JP2019029696A - Sensor control device, sensor system and sensor control method - Google Patents

Abstract

To reduce a power consumption while suppressing precision of data from lowering.SOLUTION: A sensor control device has: a radio communication part which receives measurement data on measurements that a plurality of sensors located within a range of communication with the device obtain; a regressive estimation part which performs regressive estimation using the measurement data received by the radio communication part; a calculation part which calculates independence levels indicative of levels of correlation of the measurement data with other sensors for each sensor from a regressive estimation result obtained by the regressive estimation part; a transmission part which transmits a request to accommodate one sensor according to independence levels calculated by the calculation part; and a setting part which sets, when a sensor is accommodated in another sensor control device as the transmission part sends the request to accommodate the sensor, measurement frequencies of the respective sensors based upon the regressive estimation result of measurement data of the plurality of sensors except the sensor.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a sensor control device, a sensor system, and a sensor control method.

  In recent years, an IoT (Internet of Things) service is known that collects measurement data related to the environment by a sensor system having sensor nodes that are distributed and provides information based on the analysis content of the measurement data. In order to enable measurement at an arbitrary place in a wide area, a battery as a main power source is usually attached to the sensor node. In the sensor system, it is desired to reduce the power consumption of each sensor node in order to reduce the maintenance load due to frequent battery replacement at the sensor node.

  Therefore, in order to reduce power consumption, for example, it is conceivable to reduce the frequency of measurement by the sensor node. However, if the frequency of measurement is simply reduced, there is a high possibility that a change in the state of the measurement target will be overlooked, and the accuracy of measurement data may deteriorate. For this reason, a method of predicting future data from the measurement data by the sensor node and controlling the operation interval of each sensor node according to the prediction error has been studied.

JP 2014-138245 A Japanese Patent Laid-Open No. 10-131973

  By the way, when the sensor system has a plurality of sensor nodes (hereinafter simply referred to as “nodes”), the measurement data of each node may have a correlation. Based on the correlation of the measurement data, the measurement data of some nodes can be estimated from the measurement data of other nodes. In this way, it is possible to reduce the frequency of measurement by redundant nodes by estimating the measurement data of redundant nodes based on the correlation of the measurement data, rather than predicting the measurement data of the nodes individually. This makes it possible to further reduce power consumption.

  That is, based on the correlation of the measurement data of a plurality of nodes, the power consumption of the entire sensor system is reduced by concentrating the measurement on the key node having low redundancy and estimating the measurement data of the redundant node. It becomes possible.

  A sensor system having a plurality of nodes may adopt a configuration in which the node 10 is accommodated in the gateway device 20 as shown in FIG. That is, the gateway device 20 accommodates the nodes 10 located in a range where wireless communication is possible, and collects measurement data from the nodes 10 accommodated by the respective gateway devices 20. The gateway device 20 transmits the collected measurement data to the server device 30 connected by wire.

  In the sensor system having such a configuration, since the measurement data collected by all the gateway devices 20 is aggregated in the server device 30, the redundancy calculation based on the correlation of the measurement data described above is performed by the server device 30. Executed.

  However, when the number of nodes 10 and gateway devices 20 is large, there is a problem that the amount of measurement data used for redundancy calculation increases and the processing load on the server device 30 increases.

  In order to reduce the processing load on the server device 30, it is conceivable that the gateway device 20 performs redundancy calculation. However, the node 10 accommodated in the gateway device 20 is located in a range where wireless communication with the gateway device 20 is possible. Only some of the nodes 10. For this reason, the correlation of the measurement data is not correctly calculated from the measurement data of the node 10 accommodated in one gateway device 20, and it is difficult to calculate the redundancy of the node 10 with high accuracy. As a result, the node 10 that reduces the frequency of measurement is not properly selected, and a certain limit is imposed on the reduction of power consumption.

  The disclosed technology has been made in view of the above points, and provides a sensor control device, a sensor system, and a sensor control method capable of reducing power consumption while suppressing a decrease in accuracy of data. Objective.

  In one aspect, a sensor control device disclosed in the present application is a wireless communication unit that receives measurement data obtained by measurement by a plurality of sensors located in a range where wireless communication with the own device is possible, and is received by the wireless communication unit. A regression estimation unit that performs regression estimation using the measured data, and a regression estimation result that is executed by the regression estimation unit. According to the degree of independence calculated by the calculation unit, a calculation unit that calculates, a transmission unit that transmits an accommodation request for requesting accommodation of any sensor to another sensor control device, and an accommodation request by the transmission unit When the sensor is accommodated by the other sensor control device as a result of the transmission, the measurement frequency of each sensor is determined based on the regression estimation result of the measurement data of a plurality of sensors excluding the sensor. And a setting unit for setting.

  According to one aspect of the sensor control device, the sensor system, and the sensor control method disclosed in the present application, there is an effect that it is possible to reduce power consumption while suppressing a decrease in data accuracy.

FIG. 1 is a diagram illustrating an example of a sensor system according to the first embodiment. FIG. 2 is a block diagram showing a configuration of the node according to the first embodiment. FIG. 3 is a block diagram illustrating a configuration of the gateway device according to the first embodiment. FIG. 4 is a diagram illustrating a specific example of the regression model table. FIG. 5 is a diagram for explaining the degree of independence. FIG. 6 is a block diagram showing a configuration of another gateway apparatus according to the first embodiment. FIG. 7 is a sequence diagram illustrating the operation of the sensor system according to the first embodiment. FIG. 8 is a flowchart showing the sensor control method according to the first embodiment. FIG. 9 is a flowchart illustrating the accommodation sensor determination method according to the first embodiment. FIG. 10 is a block diagram illustrating a configuration of the gateway device according to the second embodiment. FIG. 11 is a flowchart showing a neighbor GW determination method. FIG. 12 is a flowchart showing another neighbor GW determination method. FIG. 13 is a diagram illustrating an example of a sensor system.

  Hereinafter, embodiments of a sensor control device, a sensor system, and a sensor control method disclosed in the present application will be described in detail with reference to the drawings. In addition, this invention is not limited by this embodiment.

(Embodiment 1)
FIG. 1 is a diagram illustrating an example of a sensor system according to the first embodiment. The sensor system shown in FIG. 1 includes a plurality of sensor nodes (nodes) 10, a plurality of gateway devices 100 and 200, a measurement server device 30, an application server device 40, and a user terminal 50. It is assumed that the number of nodes and gateway devices included in the sensor system is arbitrary.

  The node 10 is a terminal device having a sensor function for measuring the state of the measurement target. The node 10 measures the state of an object to be measured in an outdoor environment or the inside of a structure, such as temperature, humidity, water level, wind direction, or human movement. Each node 10 is scattered for each measurement point in the space where the measurement is performed.

  Each of the gateway devices 100 and 200 accommodates the node 10 located in a range where wireless communication is possible, and collects measurement data obtained by the measurement of the accommodated node 10. Then, the gateway apparatuses 100 and 200 execute regression estimation using the collected measurement data, calculate the redundancy of each node 10, and set the measurement frequency of each node 10 according to the redundancy. That is, the power consumption of the entire sensor system is reduced by increasing the measurement frequency of the node 10 having low redundancy and decreasing the measurement frequency of the node 10 having high redundancy.

