Space state prediction method and space state prediction system

Abstract

To predict a quantity of state within a space using a sensor located at a position where accurate capturing of the state of the space is possible.SOLUTION: Measurement values from multiple sensors 200 located within a space 10 are acquired. A simulated quantity of state at multiple positions within the space 10 at a first time t1 is subjected to data assimilation processing using the measurement values from the multiple sensors 200. Data obtained through the data assimilation processing is input to a prediction model, and a time-dependent quantity of state after assimilation, which is a quantity of state at a predetermined position within the space 10 at a second time t2, is predicted. A simulated quantity of state at the predetermined position at the first time t1 is input to the prediction model, and a time-dependent quantity of state after non-assimilation, which is a quantity of state at the predetermined position at the second time t2, is predicted. Appropriateness of a position of a sensor 200 located at the predetermined position is determined based on the time-dependent quantity of state after assimilation, the time-dependent quantity of state after non-assimilation, and the measurement value of the sensor 200 at the second time t2.SELECTED DRAWING: Figure 9

Description

Technical Field

[0001] The present disclosure relates to a space state prediction method for predicting a state quantity of a space in a building, and a space state prediction system therefor.

Background Art

[0002] A method for updating a model of the air flow state of an air conditioning system in real time using sensors is known. For example, in the method disclosed in Patent Document 1, all modes of a prepared physical model are generated, compared with modes reconstructed from the air flow state measured by sensors arranged in a space, and a reduced-order model associated with the dominant mode of the physical model is selected to predict the space state.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] For example, in the conventional method, after the positions of the sensors are determined once, the space state is predicted using the sensors installed at the same positions. However, when the state of the space changes due to opening and closing of doors and windows, movement of people, changes in solar radiation, etc., it is difficult for the sensors installed at the same positions to capture the changed state of the space. Therefore, there is a need for a method for predicting the state quantity in a space using sensors arranged at positions where the state of the space can be appropriately captured.

[0005] The present disclosure solves the above problems, and provides a space state prediction method and the like capable of predicting the state quantity in a space using sensors arranged at positions where the state of the space can be appropriately captured.

Means for Solving the Problems

[0006] A method for predicting a spatial state according to an aspect of the present disclosure is a method for performing data assimilation processing using measurement values of sensors that detect state quantities in a space and predicting the state quantities in the space. The method includes a measurement value acquisition step of acquiring measurement values of a plurality of the sensors arranged at a plurality of positions in the space, an assimilation result data acquisition step of performing data assimilation processing on simulation state quantities at a plurality of positions in the space at a first time using the measurement values of the plurality of sensors at the first time, and inputting the assimilation result data into a prediction model of the state quantity in the space and time-evolving the data to predict an assimilated time-evolved state quantity that is the state quantity at a predetermined position in the space at a second time; a non-assimilated prediction step of inputting a simulation state quantity at the predetermined position at the first time into the prediction model and time-evolving the data without performing the data assimilation processing to predict a non-assimilated time-evolved state quantity that is the state quantity at the predetermined position at the second time; and a determination step of determining the suitability of the position of the sensor arranged at the predetermined position based on the assimilated time-evolved state quantity, the non-assimilated time-evolved state quantity, and the measurement value of the sensor at the second time.

[0007] A spatial state prediction system according to an aspect of the present disclosure is a spatial state prediction system that performs data assimilation processing using measurement values of sensors that detect state quantities in a space and predicts the state quantities in the space, the information acquisition unit that acquires measurement values of the plurality of sensors arranged at a plurality of positions in the space, and the simulated state quantities at the plurality of positions in the space at the first time are used as the measurement values of the plurality of sensors 200 at the first time. An assimilation unit that obtains assimilation result data by performing data assimilation processing and inputs the assimilation result data into a prediction model of the state quantity in the space and develops it in time to predict the assimilated time-developed state quantity that is the state quantity at a predetermined position in the space at the second time; and without performing the data assimilation processing, the simulated state quantity at the predetermined position at the first time is input into the prediction model and developed in time to predict the non-assimilated time-developed state quantity that is the state quantity at the predetermined position at the second time. A simulation unit, and a determination unit that determines the suitability of the position of the sensor arranged at the predetermined position based on the assimilated time-developed state quantity, the non-assimilated time-developed state quantity, and the measurement value of the sensor at the second time.

[0008] Note that the general or specific aspects of the present disclosure may be implemented by a system, a method, an integrated circuit, a computer program, or a recording medium such as a computer-readable CD-ROM, and may also be implemented by any combination of a system, a method, an integrated circuit, a computer program, and a recording medium.

Effect of the Invention

[0009] According to the spatial state prediction method and the like of the present disclosure, it is possible to predict the state quantity in the space using a sensor arranged at a position where the state of the space can be appropriately grasped.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Mode for Carrying Out the Invention

[0011] Hereinafter, embodiments and the like will be described with reference to the drawings. All of the embodiments and the like described below show comprehensive or specific examples. The numerical values, shapes, materials, components, arrangement positions and connection forms of the components, steps, order of steps, etc. shown in the following embodiments and the like are examples and are not intended to limit the present disclosure. In addition, among the components in the following embodiments and the like, the components not described in the independent claims are described as optional components.

