes a predetermined correlation calculation process based on sensor measurement data to calculate a combination of sensors having a predetermined relationship. When the detection unit 216 determines that a predetermined detection condition indicating that the calculated change in the combination has reached a predetermined monitoring level is satisfied, the setting change sensor selection unit 218 determines that the change has been satisfied. A sensor whose measurement frequency is changed is selected based on a combination of sensors related to the detection condition. Then, the setting change control unit 220 performs control to change the setting so that the measurement frequency of the selected sensor becomes higher than when it is determined that the detection condition is not satisfied. [Selection] Figure 9

Description

  The present invention relates to a sensor network technique in which sensors are arranged in a distributed manner on a monitoring object and the state of the monitoring object is monitored based on measurement data of each sensor.

  A technique for monitoring the state of an object to be monitored using a wireless sensor network including a large number of sensors dispersedly installed on an object to be monitored such as a tunnel, a bridge, an overhead line, a building, the ground, and an inclined land (for example, Patent Document 1).

JP 2013-109431 A

  In the wireless sensor network, when the measurement result of a certain sensor indicates a sign of certain deformation, for example, the following countermeasure is desired. That is, the first countermeasure is to increase the measurement frequency by shortening the measurement time interval of the sensor that shows the sign of the deformation. The second countermeasure is to shorten the measurement time interval and increase the measurement frequency for sensors other than the sensor that has shown signs of deformation.

  Each of the countermeasures is a measure for acquiring detailed data with high temporal density by increasing the measurement frequency. The purpose of this is to make it possible to detect the occurrence of an event requiring attention or alert as soon as possible, or to increase the accuracy (accuracy) of detection.

  Therefore, as a simple method for specifically realizing the above-described countermeasure, it is considered that the measurement frequency of all the sensors may be increased when any signs of deformation appear.

  However, sensors used in wireless sensor networks are typically battery powered. Therefore, there is a demand to reduce the operation cost related to measurement as much as possible to suppress power consumption and reduce the maintenance cost required for battery replacement.

Therefore, a technique for appropriately selecting a sensor that increases the measurement frequency and a technique for appropriately determining when to increase the measurement frequency are required.
The present invention has been devised in view of the above circumstances.

A first invention for solving the above-described problem includes a plurality of measurement frequencies having a battery (for example, the battery 39 in FIG. 2) and a wireless communication unit (for example, the wireless communication unit 33 in FIG. 2) capable of setting and changing the measurement frequency. Sensors (for example, the sensor S in FIG. 1) are arranged in a distributed manner on a monitoring object (for example, the monitoring object 10 in FIG. 1), and a management device (for example, the management device 20 in FIG. 1) based on the measurement data of each sensor. ) Is the management device of the sensor network for monitoring the state of the monitoring object,
A calculation unit (for example, the relationship sensor set calculation unit 212 in FIG. 9) that executes a predetermined correlation calculation process based on the measurement data and calculates a combination of the sensors having a predetermined relationship;
Change determination means (for example, the detection unit 216 in FIG. 9) for determining whether a predetermined detection condition indicating that the change in the combination has reached a predetermined monitoring level is satisfied;
A selection unit (for example, a setting change sensor selection unit 218 in FIG. 9) that selects a sensor that changes the measurement frequency based on the combination when it is determined that the change determination unit satisfies the condition;
A measurement frequency setting control means (for example, FIG. 9) that performs control to change the setting so that the measurement frequency of the sensor selected by the selection means becomes higher than when it is determined that the change determination means does not satisfy it. Setting change control unit 220),
It is the management apparatus provided with.

Further, as another invention, a plurality of sensors having a battery and a wireless communication unit capable of setting and changing the measurement frequency are distributed and arranged on the monitoring target, and the management device determines whether the monitoring target is based on the measurement data of each sensor. A sensor network control method for monitoring a state,
Executing a predetermined correlation calculation process based on the measurement data to calculate a combination of the sensors having a predetermined relationship (for example, steps T5 to T7 in FIG. 3);
Determining whether a predetermined detection condition indicating that the change in the combination has reached a predetermined monitoring level is satisfied (for example, step T9 in FIG. 3);
Selecting a sensor that changes the measurement frequency based on the combination when it is determined that the detection condition is satisfied (for example, step T13 in FIG. 3);
Performing control to change the setting of the measurement frequency of the selected sensor so as to be higher than when it is determined that the detection condition is not satisfied (for example, step T15);
It is good also as comprising the control method containing these.

  According to the first aspect of the invention, a combination of sensors having a predetermined relationship based on measurement data is calculated, and it is detected that a change in the combination has reached a predetermined monitoring level by satisfying a predetermined detection condition. In the case where the sensor frequency is changed, a sensor for changing the measurement frequency is selected based on the combination, and the setting is changed so that the measurement frequency of the sensor becomes high. In other words, if the change in the combination of sensors having a predetermined relationship based on the measurement data is a change that satisfies a predetermined detection condition, some change or a sign of a change occurs in the monitored object. Therefore, it is possible to appropriately detect this moment. And since the sensor which changes measurement frequency is selected based on the combination changed so that detection conditions may be satisfied, the sensor which changes measurement frequency can be selected appropriately.