  The gateway devices 100 and 200 are connected to each other via the network N, and exchange the accommodated nodes 10 according to the correlation of measurement data by the nodes 10 in a range where wireless communication is possible. In the present embodiment, a case will be described in which the gateway device 100 accommodates the node 10 in a range in which the gateway device 100 can perform wireless communication.

  The measurement server device 30 is connected to all gateway devices including the gateway devices 100 and 200 via the network N, and collects sensor data from each gateway device. The collected sensor data includes measurement data acquired by each node 10 through measurement and estimated data estimated by the gateway device from the measurement data. Then, the measurement server device 30 transmits the collected sensor data to the application server device 40.

  The application server device 40 and the user terminal 50 are connected via a network N so that they can communicate with each other. It is assumed that the number of user terminals 50 connected to the application server device 40 is arbitrary. The application server device 40 performs statistical processing of the sensor data, and provides, for example, a graphic display of the statistical processing result to the user terminal 50.

  FIG. 2 is a block diagram illustrating a configuration of the node 10 according to the first embodiment. The node 10 illustrated in FIG. 2 includes a sensor 11, a battery 12, a processor 13, a communication interface (hereinafter abbreviated as “communication I / F”) 14, and a memory 15.

  The sensor 11 is an element that measures a state of a measurement target and outputs a measurement value that is a result of the measurement. The battery 12 is a power source of the node 10 and is, for example, a primary battery. Each part of the node 10 is driven using electric power from the battery 12. The battery 12 may include a secondary battery that stores electric power generated by various power generation elements in addition to the primary battery.

  The processor 13 includes, for example, a central processing unit (CPU), a field programmable gate array (FPGA), or a digital signal processor (DSP), and controls the entire node 10. The processor 13 acquires a measurement value obtained by the sensor 11 at a measurement timing corresponding to the setting, and generates measurement data including the acquired measurement value. In the present embodiment, a learning period and an operation period are provided, and in the learning period, a measurement value by the sensor 11 is acquired at a frequency common to all the nodes 10. On the other hand, during the operation period, the frequency of measurement is set according to the redundancy of the node 10, and the measurement value obtained by the sensor 11 is acquired at a different measurement timing for each node 10.

  The communication I / F 14 communicates with a communication destination outside the node 10. Specifically, the communication I / F 14 transmits, for example, measurement data generated by the processor 13 to a gateway device in a range where wireless communication with the own node is possible. Further, the communication I / F 14 receives measurement timing control information transmitted from, for example, a gateway device in a range where wireless communication with the own node is possible. The memory 15 includes, for example, a RAM (Random Access Memory) or a ROM (Read Only Memory), and stores various types of information when processing is executed by the processor 13.

  FIG. 3 is a block diagram showing a configuration of gateway apparatus 100 according to the first embodiment. The gateway device 100 is a gateway device that requests another gateway device (gateway device 200) to accommodate the node 10 in a range in which wireless communication can be performed from its own device. The gateway device 100 illustrated in FIG. 3 includes a wireless communication unit 100a, a communication I / F 100b, a processor 100c, and a memory 100d.

  The wireless communication unit 100a receives measurement data transmitted from the node 10 in a wireless communicable range, and performs predetermined wireless reception processing on the received measurement data.

  The communication I / F 100b is communicably connected to another gateway apparatus via the network N. For example, the communication I / F 100b transmits an accommodation request for the node 10 to the gateway apparatus 200, or information on availability of accommodation as a response to the accommodation request. Or receive.

  The processor 100c includes, for example, a CPU, FPGA, DSP, or the like, and controls the entire gateway device 100. Specifically, the processor 100c includes a measurement data acquisition unit 110, a regression estimation unit 120, an independence degree calculation unit 130, an accommodation request generation unit 140, an accommodation availability acquisition unit 150, a measurement necessity acquisition unit 160, a node setting unit 170, and A measurement data transmission unit 180 is included.

  The measurement data acquisition unit 110 acquires measurement data transmitted from the node 10 that wirelessly communicates with the gateway device 100 and received by the wireless communication unit 100a. In the present embodiment, since a learning period and an operation period are provided, the measurement data acquisition unit 110 performs measurement data at a frequency common to all the nodes 10 that wirelessly communicate with the gateway device 100 during the learning period. To get. Further, the measurement data acquisition unit 110 acquires measurement data at a different frequency for each node 10 during the operation period.

  The regression estimation unit 120 performs regression estimation using measurement data. That is, the regression estimation unit 120 generates a regression model for estimating the measurement data of one node 10 from the measurement data of another node 10 and calculates a regression coefficient in the regression model. Specifically, the regression estimation unit 120 obtains a regression model and a regression coefficient between the measurement data of each node 10 using the measurement data acquired during the learning period. The obtained regression coefficient is stored, for example, in a regression model table shown in FIG.

  The regression model table shown in FIG. 4 holds regression coefficients between measured values obtained by a plurality of sensors including sensors # 1 to # 4. For example, when the estimation sensor to be estimated is sensor # 1, the measurement value of sensor # 1 can be estimated from the measurement data of sensor # 1 and sensor # 4, which are measurement sensors, and the regression coefficient of each measurement data Is found to be 0.74 and 0.24. Therefore, the current measurement value of sensor # 1 can be estimated by performing an operation using these regression coefficients on the measurement data of sensor # 1 and sensor # 4 at a certain past time.

  Returning to FIG. 3, the regression estimation unit 120 uses the regression coefficient obtained in the learning period and stored in the regression model table, from the measurement data of some nodes 10 acquired in the operation period, to other nodes 10. The measured value at is estimated, and the obtained estimated data and sensor data including the measured data are output.

The independence calculating unit 130 calculates the independence of each node 10 based on the regression model obtained during the learning period. That is, the independence calculating unit 130 calculates the independence indicating the low correlation of the measurement data with the other nodes 10 for each node 10. In the regression model, a measurement value of a certain node 10 is referred to in order to estimate a measurement value of another node 10 or is estimated with reference to measurement data of another node 10. Therefore, for a certain node 10, the reference value indicating the number of nodes 10 to which the measurement data is referred in order to estimate the measurement value of this node 10, and the measurement value is estimated with reference to the measurement data of this node 10. The degree of independence is defined by the following equation (1) based on the number of references indicating the number of nodes 10.
Independence = 1 / (number of references x number of referenced) (1)
In addition, the presence or absence of the reference and the referenced relationship between the nodes 10 is determined by whether or not the regression coefficient corresponding to each node 10 satisfies a certain condition. For example, when LASSO regression is used as a regression model of a certain node A, a method is considered in which it is assumed that there is a reference / reference relationship between this node A and node B having a regression coefficient that is not 0 in the regression model. It is done.