[0012] Also, each figure is a schematic diagram and is not necessarily drawn precisely. In each figure, the same reference numerals are given to substantially the same configurations, and duplicate descriptions may be omitted or simplified. Also, in each figure, even when the same object is illustrated, the scale may be changed for convenience.

[0013] [Basic Configuration of Space State Prediction System] FIG. 1 is a block diagram showing the basic configuration of a space state prediction system.

[0014] A space state prediction system is a system that predicts the state quantities of spaces within a building. The spaces to be predicted are, for example, spaces partitioned by walls in buildings such as houses, offices, stores, public facilities, entertainment facilities, art museums, museums, factories, warehouses, etc. The state quantities of a space are physical quantities indicating the state of the space, such as the temperature distribution, humidity distribution, wind speed (including wind direction) distribution, gas concentration distribution, and PM2.5 (fine particulate matter) distribution of the space.

[0015] As shown in FIG. 1, the space state prediction system 1 includes a space state prediction device 500, and one or more sensors 200 and devices 600 that can communicate via a network 2.

[0016] Each sensor 200 is a device that detects the state quantities and boundary conditions at a predetermined position within the building. The state quantities at the predetermined position are, for example, the temperature, humidity, wind speed, gas concentration, and amount of PM2.5 at the predetermined position. The boundary conditions at a predetermined position are physical quantities indicating the external environment that affects the state quantities of the space, or the state of the boundary region between the space and the outside. The external environment that affects the state quantities of the space is, for example, the outside air temperature, outer wall temperature, outside air humidity, outside wind speed, outside gas concentration, outside amount of PM2.5, and solar radiation amount. The state of the boundary region between the space and the outside is, for example, the opening area (or opening angle) of a door or window provided in the building. The sensor 200 is, for example, a thermometer, a hygrometer, an anemometer, a gas concentration meter, a PM2.5 measuring instrument, a pyranometer, a door sensor, and a window sensor, and is provided inside the space, outside the space, or in the boundary region between the space and the outside. The sensor 200 is attached to a moving device such as a self-propelled robot or a drone. The configurations of the sensor 200 and the moving device will be described later.

[0017] The detection information detected by the sensor 200 is transmitted to the space state prediction device 500 via the network 2. When the sensor 200 is an image sensor, information regarding the position of a person in the space may be transmitted to the space state prediction device 500.

[0018] Each device 600 is a device that forms the environment of a space within a building, for example, an air conditioner, a ventilation fan, an air purifier, a circulator, or a gas diffuser. The device 600 is provided within the space of the building or in the boundary area between the space and the outside. Further, the device 600 transmits operation information including the current operating conditions and past operation history to the space state prediction device 500. The operating conditions of the device 600 include physical quantities such as the temperature, humidity, air volume, and wind direction of the air sent out from the device 600. When the device 600 is an air conditioner, information regarding the set temperature of the air, the blowing air volume, the suction air volume, the rotational speed of the fan, and the power supply amount to the heat exchanger may be included in the operation information. When the device 600 is a gas diffuser that releases diffusion substances such as hypochlorous acid, antibacterial ions, and fragrances, information regarding the released gas concentration, release amount, and release direction of the diffusion substance may be included in the operation information.

[0019] Further, the space state prediction device 500 is communicatively connected to the information terminal 310 and the external information source 320 via the network 2.

[0020] The information terminal 310 is a terminal device owned and carried by a user, and may be, for example, a smart device such as a smartphone, a tablet terminal, and a wearable terminal, or a portable terminal having portability such as a personal computer. The information terminal 310 may receive information regarding the prediction system held by the space state prediction device 500 via the network 2 and be used to notify the user.

[0021] The external information source 320 is IoT (Internet of Things) data existing on the Internet. For example, the IoT data includes weather data.

[0022] The spatial state prediction device 500 is provided in the computer 100 described later. Note that the spatial state prediction device 500 may be provided in a computer on the cloud. The spatial state prediction device 500 has an assimilation unit 510 that performs data assimilation processing and a storage unit 105 that stores various types of information for performing simulations.

[0023] In the assimilation unit 510, the detection information detected by the sensor 200 and the simulation model (prediction model) are integrated, and appropriate initial values, boundary values, and parameters are determined so as to reproduce the phenomenon. In the storage unit 105, the layout information of the space is stored in advance. The layout information includes information regarding the shape and size of the space, and information regarding objects arranged in the space, such as desks and partitions, and the positions of the objects. For example, the information on the shape of the space is data obtained by converting a 3D model of the space to be analyzed into a point cloud by the finite volume method.

[0024] FIG. 2 is a diagram showing an example of the space 10 in a building.

[0025] FIG. 2 shows a view of the space 10 as seen from above. As sensors 200 for detecting state quantities in the space 10, a thermometer, a hygrometer, and an anemometer are provided, and as devices 600 for forming the environment of the space 10, an air conditioner, a circulator, a ventilation fan, and an air purifier are shown as an example. Also shown in the figure are a door sensor for detecting the open / closed state of the door and a window sensor for detecting the open / closed state of the window as sensors 200 for detecting boundary conditions. Each of the door sensor and the window sensor can detect not only the presence or absence of opening / closing but also the amount of opening / closing. In the figure, the illustration of the gas concentration meter and the gas diffusion device is omitted. The detection information detected by the sensor 200 and the operation information of the device 600 are transmitted to the spatial state prediction device 500 via the network 2.