The second invention is
The calculation means calculates the correlation index value between the sensors by executing the correlation calculation process, determines the presence or absence of the relationship between the sensors based on the correlation index value,
The measurement frequency setting means determines a measurement frequency to be changed based on the correlation index value of the sensor selected by the selection means.
It is the management apparatus of 1st invention.

  According to the second aspect of the invention, the measurement frequency for setting change can be determined based on the correlation index value used for determining the presence or absence of the relationship between the sensors. Accordingly, the measurement frequency can be set and changed by associating the relationship between the sensors and the measurement frequency, and thus the setting can be changed to a more appropriate measurement frequency.

The third invention is
The calculation means calculates a Mahalanobis distance as the correlation index value in the correlation calculation process, and determines that the relationship between the sensors satisfying a predetermined long threshold condition indicating that the Mahalanobis distance is long ,
It is the management apparatus of 2nd invention.

  According to the third aspect of the invention, the Mahalanobis distance can be used as the correlation index value.

The fourth invention is:
The calculating means determines that the correlation exists between the sensors in which the correlation index value calculated by the correlation calculation process satisfies a predetermined dependency relationship;
It is the management apparatus of 2nd invention.

  According to the fourth aspect of the present invention, when the correlation index value satisfies a predetermined dependency condition, it can be determined that there is a relationship between the sensors, and the sensors are determined to have a relationship. It becomes like this. Therefore, the measurement frequency can be increased when a change in the dependency relationship that satisfies the detection condition occurs.

The fifth invention is:
The change determination means determines whether the detection condition for each of the plurality of monitoring levels is satisfied;
The measurement frequency setting unit changes the setting so as to increase the measurement frequency as the monitoring level related to the detection condition determined to be satisfied by the change determination unit is higher (for example, the measurement frequency setting data 370 in FIG. 8). ),
The management apparatus according to any one of the first to fourth aspects.

  According to the fifth aspect, the measurement frequency can be increased as the detection condition of a higher monitoring level is satisfied.

The sixth invention is:
A plurality of the detection conditions are determined for each of the abnormal events to be monitored and / or for each of the predetermined reference sensors among the plurality of sensors (for example, the first detection condition data 361 in FIG. 6). , Second detection condition data 362) in FIG.
The change determination means determines whether each of the detection conditions is satisfied,
The selection means selects a sensor for changing the measurement frequency based on the combination relating to the detection condition determined to be satisfied by the change determination means.
The management device according to any one of the first to fifth inventions.

  According to the sixth aspect of the invention, the detection condition is determined for each of the abnormal events to be monitored and / or for each reference sensor determined in advance from a plurality of sensors, and satisfies any detection condition. In this case, a sensor that changes the measurement frequency is selected based on the combination of sensors according to the satisfied detection condition. Therefore, it is possible to more appropriately select a sensor that changes the measurement frequency.

More specifically, as a seventh invention,
A plurality of the detection conditions are determined for at least the reference sensor,
Of the combinations related to the detection condition determined to be satisfied by the change determination means, the selection means is determined to have the relationship between the reference sensor related to the detection condition and the reference sensor. Selected as a sensor for changing the measurement frequency,
It is good also as comprising the management apparatus of 6th invention.

The eighth invention
The detection condition is determined based on a time series change pattern indicating a time series change progress of the combination that can be taken when the monitoring level is reached,
The change determination means determines whether the detection condition is satisfied by comparing the time series progress of the sensor combination calculated by the calculation means with the time series change pattern;
The management apparatus according to any one of the first to seventh inventions.

  According to the eighth aspect of the present invention, since it is determined whether the detection condition is satisfied by the time series of the combination of sensors having a predetermined relationship, it is possible to more accurately determine the timing for increasing the measurement frequency. Is possible.

The figure which shows an example of a structure of a sensor network system. The figure for demonstrating the structure of a sensor. The flowchart for demonstrating the flow of a sensor network control process. The figure for demonstrating an example of the determination method of a relationship. The figure for demonstrating the time series change of a related sensor group. The figure which shows the example of 1st detection condition data. The figure which shows the example of 2nd detection condition data. The figure which shows the example of measurement frequency setting data. The figure which shows an example of a structure of a management apparatus.