By the above equation (1) is used, for example, As shown in FIG. 5, the sensor # 1 to # respectively from the measurement data x ₁ ~x ₅ of 5 sensors # 1 to # estimation data y ₁ ~y ₅ of 5 Is obtained, the degree of independence of the sensors # 1 to # 5 is calculated from the respective reference numbers and referenced numbers. As a specific example, the measurement data x ₁ sensor # 1 will be referred to for calculating the estimated data y ₁ ~y ₃ of the three sensors of sensor # 1 to # 3, the sensor # 1 The number of referenced is 3. On the other hand, since the estimated data y ₁ of the sensor # ₁ is calculated with reference to the measurement data x ₁ and x ₄ of the two sensors # 1 and # 4, the reference number of the sensor # 1 is 2. When these reference numbers and referenced numbers are applied to the above equation (1), the degree of independence of sensor # 1 can be calculated as 0.17. The degree of independence becomes larger as the number of sensors to be referred to is smaller and the number of sensors to be referred to is smaller, and therefore the degree of independence is larger as the sensor has a lower correlation with the measured values of other sensors. For example, in the example illustrated in FIG. 5, the sensor # 5 is not referred to by another sensor and does not refer to another sensor, and thus the degree of independence is 1.00 at the maximum.

  Returning to FIG. 3, the accommodation request generation unit 140 generates an accommodation request for requesting another gateway device to accommodate the node 10 having a high degree of independence calculated by the independence degree calculation unit 130. Specifically, the accommodation request generation unit 140 includes the identification information of the node 10 having the highest degree of independence among the nodes 10 wirelessly communicating with the gateway device 100 and the measurement data acquired from the node 10 during the learning period. Generate a request. The accommodation request generation unit 140 may generate an accommodation request not only for the node 10 having the highest degree of independence but also for a predetermined number of nodes 10 in descending order of independence. An accommodation request for 10 may be generated.

  Then, the accommodation request generating unit 140 transmits the generated accommodation request to another predetermined gateway device via the communication I / F 100b. That is, since other gateway devices including the gateway device 200 are predetermined as the transmission destination of the accommodation request by the gateway device 100, the accommodation request generation unit 140 transmits the accommodation request to one of these gateway devices.

  The accommodation availability obtaining unit 150 obtains accommodation availability information transmitted from the gateway device that is the accommodation request transmission destination and received by the communication I / F 100b. The information on whether or not accommodation is possible indicates whether or not the node 10 that requested the accommodation to another gateway device can be accommodated. Here, when the node 10 that requested accommodation can be accommodated, it means that the correlation between the measurement data of this node 10 and the measurement data of the node 10 that can be wirelessly communicated with other gateway devices is high.

  When the node 10 can be accommodated, the accommodation availability acquisition unit 150 notifies the measurement necessity acquisition unit 160 of the fact, and when the node 10 cannot be accommodated, the accommodation request generation unit 150 140 is notified. The accommodation request generation unit 140 notified that the node 10 cannot be accommodated transmits the accommodation request to another gateway apparatus that has not yet transmitted the accommodation request.

  The measurement necessity acquisition unit 160 is notified that the node 10 can be accommodated by another gateway device from the accommodation availability acquisition unit 150, and then transmitted from the gateway device and received by the communication I / F 100b. Get sex information. The information on the necessity of measurement indicates whether or not the measurement data of the node 10 accommodated in another gateway device is necessary for the regression estimation for obtaining the estimation data. The measurement necessity acquisition unit 160 notifies the node setting unit 170 and the measurement data transmission unit 180 of the necessity of the measurement data of the node 10 accommodated in another gateway device.

  The node setting unit 170 finally determines the redundancy of each node 10 based on the regression estimation result of the measurement data during the learning period of the node 10 accommodated by the gateway device 100, and performs measurement during the operation period of each node 10. Set the timing. That is, the node setting unit 170 determines that the redundancy of the node 10 that acquires measurement data used for estimating the measurement values of many nodes 10 is low in the regression estimation by the regression estimation unit 120. Then, the node setting unit 170 sets the measurement frequency of the node 10 with low redundancy to be high. On the other hand, the node setting unit 170 determines that the redundancy of the node 10 that obtains the measurement data having a high correlation with the measurement data of the other node 10 is high in the regression estimation by the regression estimation unit 120. The node setting unit 170 then sets the measurement frequency of the node 10 with high redundancy to be low. The determination of redundancy by the node setting unit 170 may be executed based on, for example, the result of the regression estimation unit 120 optimizing the regression coefficient between measurement data using the GroupLasso technique.

  Also, the node setting unit 170 sets the measurement frequency of the node 10 that needs measurement data among the nodes 10 accommodated in other gateway devices in accordance with the notification from the measurement necessity acquisition unit 160. . On the other hand, the node setting unit 170 sets the measurement frequency of the node 10 that does not require measurement data to be low.

  In accordance with the notification from the measurement necessity acquisition unit 160, the measurement data transmission unit 180 acquires the measurement data of the node 10 that needs the measurement data among the nodes 10 accommodated in other gateway devices, and communicates with the communication I / F 100b. To other gateway devices. That is, the measurement data transmission unit 180 extracts the measurement data of the node 10 accommodated in another gateway device from the measurement data acquired by the measurement data acquisition unit 110, and transmits the measurement data to the gateway device that accommodates this node 10. .

  The memory 100d includes, for example, a RAM or a ROM, and stores various information when processing is executed by the processor 100c.

  Next, the configuration of the gateway apparatus 200 that is the transmission destination of the accommodation request from the gateway apparatus 100 will be described. FIG. 6 is a block diagram illustrating a configuration of gateway apparatus 200 according to the first embodiment. The gateway device 200 is a gateway device that is requested to accommodate the node 10 from another gateway device (gateway device 100). The gateway device 200 illustrated in FIG. 6 includes a wireless communication unit 200a, a communication I / F 200b, a processor 200c, and a memory 200d.

  The wireless communication unit 200a receives measurement data transmitted from the node 10 in a wireless communicable range, and performs predetermined wireless reception processing on the received measurement data.

  The communication I / F 200b is communicably connected to another gateway device via the network N. For example, the communication I / F 200b receives an accommodation request for the node 10 from the gateway device 100 or transmits information on availability of accommodation as a response to the accommodation request. To do.

  The processor 200c includes, for example, a CPU, FPGA, DSP, or the like, and controls the entire gateway device 200. Specifically, the processor 200c includes a measurement data acquisition unit 210, a regression estimation unit 220, an accommodation request acquisition unit 230, an independence degree calculation unit 240, an accommodation availability determination unit 250, a measurement necessity notification unit 260, a node setting unit 270, and A measurement data receiving unit 280 is included.

  The measurement data acquisition unit 210 acquires measurement data transmitted from the node 10 that wirelessly communicates with the gateway device 200 and received by the wireless communication unit 200a. Similar to the measurement data acquisition unit 110 of the gateway device 100, the measurement data acquisition unit 210 acquires measurement data at a frequency common to all the nodes 10 that wirelessly communicate with the gateway device 200 during the learning period. Further, the measurement data acquisition unit 210 acquires measurement data at a different frequency for each node 10 during the operation period.

  In addition, the measurement data acquisition unit 210 acquires measurement data transmitted from another gateway device and received by the measurement data reception unit 280 during the operation period.