[0026] The spatial state prediction device 500 performs a simulation of the space 10 based on various types of information stored in the storage unit 105, detection information acquired from the sensor 200, and IoT data and the like acquired from the external information source 320, and predicts the state quantities of the space 10.

[0027] FIG. 3 is a diagram showing a computer 100 that constitutes a space state prediction device 500 included in the space state prediction system 1.

[0028] The computer 100 includes an input unit 101, an arithmetic circuit 102, a memory 103, an output unit 104, a storage unit 105, and a communication unit 106.

[0029] The communication unit 106 communicates wirelessly or wiredly with the sensor 200, the device 600, the information terminal 310, and the external information source 320 via the network 2. The communication method of the wireless communication may be Wi-Fi (registered trademark), Bluetooth (registered trademark), or ZigBee (registered trademark), or other methods.

[0030] The input unit 101 has a function as an HMI (Human Machine Interface) that receives input operations by the user, and includes, for example, a keyboard, a mouse, a touch sensor, a touch pad, etc. A part of the layout information of the space 10 may be input to the computer 100 via the input unit 101.

[0031] The output unit 104 has a display that displays an image, characters, etc., and the display is, for example, a liquid crystal display, a plasma display, an organic EL (Electro-Luminescence) display, etc. Note that the output unit 104 may have a printer that prints an image, characters, etc., and may have a function of storing the data output from the arithmetic circuit 102 in the storage unit 105 in a file format.

[0032] The memory unit 105 stores a program (i.e., a computer program) 105a in which each instruction to the arithmetic circuit 102 is described. The program 105a is stored in the memory unit 105 via, for example, a removable medium or a network. The removable medium is, for example, a CD-ROM (Compact Disc Read Only Memory), a flash memory, or the like. Therefore, the communication unit 106 may include an interface for reading the program 105a of the removable medium.

[0033] In addition, the memory unit 105 stores simulation software for performing numerical analysis. Examples of the simulation software include CFD (Computational Fluid Dynamics) and BIM (Building Information Modeling).

[0034] In addition, the memory unit 105 stores each temporary data 105b temporarily generated by the processing of the arithmetic circuit 102. Such a memory unit 105 is a non-volatile recording medium, such as a magnetic storage device such as a hard disk, an optical disk, or a semiconductor memory. In the present embodiment, the layout information and simulation information of the space 10 are stored in the memory unit 105. In addition, information such as boundary conditions, state quantities of the space 10, and operating conditions of devices is stored in the memory unit 105 during the process of the operation.

[0035] The memory 103 temporarily stores the program 105a read out and expanded by the arithmetic circuit 102. Such a memory 103 is, for example, a volatile RAM (Random Access Memory).

[0036] The arithmetic circuit 102 is a circuit that executes the program 105a developed in the memory 103, and is, for example, a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or the like. When executing the program 105a, the arithmetic circuit 102 may use each temporary data 105b stored in the storage unit 105.

[0037] The arithmetic circuit 102 is a circuit for realizing the function of the spatial state prediction device 500. The arithmetic circuit 102 performs a simulation of the space 10 using simulation software and predicts the state quantity of the space 10.

[0038] [Spatial state prediction system of Example 1] The spatial state prediction system 1A of Example 1, which is an example of the above spatial state prediction system 1, will be described.

[0039] The spatial state prediction system 1A is a system that performs data assimilation processing using the measurement values of the sensor 200 that detects the state quantity in the space 10 and predicts the state quantity in the space 10.

[0040] The state of the space 10 in the building changes, for example, due to the operating conditions of the air conditioning equipment provided in the space 10, the opening and closing of windows or doors, the surrounding environment such as the weather, and the movement of people. Along with this change, the state quantities in the space 10, such as gas concentration, air flow, and temperature, also fluctuate. Therefore, the current position of the sensor 200 arranged in the space 10 may be an inappropriate position as a measurement position for predicting the state quantity of the space 10. In that case, the prediction accuracy of the state quantity of the space 10 obtained by the data assimilation processing decreases. The spatial state prediction system 1A of Example 1 has a configuration that can determine whether the sensor 200 is arranged at an appropriate position in response to changes such as the surrounding environment and the movement of people.

[0041] FIG. 4 is a block diagram showing the configuration of the spatial state prediction system 1A of Example 1.

[0042] The spatial state prediction system 1A of Example 1 includes a spatial state prediction device 500A, a sensor mobile body 250, a device 600, etc. The configuration of the device 600 is the same as that in the embodiment shown in FIG. 1.

[0043] FIG. 5 is a diagram showing an example of the sensor mobile body 250 disposed in the space 10.

[0044] As shown in the figure, the sensor mobile body 250 is composed of a sensor 200 and a moving device 210 to which the sensor 200 is attached.

[0045] The sensor 200 is at least one of a thermometer, a hygrometer, an anemometer, a gas concentration meter, a PM2.5 meter, and a pyranometer. The sensor 200 measures the state quantity in the space 10. The measurement data obtained by the sensor 200 is output to the moving device 210. Also, the measurement data obtained by the sensor 200 is output to the spatial state prediction device 500A via the moving device 210.