  FIG. 1 is a diagram illustrating an example of a configuration of a sensor network system 1 according to the present embodiment. FIG. 2 is a diagram illustrating a functional configuration of the sensor S used in the sensor network system 1. The sensor network system 1 is configured to include a sensor network including a large number of sensors S distributed in various places on the monitoring target 10, a relay device 15, and a management device 20. Each sensor S and the relay device 15 are connected by wireless communication, and the relay device 15 and the management device 20 are connected by a network N. The network N means a communication path capable of data communication. In addition to a LAN (Local Area Network) such as a dedicated line (dedicated cable) or Ethernet (registered trademark), a communication network such as a telephone communication network, a cable network, or the Internet. In addition, the communication method may be wired or wireless.

  The monitoring object 10 may be any object that is supposed to be monitored by the sensor network, such as a tunnel, a bridge, an elevated line, a building, the ground, or an inclined land.

  The sensor S includes a plurality of types of sensors such as a temperature sensor, an illuminance sensor, a strain sensor, a vibration sensor, an acceleration sensor, and a gyro sensor. Each sensor S measures at a measurement time interval corresponding to the measurement frequency set in the measurement frequency setting value 38 and transmits measurement data to the relay device 15. The relay device 15 transmits the measurement data transmitted from each sensor S to the management device 20 as needed. In the present embodiment, the relay device 15 is configured to immediately transmit the measurement data transmitted from each sensor S to the management device 20, but is higher than the highest measurement frequency (shortest measurement time interval) of each sensor S. As transmission at a high frequency (short time interval), measurement data may be accumulated and transmitted to the management apparatus 20 during that time.

  As shown in FIG. 2, the sensor S includes a measurement unit 31, a wireless communication unit 33, a timer 35, a memory 37, and a battery 39. The measurement unit 31 is a main functional unit of the sensor S, and is a device that measures a physical quantity to be measured such as an acceleration if the sensor is an acceleration sensor or a change amount thereof. The timer 35 is a time measuring unit that measures a measurement time interval corresponding to the measurement frequency set in the measurement frequency setting value 38 stored in the memory 37. The battery 39 is a secondary battery, a dry battery or the like for supplying the operating power of the sensor S. Before the stored electric power of the battery 39 runs out, the sensor S is replaced or the battery 39 is replaced.

  The management apparatus 20 is realized by a computer, and the system configuration of the computer may be configured by a single computer apparatus, or may be configured by connecting a plurality of computer apparatuses via a communication network. . The management device 20 may be configured to perform wireless communication with each sensor S without using the relay device 15. In that case, the relay device 15 becomes unnecessary.

  Based on the measurement data of each sensor S, the management device 20 detects the occurrence of a change in the monitoring target 10 or a sign of the change and issues an alarm or a warning. Further, control for changing the setting of the measurement frequency of each sensor S is performed. Specifically, a predetermined correlation calculation process based on measurement data of each sensor S is executed to calculate a combination of sensors S having a predetermined relationship, and the change in the calculated combination reaches a predetermined monitoring level. When a predetermined detection condition indicating that the measurement has been performed is satisfied, a sensor S that changes the measurement frequency is selected based on the combination. Then, the measurement frequency of the selected sensor S is changed so as to be higher than when it is determined that the detection condition is not satisfied.

Processing performed by the management device 20 will be described in more detail. FIG. 3 is a flowchart for explaining the flow of a sensor network control process that is a process performed by the management apparatus 20.
First, the management apparatus 20 starts acquisition of measurement data from each sensor S as an initial operation (step T1). Then, in accordance with the inspection frequency for monitoring (which can also be referred to as an inspection cycle), the processing of steps T5 to T27 is performed every time the inspection timing comes (step T3: YES). The inspection timing is set according to the highest measurement frequency (shortest measurement time interval) set for each sensor S (steps T17 and T27), and can be set to be equivalent to the highest measurement frequency, for example. .

  When the inspection timing arrives (step T3: YES), the management device 20 executes a correlation calculation process for calculating a correlation index value between the sensors S (step T5), and uses the calculated correlation index value. A relational sensor set that is a combination of sensors S having a predetermined relation is determined (step T7).

  Two specific methods for the correlation calculation processing and the determination of the relationship sensor set will be described. The first method is a method using the MT (Mahalanobis Taguchi) method, and the second method is a method based on dependency relations.

The first method will be described. FIG. 4 is a diagram for explaining the determination of the relationship between the sensors S based on the first method. FIG. 4 shows an example in which two sensors S1 and S2 are determined. In the first method, the relationship between the measurement data of the two sensors S is determined using an MT (Mahalanobis Taguchi) method in which the measurement data of each sensor S in the past predetermined period is a multivariable. Since the MT method itself is publicly known, detailed description thereof is omitted, but specific application contents in applying the present embodiment to the first method will be described. First, in FIG. 4, the time of the determination target in the inspection timing measurement time t _0. Each measurement time _t -1 in a predetermined period in the past going back from the determination target time _{t 0, t -2,} the sensor S1 at · ·, S2 each measurement data _{D 1-1, D 1-2, ··,} D _2-1 , D _2-2 ,... Are unit spaces, and the measurement data D _1-0 and D _2-0 of the sensors S 1 and S 2 at time t ₀ are the determination target samples. Calculate the Mahalanobis distance between. If the Mahalanobis distance is less than or equal to a predetermined threshold, it is determined that there is no relationship, and if it exceeds the threshold, it is determined that there is a relationship. In other words, “there is a relationship” as used herein indicates a state in which the state between the sensors S1 and S2 exceeds the relationship of the normal state (normal state) (for example, a deformed state or a sign of a deformed state). State). In other words, “no relationship” means that the state between the sensors S1 and S2 is in a state corresponding to a normal state (normal state).