  The regression estimation unit 220 performs regression estimation using measurement data. That is, the regression estimation unit 220 generates a regression model for estimating the measurement data of one node 10 from the measurement data of another node 10 and calculates a regression coefficient in the regression model. Specifically, the regression estimation unit 220 obtains a regression model and a regression coefficient between the measurement data of each node 10 using the measurement data acquired during the learning period.

  The accommodation request acquisition unit 230 acquires an accommodation request transmitted from another gateway device and received by the communication I / F 200b. Specifically, the accommodation request acquisition unit 230 includes the identification information of the node 10 having the highest degree of independence among the nodes 10 wirelessly communicating with the gateway device 100 and the measurement data acquired from the node 10 during the learning period. Get the request. The accommodation request acquisition unit 230 may acquire not only the node 10 having the highest degree of independence but also the accommodation requests for the predetermined number of nodes 10 in the order of the degree of independence. An accommodation request for 10 may be acquired.

  The independence calculating unit 240 calculates the independence of each node 10 based on the regression model obtained during the learning period. In addition, when the accommodation request is received, the independence degree calculation unit 240 determines whether each of the nodes 10 is based on the result of regression estimation using the measurement data acquired in the learning period and the measurement data included in the accommodation request. Calculate independence. That is, the independence level calculation unit 240 calculates the independence level from the result of the regression estimation not including the measurement data of the node 10 requested to be accommodated, and also from the result of the regression estimation including the measurement data of the node 10 requested to be accommodated. Calculate independence. In calculating the degree of independence, the above equation (1) can be used.

  The accommodation availability determination unit 250 determines whether or not the gateway device 200 accommodates the node 10 requested to be accommodated by the gateway device 100 based on the independence calculated by the independence degree calculation unit 240. Specifically, the accommodation determination unit 250 includes the measurement data of the node 10 requested to be included in the regression estimation, and determines whether the accommodation is possible depending on whether the average degree of independence is small. Therefore, the accommodation determination unit 250 determines the average value of the independence calculated from the result of the regression estimation including the measurement data of the node 10 requested to be accommodated and the regression estimation that does not include the measurement data of the node 10 requested to be accommodated. The average value of independence calculated from the results is compared. As a result of this comparison, the accommodation determination unit 250 determines that the accommodation is possible when the average value of independence calculated from the result of regression estimation including the measurement data of the node 10 requested to be accommodated is smaller. . Then, the accommodation availability determination unit 250 transmits information indicating that the requested node 10 can be accommodated from the communication I / F 200b to the gateway device 100. In addition, the accommodability determination unit 250 notifies the measurement necessity notification unit 260 that the requested node 10 can be accommodated.

  On the other hand, the accommodation possibility determination unit 250 determines that the accommodation is impossible when the average value of independence calculated from the result of the regression estimation including the measurement data of the node 10 for which accommodation is requested is larger. Then, the accommodation availability determination unit 250 transmits information indicating that the requested node 10 cannot be accommodated from the communication I / F 200b to the gateway device 100.

  When the node 10 requested to be accommodated can be accommodated, the measurement necessity notification unit 260 needs the measurement data of the node 10 requested to be accommodated based on the regression estimation result including the measurement data of the node 10. It is determined whether or not. That is, the measurement necessity notification unit 260 determines the redundancy of the requested node 10 based on the regression estimation result by the regression estimation unit 220, and determines that the measurement data is necessary when the redundancy is low. . Then, the measurement necessity notification unit 260 transmits information indicating that measurement data is necessary from the communication I / F 200b to the gateway device 100. Further, the measurement necessity notification unit 260 determines that the measurement data is unnecessary when the redundancy of the node 10 requested to be accommodated is high, and transmits information indicating that to the gateway device 100.

  The node setting unit 270 finally determines the redundancy of each node 10 based on the regression estimation result of the measurement data during the learning period of the node 10 accommodated by the gateway device 200, and performs measurement during the operation period of each node 10. Set the timing. That is, the node setting unit 270 determines that the redundancy of the node 10 that acquires the measurement data used for estimating the measurement values of many nodes 10 is low in the regression estimation by the regression estimation unit 220. And the node setting part 270 sets so that the measurement frequency of the node 10 with low redundancy may be made high. On the other hand, in the regression estimation by the regression estimation unit 220, the node setting unit 270 determines that the redundancy of the node 10 that acquires the measurement data having a high correlation with the measurement data of the other nodes 10 is high. And the node setting part 270 sets so that the measurement frequency of the node 10 with high redundancy may be made low.

  The measurement data receiving unit 280 receives measurement data of the node 10 that can wirelessly communicate with the gateway device from another gateway device that is a transmission destination of information indicating that the measurement data is necessary, and the measurement data acquisition unit 210. Output to. That is, the measurement data receiving unit 280 receives, from the gateway device 100, measurement data of the node 10 requested to be accommodated by the gateway device 100, for example.

  The memory 200d includes, for example, a RAM or a ROM, and stores various information when processing is executed by the processor 200c.

  Next, the operation of the gateway devices 100 and 200 configured as described above will be described with reference to the sequence diagram shown in FIG. FIG. 7 shows an operation when the gateway device 200 accommodates the node 10 capable of wireless communication with the gateway device 100.

  When the measurement data of the node 10 capable of wireless communication with the gateway device 100 is acquired during the learning period, the regression estimation unit 120 performs regression estimation of the measurement data. Then, the degree of independence of each node 10 is calculated from the result of regression estimation, and a node 10 (hereinafter referred to as “independent node”) having a high degree of independence is specified. The measurement data of the independent node has a low correlation with the measurement data of the other nodes 10 and becomes a factor that reduces the accuracy of regression estimation. Therefore, an accommodation request for requesting accommodation of the independent node is generated and transmitted to the predetermined gateway apparatus 200 (step S101). The accommodation request includes identification information of the independent node and measurement data.

  When the accommodation request is received by the gateway device 200, regression estimation using the measurement data of the independent node and the measurement data of the node 10 capable of wireless communication with the gateway device 200 is executed, and the degree of independence of each node 10 is calculated. The Then, by including the measurement data of the independent node in the regression estimation, it is determined whether or not the independent node can be accommodated by the gateway device 200 depending on whether or not the average value of the independence level is small (step S102). . That is, if the average value of the independence calculated from the regression estimation result including the measurement data of the independent node is smaller than the average value of the independence calculated from the regression estimation result not including the measurement data of the independent node, the independent node Is determined to be accommodated by the gateway device 200. Here, the description will proceed assuming that the independent node is determined to be accomodable by the gateway device 200.

  If it is determined that the independent node can be accommodated, information indicating that fact is transmitted from the gateway device 200 to the gateway device 100 (step S103). When it is determined that the independent node cannot be accommodated, information indicating that is transmitted from the gateway device 200 to the gateway device 100.

  When it is determined that the independent node can be accommodated, the gateway device 200 determines whether or not the measurement data of the independent node is necessary from the regression estimation result including the measurement data of the independent node (step S104). ). That is, the redundancy of the independent node in the regression estimation in the gateway device 200 is determined, and when the redundancy of the independent node is low, it is determined that the measurement data of the independent node is necessary. On the other hand, when the redundancy of the independent node is high, since the measurement value by the independent node can be estimated, it is determined that the measurement data of the independent node is unnecessary. Here, the description will be made assuming that the measurement data of the independent node is determined to be necessary.