[0046] The moving device 210 is a device for moving the sensor 200. The moving device 210 is, for example, a drone. Note that the moving device 210 may be a self-propelled robot that freely moves around in the space 10, or a moving robot that moves along a rail laid in the space 10.

[0047] The moving device 210 has an information acquisition unit 220, a control unit 230, a drive mechanism 240, and a positioning sensor 201.

[0048] The drive mechanism 240 is composed of a plurality of parts for moving the sensor mobile body 250, and includes parts such as a motor, a gear, and a propeller.

[0049] The control unit 230 controls the operations of the sensor 200, the positioning sensor 201, the information acquisition unit 220, and the drive mechanism 240. The control unit 230 has a communication function for communicating with the spatial state prediction device 500A.

[0050] The information acquisition unit 220 acquires information regarding the measurement values of the sensor 200 output from the sensor 200. The information regarding the measurement values of the sensor 200 is transmitted to the space state prediction device 500A via the control unit 230.

[0051] The positioning sensor 201 has a GPS (Global Positioning System) function for acquiring its own position information. The positioning sensor 201 may include a gyro sensor and an acceleration sensor. The position information of the sensor 200 obtained by the positioning sensor 201 is transmitted to the space state prediction device 500A via the control unit 230.

[0052] FIG. 6 is a diagram showing an example of the space 10 in a building to which the prediction by the space state prediction system 1A of the first embodiment is applied.

[0053] In the figure, an example where the space 10 is an office space is shown. Also, in the figure, the device 600 arranged on the ceiling of the space 10 is a gas diffusion device that emits a predetermined diffusing substance, and examples are shown where the sensors 200a, 200b, 200c, 200d, 200e, 200f respectively mounted corresponding to the sensor mobile bodies 250a, 250b, 250c, 250d, 250e, 250f are gas concentration meters for detecting the gas concentration of the predetermined diffusing substance. Hereinafter, all or any one of the sensor mobile bodies 250a to 250f may be referred to as the sensor mobile body 250, and all or any one of the sensors 200a to 200f may be referred to as the sensor 200.

[0054] In the space state prediction system 1A of the first embodiment, in order to measure the state quantity of the space 10 at an appropriate location according to changes in the situation within the space 10, the sensor 200 arranged at an inappropriate position is moved to a position different from the current position.

[0055] In Fig. 6(a), the positions of sensors 200a to 200f between the first time t1 and the second time t2 are shown. The period from the first time t1 to the second time t2 is, for example, a period of 10 minutes or more and 50 minutes or less. Each of the sensors 200a to 200f measures the state quantity in space 10 at a predetermined sampling period. The sampling period is, for example, a period that is 1 / 10 or less of the period between the first time t1 and the second time t2.

[0056] Between the first time t1 and the second time t2, the positions of sensors 200a to 200f are fixed at predetermined positions. Being fixed at a predetermined position does not mean being fixed to a wall or a floor, but means being fixed as position coordinates. The positions of sensors 200a to 200f in this embodiment may be changed to other positions according to predetermined conditions from the initially set positions. In that case, the other positions after the change become new predetermined positions.

[0057] In Fig. 6(b), the positions of sensors 200a to 200f at the third time t3, which is a time after the second time t2, are shown. In this example, the positions of sensors 200a, 200c to 200f are determined to be appropriate, and the arranged positions of sensors 200a, 200c to 200f are maintained. On the other hand, the position of sensor 200b is determined to be inappropriate, and the position of sensor 200b is moving to another position in the direction of the arrow.

[0058] The space state prediction device 500A of Example 1 has the following configuration for determining the appropriateness of the position of sensor 200.

[0059] As shown in Fig. 4, the space state prediction device 500A includes the memory unit 105 described above, and an assimilation unit 510A, an information acquisition unit 520, a simulation unit 530, a determination unit 540, and a control unit 550.

[0060] The information acquisition unit 520 acquires information regarding the measurement values of the sensor 200, which is the information output from the sensor mobile body 250, and the position information of the sensor 200. The simulation unit 530 performs a normal simulation using the input conditions. The assimilation unit 510 performs data assimilation processing using the measurement values of the sensor 200 at a time during the simulation. The determination unit 540 determines whether it is necessary to change the position of the sensor 200 based on the various information acquired by the information acquisition unit 520, the calculation processing result in the simulation unit 530, and the calculation processing result in the assimilation unit 510.

[0061] FIG. 7 is a diagram showing an example of a simulation and data assimilation processing for predicting the state quantity of the space 10.

[0062] In FIG. 7(a), an example is shown in which the state quantity in the simulation at the initial time t0 is used as input data, and a normal simulation is performed from the initial time t0 to the third time t3, that is, an example in which data assimilation processing is not performed.

[0063] In FIG. 7(b), an example in which data assimilation processing is performed is shown. Specifically, in the figure, an example is shown in which the state quantity in the simulation at the initial time t0 is used as input data, and a simulation is performed from the initial time t0 to the first time t1. Also, an example is shown in which the state quantity in the simulation at the first time t1 is assimilated using the measurement value r1 of the sensor 200 at the first time t1, and a simulation is performed from the first time t1 to the second time t2 using the assimilated state quantity as input data. Further, an example is shown in which the state quantity in the simulation at the second time t2 is assimilated using the measurement value r2 of the sensor 200 at the second time t2, and a simulation is performed from the second time t2 to the third time t3 using the assimilated state quantity as input data.