  In the first method, the Mahalanobis distance corresponds to a correlation index value, and the calculation of the Mahalanobis distance corresponds to a correlation calculation process. Then, a related sensor set is determined based on a combination between the sensors S determined to have a relationship.

  For example, as shown in FIG. 5, a related sensor set for each measurement time t can be calculated from the collected measurement data. In FIG. 5, at time t1, it is determined that there is a relationship between the sensors S1 and S2, between S1 and S5, and between S2 and S5. A combination between sensors determined to have a relationship at this time is defined as a relationship sensor set α. The same relational sensor set α is determined at times t2 and t3, but at time t4, the combination between the sensors determined to have a relationship changes to change between sensors S1 and S2, between S1 and S5, and S1. Between S3 and S3, and between S3 and S5. A combination between sensors determined to have a relationship at this time is defined as a relationship sensor set γ. In this way, the related sensor set is determined every time the inspection timing arrives.

  Moreover, the numerical value described between the sensors of FIG. 5 has shown an example of the normalized Mahalanobis distance. Here, the threshold value of the Mahalanobis distance for which there is a relationship is set to 0.6, and the relationship is determined to be relevant when the Mahalanobis distance is longer than this threshold value.

  Next, the second method will be described. In the second method, the multivariate autoregressive model (VAR) is applied to the measurement data between the two sensors S, and the dependence between the two sensors S is based on the deviation between the predicted value obtained by VAR and the actual measurement value. Determine the relationship index value. This dependency index value corresponds to the correlation index value.

A case where the dependency relationship index value between the sensor Sa and the sensor Sb is calculated will be described as an example. First, a time series model for calculating the predicted value a _t measured by the sensor Sa is constructed. Specifically, using the predicted value a ^ _t as an objective variable, past measurement data {a _t-1 , a _t-2 ,..., A _tk } of the sensor Sa and past measurement data {b of the sensor Sb _t-1 , b _t-2 ,..., b _t−k } are explanatory variables, and constants β (β _a1 , β _a2 ,..., β _ak , β _b1 , β _b2 , ..., β _bk ) is determined.
a ^ _t = β ₀
+ Β _a1 × a _t−1 + β _a2 × a _t−2 +...
+ Β _b1 × b _t−1 + β _b2 × b _t−2 +... Formula (1)

Then, an average value a _m of the actual measurement value measured by the sensors Sa to the time t~ past predetermined period of time is an inspection timing (measurement data), the coefficient of determination of the time-series model obtained by the following formula (2) calculate.
Determining coefficient = Σ _t (a _t −a _t ) ² / Σ _t (a _t −a _m ) ² Equation (2)

Then, the calculated determination coefficient value is set as a dependency index value between the sensor Sa and the sensor Sb. The dependency index value can also be said to be a value indicating the strength of the correlation between the sensors S.
In the second method, the dependency index value (determination coefficient value) corresponds to the correlation index value, and the calculation of the dependency index value corresponds to the correlation calculation process. Then, the relationship sensor set is determined by determining that there is a relationship when the dependency index value is equal to or greater than a predetermined reference value.

  The relational sensor set determined by the second method can also be shown, for example, as shown in FIG. The numerical value between the sensors S shown in FIG. 5 in this case is a dependency index value (its normalized value). A combination between the sensors S whose dependency index value is equal to or greater than a predetermined reference value is determined as a related sensor set at each time (each inspection timing).

  Returning to FIG. 3, after completing the determination of the related sensor set at the inspection timing (step T <b> 7), the management device 20 determines a predetermined detection condition indicating that the change of the related sensor set has reached a predetermined monitoring level. Is satisfied (step T9). Examples of detection condition data used for this determination are shown in FIGS.

  FIG. 6 is a diagram illustrating an example of the first detection condition data 361 (detection condition data 360). The first detection condition data 361 includes, for each abnormal event that is assumed to occur in the monitoring target 10, a reference sensor for the abnormal event and a time-series change pattern of a related sensor set corresponding to each monitoring level. Is stipulated. The deformation event is defined, for example, for each position of the monitoring target 10 and a state change (for example, crack, distortion, displacement, etc.) that may occur at the position. Each deformation event is associated with a sensor S that plays the main role of detecting occurrence of the deformation or a sign of the deformation as a reference sensor. In the example of FIG. 6, “S1” is associated with the deformation event “a” as a reference sensor.