  If it is determined that the measurement data of the independent node is necessary, information indicating that is transmitted from the gateway device 200 to the gateway device 100 (step S105). When it is determined that the measurement data of the independent node is unnecessary, information indicating that is transmitted from the gateway device 200 to the gateway device 100.

  When notified that the measurement data of the independent node is necessary, the gateway device 100 sets the measurement timing of the node 10 capable of wireless communication with the gateway device 100 (step S106). That is, the measurement timing of each node 10 is set so as to increase the measurement frequency of the node 10 with low redundancy from the regression estimation result regarding the node 10 finally accommodated in the gateway device 100. In addition, since the measurement data of the independent node accommodated in the gateway device 200 is notified, the measurement frequency of the independent node is set high. Thereby, the measurement data of the node 10 used for the regression estimation in the gateway device 100 and the measurement data of the independent node used for the regression estimation in the gateway device 200 are acquired with high frequency.

  On the other hand, in the gateway device 200, the measurement of each node 10 is performed so as to increase the measurement frequency of the node 10 having low redundancy from the regression estimation result regarding the node 10 finally accommodated in the gateway device 200 including the independent node. Timing is set (step S107). Thereby, the measurement data of the node 10 used for regression estimation in the gateway apparatus 200 comes to be acquired with high frequency.

  In these node settings, since the independent node is accommodated in the gateway device 200, the redundancy of each node 10 is determined based on the regression estimation result using measurement data with high correlation. For this reason, the accuracy of regression estimation in each gateway device is high, and the redundancy determination of the node 10 is also highly accurate. As a result, the node 10 whose measurement frequency is lowered can be appropriately determined, and the power consumption of the entire sensor system can be reduced.

  After the measurement timing of the node 10 is set, measurement data is acquired from the node 10 set with a high measurement frequency in the operation period (step S108). Of the measurement data acquired by the gateway device 100, the measurement data of the independent node is transferred to the gateway device 200 (step S109) and used for regression estimation in the gateway device 200 (step S110). Also in the gateway device 100, regression estimation using measurement data of the node 10 accommodated in the gateway device 100 is executed except for independent nodes. These regression estimations are not limited to the nodes 10 with which the respective gateway devices can directly communicate with each other, but are executed by a single gateway device accommodating the nodes 10 with high correlation of measurement data. For this reason, the accuracy of regression estimation is high, and accurate estimation data can be obtained from measurement data. As a result, even if measurement data is not acquired from all the nodes 10, accurate sensor data of all the nodes 10 can be acquired by regression estimation.

  Next, a sensor control method by the gateway device 100 that requests accommodation of an independent node will be described with reference to the flowchart shown in FIG.

  In the learning period, the measurement data of the node 10 that can wirelessly communicate with the gateway device 100 is received by the wireless communication unit 100 a and acquired by the measurement data acquisition unit 110. The measurement data acquired during the learning period is accumulated as learning data (step S201). Then, regression estimation using the accumulated learning data is performed by the regression estimation unit 120. That is, a regression model and a regression coefficient for estimating the current measurement value from the measurement data of each node 10 in the past are obtained for each node 10. In the regression model, in order to estimate a measurement value of a certain node 10, measurement data of another node 10 having a high correlation is referred to.

  When the regression estimation is executed, the independence level of each node 10 is calculated by the independence level calculation unit 130 (step S202). Specifically, the degree of independence is calculated for each node 10 by applying the number of references and the number of referenced in the regression model for each node 10 to the above equation (1). The degree of independence of a certain node 10 indicates a low correlation between the measurement data of this node 10 and the measurement data of another node 10. Therefore, the accommodation request generation unit 140 selects an independent node having a low correlation with the measurement data of the other nodes 10 and a high degree of independence. Specifically, for example, the node 10 having the maximum degree of independence may be selected as the independent node, and the node 10 having the degree of independence of a predetermined threshold or more may be selected as the independent node.

  When the independent node is selected, the accommodation request generation unit 140 generates an accommodation request including identification information of the independent node and measurement data, and transmits the accommodation request to the predetermined gateway device 200 (step S203). After the accommodation request is transmitted, information indicating whether or not the independent node can be accommodated is transmitted from the destination gateway apparatus 200 and received by the communication I / F 100b (step S204).

  Information indicating availability is acquired by the availability determination unit 150, and it is determined whether the independent node can be accommodated by the gateway device 200 (step S205). When the information acquired by the storage availability acquisition unit 150 indicates that the storage is impossible (No in step S205), the storage request generation unit 140 is notified of this. Then, the accommodation request generation unit 140 determines whether or not the accommodation request has been transmitted to all gateway devices predetermined as the transmission destination of the accommodation request (step S206). As a result of the determination, if there is a gateway apparatus that has not yet transmitted the accommodation request (No in step S206), the accommodation request is transmitted to the gateway apparatus (step S203). On the other hand, if the accommodation request has been transmitted to all the gateway devices (Yes in step S206), there is no gateway device that can accommodate the independent node, and thus the processing ends.

  If the result of determination in step S205 is that the independent device can be accommodated by the gateway device 200 (step S205 Yes), measurement necessity information indicating whether or not measurement data of the independent node is necessary is received from the gateway device 200. Received (step S207). The measurement necessity information is acquired by the measurement necessity acquisition unit 160, and it is determined whether or not the measurement data of the independent node is necessary (step S208). The necessity of measurement related to the independent node is determined by the gateway device 200 that performs regression estimation using measurement data having a high correlation with the measurement data of the independent node. For this reason, the accuracy of the necessity for measurement is high.

  When the measurement data of the independent node is necessary (step S208 Yes), the node setting unit 170 sets the measurement timing so as to increase the measurement frequency of the independent node (step S209). As a result, during the operation period, measurement data is received from the independent node at a high frequency, so that the measurement data transmission unit 180 transmits the measurement data of the independent node to the gateway device 200 (step S210).

  On the other hand, when the measurement data of the independent node is unnecessary (No at Step S208), the node setting unit 170 sets the measurement timing so as to reduce the measurement frequency of the independent node (Step S211). As a result, the measurement data is not received from the independent node during the operation period, but the estimation data estimated by the regression estimation in the gateway device 200 is received (step S213). Since the estimation data is obtained by regression estimation using measurement data having a high correlation with the measurement data of the independent node, estimation data with high accuracy is received from the gateway device 200 as the measurement value of the independent node.

  Whether or not the measurement data of the independent node is necessary or not, the node setting unit 170 determines the redundancy of each node 10 from the regression estimation result of the learning data, and the measurement is performed with low redundancy and high frequency. The necessary measurement frequency of the node 10 is set high. In addition, the measurement frequency of the node 10 that is highly redundant and does not require high-frequency measurement is set low. Thereby, unnecessary measurement is reduced in the entire sensor system, and power consumption can be reduced.