[0064] In Fig. 7(c), an example of performing a normal simulation by changing the input data at each of the first time t1, the second time t2, and the third time t3, and an example of performing the data assimilation process shown in (b) are shown. In this embodiment, the scenario shown in Fig. 7(c) is executed.

[0065] As shown in Fig. 7(c), the spatial state prediction device 500A of this embodiment uses the state quantity in the simulation at the initial time t0 as input data and performs a normal simulation from the initial time t0 to the first time t1. Next, the state quantity in the simulation at the first time t1 is assimilated using the measured value r1 of the sensor 200 at the first time t1, and the simulation is performed from the first time t1 to the second time t2 using the state quantity obtained by the assimilation as input data. On the other hand, the state quantity in the simulation at the first time t1 is used as input data as it is, and a normal simulation is performed from the first time t1 to the second time t2. Next, the state quantity in the simulation at the second time t2 is assimilated using the measured value r2 of the sensor 200 at the second time t2, and the simulation is performed from the second time t2 to the third time t3 using the state quantity obtained by the assimilation as input data. On the other hand, the state quantity in the simulation at the second time t2 is used as input data as it is, and a normal simulation is performed from the second time t2 to the third time t3.

[0066] Fig. 8 is a diagram showing an example of the change in the state quantity in the space 10.

[0067] The state quantities f11 and f11m shown in the figure are state quantities obtained by time-evolving a predetermined state quantity on the simulation at the initial time t0 as input data. Time-evolving means performing a simulation for a predetermined period. In this example, each of the periods from the initial time t0 to the first time t1, from the first time t1 to the second time t2, and from the second time t2 to the third time t3 is shown as the period for time-evolving. Note that the state quantities f11 and f11m are state quantities on the simulation corresponding to the respective positions of the plurality of sensors 200 arranged in the space 10. In this example, the state quantities at a plurality of positions are collectively denoted as "state quantity fxxm", and the state quantity at a predetermined position among the plurality of positions is denoted as "state quantity fxx" (where xx is a number).

[0068] The assimilation unit 510A obtains assimilation result data a11 by performing data assimilation processing on the state quantity f11m on the simulation at a plurality of positions in the space 10 at the first time t1 using the measurement values r1 of the plurality of sensors 200 at the first time t1. Note that the assimilation unit 510A performs data assimilation processing using not only the measurement values of the sensors 200 but also the position information of the sensors 200 and the like.

[0069] That is, the assimilation unit 510A performs data assimilation processing on each state quantity f11 corresponding to a plurality of positions using the measurement values (actual measurement values) of the plurality of sensors 200, and obtains assimilation result data a11 after the assimilation processing. The assimilation result data a11 may be a data group composed of a plurality of state quantities, or may be data indicating the distribution of the state quantity in the space 10.

[0070] The assimilation unit 510A inputs the assimilation result data a11 obtained by the data assimilation processing into the prediction model of the state quantity in the space 10 and time-evolves it to predict the time-evolved state quantity f12 after assimilation, which is the state quantity at a predetermined position in the space 10 at the second time t2.

[0071] The simulation unit 530 predicts the non-assimilated time-evolved state quantity f22, which is the state quantity at a predetermined position at the second time t2, by inputting the simulated state quantity f11 at a predetermined position at the first time t1 into the above prediction model and evolving it over time without performing data assimilation processing. Note that the simulation unit 530 performs simulation using the state quantity f11 at a predetermined position among a plurality of positions, and does not perform simulation using the state quantities at other positions different from the predetermined position.

[0072] The determination unit 540 determines the suitability of the position of the sensor 200 disposed at a predetermined position based on the assimilated time-evolved state quantity after performing data assimilation processing, the non-assimilated time-evolved state quantity without performing data assimilation processing, and the measurement value of the sensor 200, which is the measured value.

[0073] For example, the determination unit 540 determines the suitability of the position of the sensor 200 based on which of the assimilated time-evolved state quantity f12 and the non-assimilated time-evolved state quantity f22 is closer to the measurement value r2 of the sensor 200 at the second time t2. Specifically, when comparing at the second time t2, if the non-assimilated time-evolved state quantity f22 is closer to the measurement value r2 of the sensor 200 than the assimilated time-evolved state quantity f12 (for example, when (f12 - r2) > (f22 - r2)), the determination unit 540 determines that the position of the sensor 200 disposed at the predetermined position is inappropriate. When it is determined that the position of the sensor 200 is inappropriate, the control unit 550 moves the sensor 200 using the sensor mover 250.

[0074] The control unit 550 transmits a movement signal for moving the sensor 200 to the control unit 230 of the sensor mover 250. The drive mechanism 240 of the sensor mover 250 drives based on the movement signal output from the control unit 230 and moves the sensor 200.

[0075] The sensor mobile body 250 autonomously moves within the space 10 and arranges the sensor 200 mounted thereon at a position different from the current position. For example, the sensor mobile body 250 may autonomously move with random movements like a cleaning robot. Also, the sensor mobile body 250 may determine a movement range or a movement route based on the learning result obtained by machine learning based on the past movement history and then move. Further, a movement permission area within the space 10 may be preset for the sensor mobile body 250 so that a plurality of sensor mobile bodies 250 do not exist in the same area at the same time.