  In this embodiment, there are three monitoring levels of levels 1 to 3, but may be 4 or more, or 1 or 2. In the first detection condition data 361, a plurality of time-series change patterns of related sensor sets are defined as detection conditions for each monitoring level for the abnormal event. For example, in FIG. 6, when the time series change of the related sensor set is α → α → α → γ, it indicates that the detection condition of the monitoring level 1 of the abnormal event “a” is satisfied. . Since the time series change pattern of the related sensor set shown in FIG. 5 matches this example, if the determined time series change of the related sensor set is as shown in FIG. 5, the change event “a” It corresponds to the monitoring level 1. Further, in the time series change pattern of the relational sensor set shown in FIG. 5, when the relational sensor set γ is determined at the next time, it matches the time series change pattern of the monitoring level 2 in FIG. In this case, it corresponds to the monitoring level 2 of the deformation event “a”.

  As shown in FIG. 6, a plurality of patterns are defined in the time series change pattern of the related sensor set even for the same abnormal event. If any of the time-series change patterns is matched, it corresponds to the corresponding monitoring level of the corresponding abnormal event.

  FIG. 7 is a diagram illustrating an example of the second detection condition data 362 (detection condition data 360). Similar to the first detection condition data 361 shown in FIG. 6, a reference sensor for the abnormal event is determined for each abnormal event, and a detection condition is determined for each monitoring level. However, the detection condition of the second detection condition data 362 is a condition based on a change from the related sensor set at the previous test timing to the related sensor set at the current test timing, and is a condition related to the reference sensor.

  Specifically, in the relational sensor group, the number between the sensors S determined to be related with the reference sensor as a base point is set as the order, the first determination focusing on the change amount of the order, and the reference sensor as the base point A virtual vector whose component is a correlation index value between the sensors S determined to have a relationship is obtained, and an OR condition or an AND condition with the second determination focusing on the change of the virtual vector is used as a detection condition.

  For example, time t4 in FIG. 5 is the current inspection timing, time t3 is the previous inspection timing, and it is determined whether the detection condition is satisfied by using the reference sensor as the sensor S1. In the relational sensor set α at time t3, there are two sensors S1 and S2 and S1 and S5, and the order at this time is “2”. is there. In the relational sensor set γ at time t4, the number of the sensors S determined to be related with the sensor S1 as the base point is between S1 and S2, between S1 and S3, and between S1 and S5. Is “3”. Therefore, in the first determination, the change amount of the order is “1”.

  Further, the correlation index value between the sensors S determined to have a relationship with the sensor S1 as the base point in the relationship sensor set α at time t3 is “0.7” between S1 and S2, and “0” between S1 and S5. .7 ". Similarly, the correlation index value between the sensors S determined to have a relationship with the sensor S1 as the base point in the relationship sensor set γ at time t4 is “0.9” between S1 and S2, and “ 0.6 ”and“ 0.8 ”between S1 and S5. Therefore, a virtual vector whose components are three correlation index values between S1 and S2, between S1 and S3, and between S1 and S5 is obtained from the relation sensor set α at time t3 and the relation sensor set γ at time t4. Thus, the similarity between the virtual vectors is calculated. The second determination is to use this similarity as a change in the virtual vector.

  In the example of FIG. 7, for example, as the detection condition of the monitoring level 1 with the reference sensor “S1” for the deformation event “a”, the degree change amount is “1”, or the virtual vector similarity is M1 or more M2 The case of less than is stipulated. If the time t4 in FIG. 5 is the current inspection timing, the order change amount is “1”, and thus this detection condition is satisfied.

  Two detection conditions, the detection condition based on the first detection condition data 361 and the detection condition based on the second detection condition data 362, have been described as detection conditions. Either of them may be used. Further, when both are used and one of the detection conditions is met, it may be determined that the detection condition is met.

  Returning to FIG. After determining whether or not each detection condition is satisfied, if the management apparatus 20 determines that any of the detection conditions is satisfied (step T9: YES), the abnormal event related to the detection condition becomes the detection condition. Since it indicates that the monitoring level has been reached, a process for notifying that effect is performed (step T11). A notification process such as outputting a predetermined voice as an alarm or warning, displaying a predetermined display, or transmitting a predetermined notification to the outside is performed.

  Next, the sensor S for setting and changing the measurement frequency is selected (step T13). Specifically, a sensor having a relationship with the reference sensor according to the detection condition determined to be satisfied in step T11 is selected as a measurement frequency change target.

  For example, in FIG. 5, it is assumed that the current examination timing is time t4 and it is determined that the detection level 1 detection condition of the change event “a” in the first detection condition data 361 shown in FIG. In this case, since the reference sensor related to the detection condition is the sensor S1, first, the sensor S1 is selected as a measurement frequency change target. In addition, in the related sensor set α at time t3 immediately before it is determined that the detection condition is satisfied, the sensors (sensors S2 and S5) having a relationship with the sensor S1 that is the reference sensor are selected as the measurement frequency change targets. .