  Next, an accommodation sensor determination method by the gateway device 200 that is required to accommodate an independent node will be described with reference to a flowchart shown in FIG.

  In the learning period, the measurement data of the node 10 that can wirelessly communicate with the gateway device 200 is received by the wireless communication unit 200a and acquired by the measurement data acquisition unit 210. The measurement data acquired during the learning period is accumulated as learning data (step S301). Then, the regression estimation unit 220 performs regression estimation using the accumulated learning data. That is, a regression model and a regression coefficient for estimating the current measurement value from the measurement data of each node 10 in the past are obtained for each node 10.

  The gateway device 100 also accumulates learning data, and transmits an accommodation request including identification information of independent nodes and measurement data to the gateway device 200. The accommodation request is received by the communication I / F 200b (step S302) and acquired by the accommodation request acquisition unit 230.

  The measurement data of the independent node included in the accommodation request is output to the regression estimation unit 220, and regression estimation using the learning data stored in the gateway device 200 and the measurement data of the independent node is executed. That is, regression estimation using the measurement data of the independent node in addition to the node 10 wirelessly communicating with the gateway device 200 is executed. Then, the degree of independence for each node 10 is calculated by the independence degree calculation unit 240 from the regression estimation result that does not include the measurement data of the independent node, and the independence level for each node 10 also from the regression estimation result that includes the measurement data of the independent node. The degree is calculated (step S303).

  When the independence degree is calculated by the independence degree calculation unit 240, the accommodation possibility determination unit 250 determines whether or not the independent node can be accommodated by the gateway device 200 (step S304). Specifically, the independence degree calculated from the regression estimation result including the measurement data of the independent node and the average value of the independence degree calculated from the regression estimation result not including the measurement data of the independent node by the accommodation determination unit 250. Is compared with the average value. If the average value of the degree of independence is smaller when the measurement data of the independent node is not included, it is determined that the independent node cannot be accommodated because the correlation of the measurement data is lowered by accommodating the independent node. (Step S304 No). In this case, information indicating that the accommodation of the independent node is rejected is transmitted from the accommodation availability determination unit 250 to the gateway device 100 via the communication I / F 200b (step S306).

  On the other hand, if the average value of the degree of independence is smaller when the measurement data of the independent node is included, since the correlation of the measurement data is increased by accommodating the independent node, it is determined that the independent node can be accommodated. (Step S304 Yes). In this case, information indicating that the accommodation of the independent node is permitted is transmitted from the accommodation availability determination unit 250 to the gateway apparatus 100 via the communication I / F 200b (step S305).

  When the nodes 10 accommodated by the gateway device 200 are determined, regression estimation using the measurement data of these nodes 10 including independent nodes is executed by the regression estimation unit 220 (step S307). In this regression estimation, the measurement data of the node 10 requested to be accommodated by another gateway device and permitted to be accommodated, and the measurement of the node 10 that is not accommodated in the other gateway device among the nodes 10 wirelessly communicating with the gateway device 200 Data is used. That is, the regression estimation using the measurement data of the nodes 10 accommodated in the gateway device 200 is finally executed, and the regression model and the regression coefficient for associating the measurement data of these nodes 10 are obtained.

  By obtaining the regression model, it is determined for each node 10 whether measurement data is necessary or whether a measurement value can be estimated from the measurement data of other nodes 10. Therefore, the measurement necessity notification unit 260 determines whether or not the measurement data of the independent node is necessary (step S308), and when the measurement data of the independent node is necessary (Yes in step S308), the measurement data is stored. The requested information is transmitted from the communication I / F 200b to the gateway device 100 (step S309). In the operation period, the measurement data of the independent node transmitted from the gateway device 100 is received by the measurement data receiving unit 280 (step S310). The measurement data of the independent node is output to the regression estimation unit 220 together with the measurement data acquired by the measurement data acquisition unit 210, and the measurement value of the node 10 for which no measurement data has been acquired is estimated. Sensor data including measurement data and estimation data is output from the regression estimation unit 220.

  On the other hand, if the measurement data of the independent node is unnecessary in the determination in step S308 (No in step S308), information indicating that the measurement data is unnecessary and the estimated data is sent is sent from the communication I / F 200b to the gateway apparatus 100. It is transmitted (step S311). In the operation period, the measured value of the independent node is estimated by the regression estimation unit 220, and the obtained estimated data is transmitted to the gateway device 100 (step S312).

  As described above, according to the present embodiment, a node having a high degree of independence among nodes that can be wirelessly communicated by a gateway device is accommodated in another gateway device, and regression estimation of measurement data of the node accommodated in units of gateway devices is performed. To set the node measurement frequency according to the redundancy. Therefore, it is possible to execute regression estimation by accommodating nodes having high correlation of measurement data in one gateway device, and it is possible to improve the accuracy of regression estimation. As a result, even if regression estimation is executed in units of gateway devices, it is possible to reduce power consumption while suppressing a decrease in data accuracy.

(Embodiment 2)
A feature of the second embodiment is that the gateway device specifies a gateway device in the vicinity of the own device and transmits a housing request to the neighboring gateway device.

  The configuration of the sensor system and the node according to the second embodiment is the same as that of the first embodiment (FIGS. 1 and 2), and the description thereof is omitted. In the second embodiment, it is assumed that gateway device 200 is located in the vicinity of gateway device 100.

  FIG. 10 is a block diagram showing a configuration of gateway apparatus 100 according to the second embodiment. 10, the same parts as those in FIG. 3 are denoted by the same reference numerals, and the description thereof is omitted. The gateway device 100 shown in FIG. 10 employs a configuration in which a neighboring gateway determination unit (hereinafter abbreviated as “neighboring GW determination unit”) 310 is added to the gateway device 100 shown in FIG.

The neighboring GW determination unit 310 determines a gateway device (neighboring GW) in the vicinity of the gateway device 100 based on the reception status of beacon information transmitted / received between each gateway device and each node 10. Specifically, for example, when each gateway device transmits beacon information including identification information of its own device, the neighboring GW determination unit 310 receives all the information received by the node 10 from the node 10 capable of wireless communication with the gateway device 100. The gateway device information corresponding to the beacon information is acquired. Then, for each gateway device other than the gateway device 100, the neighboring GW determination unit 310 calculates the ratio of the nodes 10 that have received the beacon information in the entire nodes 10 that can wirelessly communicate with the gateway device 100 as the area overlap rate. That is, for example, when there are three nodes 10 that can wirelessly communicate with the gateway device 100 and two of the nodes 10 receive the beacon information of the gateway device 200, the area overlap between the gateway device 100 and the gateway device 200 The rate is calculated to be about 67% (≈2 / 3).

  After calculating the area duplication rate for each gateway device, the neighboring GW determination unit 310 compares the area duplication rate with a predetermined threshold value, and determines that the gateway device having the area duplication rate equal to or higher than the predetermined threshold value is the neighboring GW. To do. Then, the neighboring GW determination unit 310 notifies the accommodation request generation unit 140 of the determined neighboring GW.

  In the present embodiment, the accommodation request generation unit 140 transmits an independent node accommodation request to the neighboring GW notified from the neighboring GW determination unit 310.