[0076] On the other hand, when comparing at the second time t2, the determination unit 540 determines that the position of the sensor 200 arranged at a predetermined position is appropriate if the time-developed state quantity f12 after data assimilation is closer to the measured value r2 of the sensor 200 than the non-assimilated time-developed state quantity f22 for which data assimilation processing was not performed (for example, when (f12 - r2) < (f22 - r2)). When it is determined that the position of the sensor 200 is appropriate, the control unit 550 maintains the position of the sensor 200 without moving it.

[0077] When the position of the sensor 200 is appropriate, the space state prediction device 500A performs the same processing from the second time t2 to the third time t3.

[0078] The state quantities f12 and f12m shown in the figure are state quantities obtained by time-developing the assimilation result data a11 at the first time t1 as input data, that is, the time-developed state quantities f12 and f12m after assimilation.

[0079] The assimilation unit 510A acquires the assimilation result data a12 by performing data assimilation processing on the simulated state quantities f12m at a plurality of positions within the space 10 at the second time t2 using the measured values r2 of the plurality of sensors 200 at the second time t2. The assimilation unit 510A inputs the assimilation result data a12 obtained by the data assimilation processing into the prediction model of the state quantity within the space 10 and time-develops it to predict the time-developed state quantity f13 after assimilation, which is the state quantity at a predetermined position at the third time t3.

[0080] Without performing data assimilation processing, the simulation unit 530 inputs the simulation state quantity f12 at a predetermined position at the second time t2 into the above prediction model and develops it over time to predict the non-assimilated time-developed state quantity f23, which is the state quantity at a predetermined position at the third time t3.

[0081] Based on the assimilated time-developed state quantity after performing data assimilation processing, the non-assimilated time-developed state quantity without performing data assimilation processing, and the measured value of the sensor 200, which is the measured value, the determination unit 540 determines the suitability of the position of the sensor 200 arranged at a predetermined position.

[0082] For example, at the third time t3, the determination unit 540 determines the suitability of the position of the sensor 200 based on which of the assimilated time-developed state quantity f13 and the non-assimilated time-developed state quantity f23 is closer to the measured value r3 of the sensor 200. Specifically, when comparing at the third time t3, if the non-assimilated time-developed state quantity f23 is closer to the measured value r3 of the sensor 200 than the assimilated time-developed state quantity f13 (for example, when (f13 - r3) > (f23 - r3)), the determination unit 540 determines that the position of the sensor 200 arranged at a predetermined position is inappropriate. When it is determined that the position of the sensor 200 is inappropriate, the control unit 550 moves the sensor 200 using the sensor moving body 250.

[0083] On the other hand, when comparing at the third time t3, if the assimilated time-developed state quantity f13 after performing data assimilation processing is closer to the measured value r3 of the sensor 200 than the non-assimilated time-developed state quantity f23 without performing data assimilation processing (for example, when (f13 - r3) < (f23 - r3)), the determination unit 540 determines that the position of the sensor 200 arranged at a predetermined position is appropriate. When it is determined that the position of the sensor 200 is appropriate, the control unit 550 maintains the position of the sensor 200 without moving it.

[0084] Thus, in this embodiment, the effect of the assimilation result data a11 assimilated using the measurement value r1 at the first time t1 is determined at the second time t2, and the effect of the assimilation result data a12 assimilated using the measurement value r2 at the second time t2 is determined at the third time t3. That is, the effect of the assimilation result data assimilated at a predetermined time is determined at a time after the predetermined time.

[0085] The spatial state prediction system 1A of this embodiment includes an information acquisition unit 520 that acquires measurement values of a plurality of sensors 200 arranged at a plurality of positions in the space 10, and a plurality of positions in the space 10 at the first time t1. The assimilation result data a11 is obtained by performing data assimilation processing on the simulated state quantity f11m using the measurement values r1 of the plurality of sensors 200 at the first time t1, and the assimilation result data a11 is input into the prediction model of the state quantity in the space 10 and time-developed. An assimilation unit 510A that predicts the time-developed state quantity f12 after assimilation, which is the state quantity at a predetermined position in the space 10 at the second time t2, and without performing data assimilation processing, the simulated state quantity f11 at a predetermined position at the first time t1 is input into the above prediction model and time-developed. A simulation unit 530 that predicts the non-assimilated time-developed state quantity f22, which is the state quantity at a predetermined position at the second time t2, and based on the time-developed state quantity f12 after assimilation, the non-assimilated time-developed state quantity f22, and the measurement value r2 of the sensor 200 at the second time t2, a determination unit 540 that determines the appropriateness of the position of the sensor 200 arranged at a predetermined position.

[0086] As in this system, by determining the appropriateness of the position of the sensor 200 based on the time-developed state quantity f12 after assimilation in which data assimilation processing has been performed, the non-assimilated time-developed state quantity f22 in which data assimilation processing has not been performed, and the measurement value r2 of the sensor 200 at the second time t2, the sensor 200 with an inappropriate position can be moved to another position different from the current position, for example, a position where the state of the space 10 can be appropriately grasped. As a result, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately grasped.

[0087] [Spatial State Prediction Method of Example 1] The spatial state prediction method of Example 1 will be described with reference to FIG. 9.