  In addition, from the relational sensor set at the time immediately before it was determined that the detection condition was satisfied, it was decided to select the sensor to be changed in measurement frequency, but from the relational sensor set at the time determined to satisfy the detection condition, It is good also as selecting the sensor made into the change object of measurement frequency. In the above example, in the related sensor set γ at time t4, sensors (sensors S2, S3, S5) having a relationship with the sensor S1 that is the reference sensor are selected as the measurement frequency change targets. Of course, even in this case, the reference sensor is selected as a measurement frequency change target.

  In addition, a sensor that is a measurement frequency change target may be selected from both the related sensor set at the time immediately before it is determined that the detection condition is satisfied and the related sensor set at the time determined to be satisfied.

  After selecting the sensor (step T13), the management device 20 performs control to change the setting of the measurement frequency of the selected sensor S (step T15). For example, based on the measurement frequency setting data 370 as shown in FIG. 8, control is performed to set the measurement frequency corresponding to the monitoring level related to the detection condition determined to be satisfied in step T9 to the sensor S selected in step T13. This is realized by transmitting an instruction signal to change the setting of the measurement frequency setting value 38 (see FIG. 2) to each selected sensor S via the relay device 15.

Note that the measurement frequency for changing the setting may be any frequency that increases (the measurement time interval becomes shorter) than when it is determined that the detection condition is not satisfied.
For example, the measurement frequency for changing the setting may be determined using a correlation index value between the reference sensor and the selected sensor S. As a specific example, in the relationship sensor set when the sensor S is selected as the measurement frequency setting change target, the correlation index value between the sensor S and the reference sensor is “0.6”. To do. For the reference sensor, since the measurement frequency determined with reference to FIG. 8, for example, the detection condition related to the monitoring level 2 is satisfied, the measurement frequency is set to “once / hour” from the normal time. It is assumed that the frequency of “24 times” is set. In this case, the frequency of “14.4 times” obtained by multiplying the correlation index value “0.6” by the measurement frequency “24 times” set for the reference sensor is set as the measurement frequency of the sensor S. I decided to.

  Returning to FIG. 3, after the measurement frequency setting change control is performed (step T15), the management apparatus 20 changes the setting of the inspection frequency according to the changed measurement frequency (step T17). For example, when the highest measurement frequency is updated among the sensors S, the setting is changed to the inspection frequency corresponding to the measurement frequency. As a result, the inspection interval to be monitored becomes an interval according to the measurement frequency of the sensor S, and when the measurement frequency increases, the inspection interval is shortened accordingly, and the occurrence of a change or a sign of a change is detected quickly. It becomes possible. Further, since the measurement frequency is changed according to the monitoring level, the detection accuracy (accuracy) is improved.

  If it is determined in step T9 in FIG. 3 that the detection condition is not satisfied (step T9: NO), the management device 20 determines whether to restore (for example, reset) the measurement frequency (step S9). T21). For example, when a state where the detection condition is not satisfied continues for a predetermined time, it can be determined that the measurement frequency is restored.

  If it is determined that the measurement frequency is to be restored (step T21: YES), the sensor to be restored is selected (step T23), and the setting change control for restoring the measurement frequency of the sensor is performed (step T25). Is further changed (step T27), and the process proceeds to step T3. The selection of the sensor S to be restored (step T23) may be performed by selecting the sensor S that has been previously changed in setting to increase the measurement frequency.

When it is determined that the measurement frequency is not restored (step T21: NO), the process proceeds to step T3.
After the processing is subsequently performed to step T3, the management device 20 repeatedly executes the processing of steps T3 to T27.

Next, the configuration of the management apparatus 20 will be described.
FIG. 9 is a diagram for explaining the configuration of the management apparatus 20. The management device 20 is a computer that includes an operation unit 410, a display unit 420, a sound output unit 430, a communication unit 440, a processing unit 200, and a storage unit 300. As described above, it may be configured as a single device, or may be configured by communication connection of a plurality of devices.

  The operation unit 410 is realized by an input device such as a keyboard, a mouse, a touch panel, and various switches, for example, and outputs an operation signal corresponding to the operation input made to the processing unit 200. The display unit 420 is realized by a display device such as a liquid crystal display or an organic EL display, and performs various displays based on display signals from the processing unit 200. The sound output unit 430 is realized by a sound output device such as a speaker, for example, and performs various audio outputs based on the audio signal from the processing unit 200. The communication unit 440 is a communication device realized by, for example, a wireless communication module, a router, a modem, a wired communication cable jack, a control circuit, and the like, and data with external devices (mainly the relay device 15 and the sensor S). Communicate.