  Next, a neighbor GW determination method according to Embodiment 2 will be described with reference to the flowchart shown in FIG.

  Each gateway device periodically transmits beacon information including its own identification information. For this reason, the gateway apparatus 100 also transmits beacon information (step S401). The transmitted beacon information is received by the node 10 in a range where wireless communication with each gateway device is possible. Then, the node 10 measures RSSI (Received Signal Strength Indicator) of the received beacon information, and acquires identification information of the gateway device from beacon information whose RSSI is equal to or greater than a predetermined threshold. The node 10 transmits gateway information (GW information) including the acquired identification information to the gateway device 100 connected to the own node. At this time, the node 10 is connected to the gateway device 100, but also receives beacon information from a gateway device in a wireless communicable range other than the gateway device 100. Therefore, the node 10 transmits GW information including identification information of not only the gateway device 100 but also other gateway devices to the gateway device 100.

  The GW information transmitted from the node 10 is received by the gateway device 100 (step S402). The received GW information is acquired by the neighboring GW determination unit 310, and the GW information transmitted from each node 10 is aggregated to calculate the area overlap rate (step S403). That is, for each gateway device other than the gateway device 100, the ratio of the number of nodes 10 that have received beacon information is calculated as the area overlap rate. The area overlap rate of a certain gateway device is a ratio of the number of nodes 10 that have received beacon information from the gateway device among the nodes 10 connected to the gateway device 100. For this reason, a gateway device with a high area overlap rate can wirelessly communicate with many nodes 10 in common with the gateway device 100 and is considered to be in the vicinity of the gateway device 100.

  When the area overlap rate for each gateway device is calculated, a gateway device having an area overlap rate equal to or greater than a predetermined value is determined to be a neighboring GW of the gateway device 100 (step S404). The determined neighboring GW is notified to the accommodation request generating unit 140, and an accommodation request for an independent node is transmitted to the neighboring GW. Since the neighboring GW is a gateway device in the vicinity of the gateway device 100, the node 10 that acquires measurement data having a high correlation with the measurement data of the node 10 that can wirelessly communicate with the gateway device 100 can wirelessly communicate with the neighboring GW. Is likely. For this reason, the possibility that an independent node is accommodated can be increased by transmitting an accommodation request for an independent node to the neighboring GW.

  Thus, in the present embodiment, gateway device 100 determines a neighboring GW and transmits an accommodation request to the neighboring GW. Other operations of the gateway device 100 are the same as those in the first embodiment.

  As described above, according to the present embodiment, the gateway device identifies a neighboring GW in the vicinity of the own device, and designates the identified neighboring GW as the accommodation request transmission destination. For this reason, it is possible to increase the possibility that an independent node is accommodated by transmitting an accommodation request to a gateway device that accommodates a node having a high correlation of measurement data.

  In the second embodiment, the method of determining the neighbor GW when the gateway device transmits beacon information has been described. However, the neighbor GW may be determined even when each node 10 transmits beacon information. Is possible. FIG. 12 is a flowchart showing a neighbor GW determination method when each node 10 transmits beacon information.

  Each node 10 periodically transmits beacon information including its own node identification information. For this reason, the node 10 capable of wireless communication with the gateway device 100 also transmits beacon information, and the gateway device 100 receives the beacon information (step S501). The received beacon information includes beacon information of the node 10 connected to another gateway device.

  The gateway device 100 measures the RSSI of the received beacon information, and acquires the identification information of the node 10 from the beacon information whose RSSI is equal to or greater than a predetermined threshold (step S502). The gateway device 100 determines whether or not the acquired identification information includes the identification information of the node 10 connected to another gateway device (step S503). It is assumed that the node 10 to be connected is registered in advance in each gateway device, and the information of the node 10 connected to each gateway device is shared by all gateway devices. Therefore, the gateway device 100 can determine whether or not the node 10 is a node 10 connected to another gateway device by referring to the beacon information received from the node 10.

  When the identification information of the node 10 connected to the other gateway device is included in the beacon information (Yes in step S503), the gateway device 100 identifies the identification information of the node 10 with respect to the other gateway device to which the node 10 is connected. The node information including is transmitted (step S504). Thereby, the gateway apparatus 100 can notify the gateway apparatus connected to this node 10 that beacon information has been received from the node 10 not connected to the own apparatus.

  After transmitting the node information or when the beacon information is not received from the node 10 connected to another gateway device (No in Step S503), the gateway device 100 receives the node information from the other gateway device (Step S505). . That is, node information including identification information of the node 10 whose beacon information is received by another gateway device among the nodes 10 connected to the gateway device 100 is received.

  Then, by summing up node information received from each gateway device, an area overlap rate is calculated (step S506). That is, for each gateway device that is the source of node information, the ratio of the number of nodes 10 that have received beacon information among the nodes 10 connected to the gateway device 100 is calculated as the area overlap rate. The area overlap rate of a certain gateway device is the ratio of the number of nodes 10 that have received beacon information by the gateway device out of the nodes 10 connected to the gateway device 100. For this reason, a gateway device with a high area overlap rate can wirelessly communicate with many nodes 10 in common with the gateway device 100 and is considered to be in the vicinity of the gateway device 100.

  When the area overlap rate for each gateway device is calculated, a gateway device having an area overlap rate equal to or greater than a predetermined value is determined to be a neighboring GW of the gateway device 100 (step S507). The determined neighboring GW is notified to the accommodation request generating unit 140, and an accommodation request for an independent node is transmitted to the neighboring GW.

  Thus, also when the node 10 transmits beacon information, the gateway apparatus 100 can determine the neighboring GW as in the case where the gateway apparatus transmits beacon information.

  In the first and second embodiments, the case where an independent node capable of wireless communication with the gateway device 100 is accommodated in the gateway device 200 has been described in detail. However, the gateway device 100 may accommodate an independent node capable of wireless communication with other gateway devices, or the gateway device 200 may request the accommodation of the independent node. And if the node 10 which each gateway apparatus accommodates differs, the regression model and regression coefficient used for regression estimation will also change. For this reason, when each gateway device transmits an accommodation request to another gateway device at an arbitrary time, the node 10 accommodated in each gateway device is different, and the regression model and the regression coefficient change, The need for measurement data can also change.

  Thus, since it is not preferable that the node 10 accommodated by each gateway device is changed to an arbitrary time, the period during which all gateway devices can transmit the accommodation request is limited to, for example, a fixed time of one day. Is preferred. That is, it is preferable that each gateway device transmits an independent node accommodation request only within this period, and the node 10 accommodated by each gateway apparatus is determined at the end of the period. Then, after the period is over, regression estimation for each gateway device may be executed to determine whether measurement data of each node 10 is necessary.

  In the first and second embodiments, the gateway device that accommodates each node 10 when the error of regression estimation in each gateway device is periodically calculated and the error becomes larger than a predetermined reference is provided. It is also possible to return to the initial state and determine the independent node again. By doing so, the accuracy of regression estimation can be improved and the power consumption can be reduced even when the environment in which each node 10 is installed changes. Even when an independent node is determined anew, each gateway device does not send an accommodation request to another gateway device at an arbitrary time, but each gateway device makes an accommodation request for an independent node within a predetermined period. It is preferable to transmit.