[0088] FIG. 9 is a flowchart showing the spatial state prediction method of Example 1.

[0089] The spatial state prediction system 1A acquires the measurement values of the sensor 200 and information regarding the current position (step S101). At this point, the position of the sensor 200 is in a fixed state.

[0090] Next, the spatial state prediction system 1A predicts the value of the assimilated time-evolving state quantity f12 at the second time t2 (step S102A). Specifically, the spatial state prediction system 1A performs data assimilation processing on the simulation state quantities f11m at a plurality of positions in the space 10 at the first time t1 using the measurement values r1 of the plurality of sensors 200 at the first time t1 to obtain assimilation result data a11. Then, by inputting the assimilation result data a11 into the prediction model of the state quantity in the space 10 and time-evolving it, the assimilated time-evolving state quantity f12, which is the state quantity at a predetermined position in the space 10 at the second time t2, is predicted.

[0091] In addition, the spatial state prediction system 1A predicts the value of the non-assimilated time-evolving state quantity f22 at the second time t2 (step S102B). Specifically, the spatial state prediction system 1A inputs the simulation state quantity f11 at a predetermined position at the first time t1 into the above prediction model and time-evolves it without performing data assimilation processing, thereby predicting the non-assimilated time-evolving state quantity f22, which is the state quantity at a predetermined position at the second time t2.

[0092] The period from the first time t1 to the second time t2 is a period appropriately selected from, for example, the range of 10 minutes or more and 50 minutes or less. The state quantity of the space 10 is, for example, the temperature, flow rate, gas concentration, etc. inside the space 10. The flow rate includes the flow rate per unit time and the direction of the air flow. Step S102B may be executed before step S102A, or may be executed simultaneously with step S102A.

[0093] Further, the space state prediction system 1A acquires the measured value r2 of the sensor 200 at the second time t2 (step S103).

[0094] Next, the space state prediction system 1A determines the suitability of the position of the sensor 200 arranged at a predetermined position based on the assimilated time-developed state quantity f12 and the non-assimilated time-developed state quantity f22, which are predicted values, and the measured value r2 of the sensor 200, which is a measured value. For example, the space state prediction system 1A determines whether the assimilated time-developed state quantity f12 is closer to the measured value r2 of the sensor 200 than the non-assimilated time-developed state quantity f22 when compared at the second time t2 (step S104).

[0095] If the assimilated time-developed state quantity f12 is farther from the measured value r2 of the sensor 200 than the non-assimilated time-developed state quantity f22 (No in S104), it means that the sensor 200 is arranged at an inappropriate position. In this case, the space state prediction system 1A moves the sensor 200 to another position using the sensor mover 250 (step S105). Then, the processing after step S101 is executed.

[0096] On the other hand, if the assimilated time-developed state quantity f12 is closer to the measured value r2 of the sensor 200 than the non-assimilated time-developed state quantity f22 (Yes in S104), it means that the sensor 200 is arranged at an appropriate position. In this case, the space state prediction system 1A does not move the sensor 200 and proceeds to the next step.

[0097] Next, the spatial state prediction system 1A predicts the value of the time-evolved state quantity f13 at the third time t3 after assimilation (step S106). Specifically, the spatial state prediction system 1A obtains the assimilation result data a12 by performing data assimilation processing on the simulated state quantities f12m at a plurality of positions in the space 10 at the second time t2 using the measurement values r2 of the plurality of sensors 200 at the second time t2. Then, by inputting the assimilation result data a12 into the prediction model of the state quantity in the space 10 and evolving it over time, the time-evolved state quantity f13 after assimilation, which is the state quantity at a predetermined position at the third time t3, is predicted.

[0098] Also, the spatial state prediction system 1A obtains the measurement value r3 at the third time t3 (step S107).

[0099] Next, the spatial state prediction system 1A determines whether to review the position of the sensor 200 based on the value of the time-evolved state quantity f13 after assimilation at the third time t3 and the measurement value r3 of the sensor 200 at the third time t3. For example, the spatial state prediction system 1A determines whether the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 at the third time t3 is within a predetermined control value (step S108).

[0100] If the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 is within the predetermined control value (Yes in S108), the spatial state prediction system 1A does not review the position of the sensor 200 and returns to step S106.

[0101] On the other hand, if the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 is greater than the predetermined control value (No in S108), the spatial state prediction system 1A reviews the position of the sensor 200 and thus executes the processing after step S101 again. In the spatial state prediction method of the first embodiment, these steps S101 to S108 are repeatedly executed.

[0102] As described above, the spatial state prediction method includes a measurement value acquisition step of acquiring measurement values of a plurality of sensors 200 arranged at a plurality of positions in the space 10, and a simulated state quantity f11m at a plurality of positions in the space 10 at the first time t1 is obtained by performing data assimilation processing using the measurement values r1 of the plurality of sensors 200 at the first time t1, and the assimilation result data a11 is obtained. By inputting the assimilation result data a11 into the prediction model of the state quantity in the space 10 and allowing it to evolve over time, an assimilated time-evolved state quantity f12, which is the state quantity at a predetermined position in the space 10 at the second time t2, is predicted in the post-assimilation prediction step. Without performing data assimilation processing, the simulated state quantity f11 at a predetermined position at the first time t1 is input into the above prediction model and allowed to evolve over time, so as to predict an unassimilated time-evolved state quantity f22, which is the state quantity at a predetermined position at the second time t2, in the post-non-assimilation prediction step. And a determination step of determining the appropriateness of the position of the sensor 200 arranged at a predetermined position based on the assimilated time-evolved state quantity f12, the unassimilated time-evolved state quantity f22, and the measurement value r2 of the sensor 200 at the second time t2.