  The processing unit 200 is realized by an arithmetic device or arithmetic circuit such as a CPU (Central Processing Unit) or FPGA (Field Programmable Gate Array), and from a program or data stored in the storage unit 300 or an external device via the communication unit 440. The overall control of the management device 20 and the overall control of the sensor network system 1 are performed based on the received data.

  The processing unit 200 includes, as functional processing blocks, a relationship sensor set calculation unit 212, a detection unit 216, a setting change sensor selection unit 218, a setting change control unit 220, an inspection frequency setting unit 222, and a notification processing unit. 224. These functional units may be realized by software by the processing unit 200 executing a program, or may be realized by a dedicated arithmetic circuit. In the present embodiment, the former software is used.

  The relationship sensor set calculation unit 212 executes a predetermined correlation calculation process based on the measurement data of each sensor S, and calculates a combination of the sensors S having a predetermined relationship as a relationship sensor set. In addition, the correlation index value calculation unit 214 that calculates the correlation index value is used for calculating the correlation. The correlation calculation process and the determination of the relationship sensor set may employ either the first method or the second method described above.

  The detection unit 216 determines whether or not a predetermined detection condition indicating that the change of the related sensor set has reached a predetermined monitoring level is satisfied. As the detection condition, the first detection condition data 361 shown in FIG. 6 may be used, or the second detection condition data 362 shown in FIG. 7 may be used.

  The setting change sensor selection unit 218 selects the sensor S that changes the measurement frequency based on the related sensor set related to the detection condition determined to be satisfied by the detection unit 216.

  The setting change control unit 220 performs control to change the setting so that the measurement frequency of the sensor S selected by the setting change sensor selection unit 218 becomes higher than when it is determined that the detection condition is not satisfied. For example, the setting change control unit 220 may determine the measurement frequency for changing the setting based on the correlation index value of the selected sensor S. The setting change control unit 220 performs control to change the setting so that the measurement frequency is increased as the monitoring level related to the detection condition determined to be satisfied by the detection unit 216 is higher. Since the setting change of the measurement frequency is realized by updating the measurement frequency setting value 38 (see FIG. 2) of the sensor S, the setting change control unit 220 changes the sensor S selected by the setting change sensor selection unit 218. The instruction signal is transmitted so as to update and store the measurement frequency setting value 38 corresponding to the determined measurement frequency.

  The inspection frequency setting unit 222 sets the frequency (time interval) for performing inspection for monitoring when monitoring the monitoring target 10 based on the measurement data of each sensor S. It is preferable to set the frequency according to the highest measurement frequency among the set sensor S measurement frequencies.

  When the detection unit 216 determines that the detection condition is satisfied, the notification processing unit 224 performs processing for reporting the abnormality event and the monitoring level related to the detection condition to the outside of the apparatus. A sound for reporting the abnormal event and the monitoring level is output from the sound output unit 430, the abnormal event and the monitoring level are displayed on the display unit 420, and a signal for reporting the abnormal event and the monitoring level is externally provided. Or output to the device.

  The storage unit 300 is realized by an IC memory such as a ROM or a RAM, or a storage device such as a hard disk. The storage unit 200 stores a program, data, or the like for the processing unit 200 to control the management device 20 in an integrated manner. It is used as a work area of the unit 200 and temporarily stores the calculation results executed by the processing unit 200, data received from the communication unit 440, and the like.

  In the storage unit 300, the sensor network control program 310, the sensor DB (database) 320, the measurement data group 330, the inter-sensor correlation index value data 340, the related sensor group information 350, and the detection condition data 360 are stored. And measurement frequency setting data 370 are stored.

  The sensor network control program 310 is a program for realizing the sensor network control process illustrated in FIG. 3, and the processing unit 200 reads and executes the sensor network control program 310, whereby the related sensor set calculation unit 212, It functions as a detection unit 216, a setting change sensor selection unit 218, a setting change control unit 220, a setting change control unit 220, an inspection frequency setting unit 222, and a notification processing unit 224.

  The sensor DB 320 stores data regarding each sensor S. For example, information such as the installation position, the type of physical quantity to be measured, the set value of the measurement frequency currently set, the communication address of wireless communication, and the like for each sensor S is stored.

  The measurement data group 330 is data in which measurement data measured by each sensor S is stored in association with time for each sensor S.

  The inter-sensor correlation index value data 340 is data in which correlation index values between sensors calculated by the correlation index value calculation unit 214 are stored in association with the calculated time for each sensor combination. is there.

  The relationship sensor set information 350 is data in which the relationship sensor set information determined by the relationship sensor set calculation unit 212 is stored in association with the determined time.

  The detection condition data 360 stores the first detection condition data 361 shown in FIG. 6 and the second detection condition data 362 shown in FIG.

  The measurement frequency setting data 370 is data of setting values of the measurement frequency for each monitoring level shown in FIG. 8 as an example.