  Note that the processing of the gateway devices 100 and 200 described in the above embodiments may be described as a program executable by a computer. In this case, this program can be stored in a computer-readable recording medium and introduced into the computer. Examples of the computer-readable recording medium include a portable recording medium such as a CD-ROM, a DVD disk, and a USB memory, and a semiconductor memory such as a flash memory.

100a, 200a Wireless communication unit 100b, 200b Communication I / F
100c, 200c Processor 100d, 200d Memory 110, 210 Measurement data acquisition unit 120, 220 Regression estimation unit 130, 240 Independence calculation unit 140 Accommodation request generation unit 150 Accommodability acquisition unit 160 Measurement necessity acquisition unit 170, 270 Node setting unit 180 measurement data transmission unit 230 accommodation request acquisition unit 250 accommodation availability determination unit 260 measurement necessity notification unit 280 measurement data reception unit 310 neighboring GW determination unit

Claims (10)

A wireless communication unit for receiving measurement data obtained by measurement by a plurality of sensors located in a range where wireless communication with the own device is possible;
A regression estimation unit that performs regression estimation using the measurement data received by the wireless communication unit;
From the regression estimation result executed by the regression estimation unit, a calculation unit that calculates the degree of independence indicating the magnitude of correlation of measurement data with other sensors for each sensor;
According to the degree of independence calculated by the calculation unit, a transmission unit that transmits an accommodation request for requesting accommodation of any sensor to another sensor control device,
As a result of the accommodation request transmitted by the transmission unit, when the sensor is accommodated by the other sensor control device, the measurement frequency of each sensor is determined based on the regression estimation results of the measurement data of a plurality of sensors excluding the sensor. A sensor control device comprising: a setting unit for setting. The calculation unit includes:
In the regression estimation result, the reference number indicating the number of sensors for which the measurement value is estimated by referring to the measurement data of one sensor, and the sensor that refers to the measurement data to estimate the measurement value of the one sensor The sensor control apparatus according to claim 1, wherein the degree of independence of the one sensor is calculated based on a reference number indicating a number. The transmitter is
The sensor control apparatus according to claim 1, wherein the degree of independence calculated by the calculation unit transmits a sensor accommodation request indicating that the correlation of measurement data with another sensor is lower than a predetermined reference. A determination unit that determines a proximity sensor control device located in the vicinity of the device;
The transmitter is
The sensor control device according to claim 1, wherein an accommodation request is transmitted to the neighboring sensor control device determined by the determination unit. The determination unit
The beacon information including the identification information of the sensor control device, the other sensor control device receives the report of the beacon information received by the sensor located in a range where wireless communication with the own device is received, and the other sensor control device The sensor control device according to claim 4, wherein a neighbor sensor control device is determined based on a ratio of sensors that have received beacon information from. The determination unit
Beacon information including sensor identification information, which is transmitted by a sensor located in a range where wireless communication can be performed with the own device and received by other sensor control devices, and received by other sensor control devices. 5. The sensor control device according to claim 4, wherein the proximity sensor control device is determined based on a ratio of sensors from which information is received. A wireless communication unit for receiving measurement data obtained by measurement by a plurality of sensors located in a range where wireless communication with the own device is possible;
A receiving unit that receives from the other sensor control device measurement data received by the other sensor control device from an external sensor located in a range in which wireless communication with the other sensor control device is possible;
A regression estimation unit that performs regression estimation using the measurement data received by the wireless communication unit and the measurement data received by the reception unit;
From the regression estimation result executed by the regression estimation unit, a calculation unit that calculates the degree of independence indicating the magnitude of correlation of measurement data with other sensors for each sensor;
A determination unit that determines whether to accommodate the external sensor based on the degree of independence calculated by the calculation unit;
And a setting unit that sets measurement frequencies of the plurality of sensors based on regression estimation results of measurement data of the external sensors and the plurality of sensors when accommodating the external sensors. . A plurality of sensors; a first gateway device that wirelessly communicates with some of the plurality of sensors; and a second gateway device that wirelessly communicates with some other sensors of the plurality of sensors. A sensor system,
The first gateway device is:
A first wireless communication unit that receives measurement data obtained by measurement by the some sensors;
A first regression estimation unit that performs regression estimation using the measurement data received by the first wireless communication unit;
A first calculation unit that calculates the degree of independence indicating the magnitude of correlation of measurement data with other sensors for each sensor from the regression estimation result executed by the first regression estimation unit;
A transmission unit that transmits an accommodation request for requesting accommodation of any sensor to the second gateway device according to the degree of independence calculated by the first calculation unit;
As a result of the accommodation request being transmitted by the transmission unit, when a sensor is accommodated by the second gateway device, measurement of each sensor is performed based on the regression estimation result of the measurement data of the part of sensors excluding the sensor. A first setting unit for setting the frequency,
The second gateway device is
A second wireless communication unit for receiving measurement data obtained by measurement by the other part of the sensors;
A receiving unit that receives the accommodation request that is the accommodation request transmitted by the transmission unit and includes measurement data of an external sensor included in the part of the sensors;
A second regression estimation unit that performs regression estimation using the measurement data received by the second wireless communication unit and the measurement data received by the reception unit;
A second calculation unit that calculates the degree of independence indicating the magnitude of correlation of measurement data with other sensors for each sensor from the regression estimation result executed by the second regression estimation unit;
A determination unit that determines whether to house the external sensor based on the degree of independence calculated by the second calculation unit;
A second setting unit configured to set the measurement frequency of the other part of the sensor based on the regression estimation result of the measurement data of the external sensor and the plurality of sensors when the external sensor is accommodated. A sensor system. A sensor control method executed by a sensor control device,
Receiving measurement data obtained by measurement by a plurality of sensors located in a range in which wireless communication with the sensor control device is possible;
Perform regression estimation using received measurement data,
Calculate the degree of independence indicating the magnitude of the correlation of measurement data with other sensors for each sensor from the results of the regression estimation performed,
Depending on the calculated degree of independence, an accommodation request for requesting accommodation of any sensor is transmitted to another sensor control device,
As a result of the accommodation request being transmitted, when the sensor is accommodated by the other sensor control device, a process for setting the measurement frequency of each sensor based on the regression estimation results of the measurement data of a plurality of sensors excluding the sensor is performed. A sensor control method comprising: Receives measurement data obtained by measuring multiple sensors located in a range where wireless communication with the sensor control device is possible,
Perform regression estimation using received measurement data,
Calculate the degree of independence indicating the magnitude of the correlation of measurement data with other sensors for each sensor from the results of the regression estimation performed,
Depending on the calculated degree of independence, an accommodation request for requesting accommodation of any sensor is transmitted to another sensor control device,
As a result of the accommodation request being transmitted, when the sensor is accommodated by the other sensor control device, a process for setting the measurement frequency of each sensor based on the regression estimation results of the measurement data of a plurality of sensors excluding the sensor is performed. A sensor control program that is executed by a computer.

Images