  With the above configuration, the management device 20 realizes the sensor network control process described above.

  According to this embodiment, a combination of sensors S having a predetermined relationship based on measurement data of each sensor S (relation sensor set) is calculated, and a predetermined monitoring is performed by a change in the combination satisfying a predetermined detection condition. When it is detected that the level has been reached, a sensor S that changes the measurement frequency is selected based on the combination, and the setting is changed so that the measurement frequency of the sensor S becomes high. That is, if the change in the combination of the sensors S having a predetermined relationship based on the measurement data is a change that satisfies a predetermined detection condition, it can be said that the monitoring object is somehow deformed or even deformed. Since it can be said that there is no sign, this timing can be detected appropriately. And since sensor S which changes measurement frequency is selected based on the combination changed so that detection conditions may be met, sensor S which changes measurement frequency can be selected appropriately.

  As mentioned above, although embodiment which applied this invention was described, the form which can apply this invention is not limited to the said form, A component can be added, abbreviate | omitted, and changed suitably.

1 Sensor network system 10 Object to be monitored S Sensor
33 Wireless communication unit
38 Measurement frequency setting section
39 Battery 20 Management Device
200 processor
212 Relational sensor set calculation unit
214 Correlation index value calculation unit
216 detector
218 Setting change sensor selector
220 Setting change control unit
222 Inspection frequency setting section
224 Notification processing unit

Claims (9)

A sensor network in which a plurality of sensors having a battery and a wireless communication unit capable of setting and changing a measurement frequency are distributed in a monitoring object, and a management device monitors a state of the monitoring object based on measurement data of each of the sensors The management device,
A calculation means for executing a predetermined correlation calculation process based on the measurement data to calculate a combination of the sensors having a predetermined relationship;
Change determining means for determining whether a predetermined detection condition indicating that the change in the combination has reached a predetermined monitoring level is satisfied;
A selection unit that selects a sensor that changes the measurement frequency based on the combination when it is determined that the change determination unit satisfies the condition;
A measurement frequency setting control means for performing control to change the setting so that the measurement frequency of the sensor selected by the selection means is higher than when it is determined not to be satisfied by the change determination means;
Management device with. The calculation means calculates the correlation index value between the sensors by executing the correlation calculation process, determines the presence or absence of the relationship between the sensors based on the correlation index value,
The measurement frequency setting means determines the measurement frequency to change the setting based on the correlation index value of the sensor selected by the selection means.
The management apparatus according to claim 1. The calculation means calculates a Mahalanobis distance as the correlation index value in the correlation calculation process, and determines that the relationship between the sensors satisfying a predetermined long threshold value condition indicating that the Mahalanobis distance is long is present. ,
The management device according to claim 2. The calculation means determines that the correlation exists between the sensors in which the correlation index value calculated by the correlation calculation process satisfies a predetermined dependency relationship;
The management device according to claim 2. The change determination unit determines whether the detection condition for each of the plurality of monitoring levels is satisfied,
The measurement frequency setting means changes the setting so as to increase the measurement frequency as the monitoring level related to the detection condition determined to be satisfied by the change determination means is higher.
The management apparatus as described in any one of Claims 1-4. A plurality of the detection conditions are determined for each abnormal event to be monitored and / or for each reference sensor determined in advance among the plurality of sensors,
The change determination means determines whether each of the detection conditions is satisfied,
The selection means selects a sensor that changes the measurement frequency based on the combination related to the detection condition determined to be satisfied by the change determination means.
The management apparatus as described in any one of Claims 1-5. A plurality of the detection conditions are determined at least for each of the reference sensors,
The selection unit is determined to have the relationship between the reference sensor according to the detection condition and the reference sensor among the combinations related to the detection condition determined to be satisfied by the change determination unit. Selected as a sensor for changing the measurement frequency,
The management device according to claim 6. The detection condition is determined based on a time-series change pattern indicating a time-series change process of the combination that can be taken when the monitoring level is reached,
The change determination means determines whether or not the detection condition is satisfied by comparing the time series progress of the combination of sensors calculated by the calculation means with the time series change pattern;
The management apparatus as described in any one of Claims 1-7. A sensor network in which a plurality of sensors having a battery and a wireless communication unit capable of setting and changing a measurement frequency are distributed in a monitoring object, and a management device monitors a state of the monitoring object based on measurement data of each of the sensors A control method,
Performing a predetermined correlation calculation process based on the measurement data to calculate a combination of the sensors having a predetermined relationship;
Determining whether a predetermined detection condition indicating that the change in the combination has reached a predetermined monitoring level is satisfied;
When it is determined that the detection condition is satisfied, selecting a sensor that changes the measurement frequency based on the combination;
Performing control to change the setting of the measurement frequency of the selected sensor so as to be higher than when it was determined that the detection condition is not satisfied;
Control method.

Images