[0103] As in this method, by determining the appropriateness of the position of the sensor 200 based on the assimilated time-evolved state quantity f12 obtained by performing data assimilation processing and the unassimilated time-evolved state quantity f22 obtained without performing data assimilation processing, it becomes possible to move the sensor 200 with an inappropriate position to another position different from the current one, for example, a position where the state of the space 10 can be appropriately grasped. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately grasped.

[0104] [Spatial State Prediction Method of Example 2] The spatial state prediction method according to Example 2 will be described with reference to FIG. 10. In Example 2, an example of attempting data assimilation processing excluding the data of the sensor 200 with an inappropriate position will be described. Note that the configuration of the spatial state prediction system 1A in Example 2 is the same as that in Example 1.

[0105] FIG. 10 is a flowchart showing the spatial state prediction method according to the second embodiment.

[0106] The spatial state prediction system 1A acquires the measurement value of the sensor 200 and information regarding the current position (step S101).

[0107] Next, the spatial state prediction system 1A predicts the value of the assimilated time-evolved state quantity f12 at the second time t2 (step S102A). Further, the spatial state prediction system 1A predicts the value of the non-assimilated time-evolved state quantity f22 at the second time t2 (step S102B).

[0108] Also, the spatial state prediction system 1A acquires the measurement value r2 at the second time t2 (step S103).

[0109] Next, the spatial state prediction system 1A determines whether the assimilated time-evolved state quantity f12 is closer to the measurement value r2 of the sensor 200 than the non-assimilated time-evolved state quantity f22 when compared at the second time t2 (step S104).

[0110] If the assimilated time-evolved state quantity f12 is farther from the measurement value r2 of the sensor 200 than the non-assimilated time-evolved state quantity f22 (No in S104), it means that this sensor 200 is placed at an inappropriate position. In this case, the spatial state prediction system 1A excludes the data of the sensor 200 placed at this inappropriate position (step S110), and returns to execute the processes such as steps S101, S102, and S103. That is, the data assimilation process is executed again using other sensors 200 excluding the sensor 200 placed at the inappropriate position.

[0111] If the assimilated time-evolved state quantity f12 is closer to the measurement value r2 of the sensor 200 than the non-assimilated time-evolved state quantity f22 (Yes in S104), it means that the other sensors 200 are placed at appropriate positions, and the process proceeds to the next step.

[0112] Thus, in the spatial state prediction system 1A of the second embodiment, when it is determined that the position of a specific sensor 200 among the plurality of sensors 200 is inappropriate, data assimilation processing is performed using the measurement values of the other sensors 200 excluding the specific sensor 200.

[0113] The spatial state prediction system 1A acquires the time-evolving state quantity f12m after assimilation at an appropriate position and the measurement value of the sensor 200 arranged at an appropriate position (step S111).

[0114] Next, the spatial state prediction system 1A calculates the RMSE (Root Mean Squared Error) using the time-evolving state quantity f12m after assimilation at an appropriate position and the measurement value of the sensor 200 arranged at an appropriate position, and determines whether the value of this RMSE is less than or equal to a predetermined value (step S112).

[0115] If the value of the RMSE is greater than the predetermined value (No in S112), the position of the specific sensor 200 that was previously determined to be in an inappropriate position is moved (step S113). Then, the processing after step S101 is executed. If the value of the RMSE is large, it may not be suitable as data for performing assimilation processing only with the data of the other sensors 200. Therefore, in this example, the position of the specific sensor 200 in the inappropriate position is moved and added to the sensor group for data collection.

[0116] On the other hand, if the value of the RMSE is less than or equal to the predetermined value (Yes in S112), it is considered that assimilation processing can be performed with the data of the sensor 200 arranged at an appropriate position, and the process proceeds to the next step.

[0117] The spatial state prediction system 1A predicts the value of the time-evolved state quantity f13 at the third time t3 after assimilation (step S114). Specifically, the spatial state prediction system 1A obtains the assimilation result data a12 by performing data assimilation processing on the simulated state quantities f12m at a plurality of positions in the space 10 at the second time t2 using the measurement values r2 of the plurality of sensors 200 at the second time t2. Then, by inputting the assimilation result data a12 into the prediction model of the state quantity in the space 10 and evolving it over time, the time-evolved state quantity f13 after assimilation, which is the state quantity at a predetermined position at the third time t3, is predicted.

[0118] Also, the spatial state prediction system 1A obtains the measurement value r3 of the sensor 200 at the third time t3 (step S115).

[0119] Next, the spatial state prediction system 1A determines whether to review the position of the sensor 200 based on the value of the time-evolved state quantity f13 after assimilation at the third time t3 and the measurement value r3 of the sensor 200 at the third time t3. For example, the spatial state prediction system 1A determines whether the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 at the third time t3 is within a predetermined control value (step S116).

[0120] If the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 is within the predetermined control value (Yes in S116), the spatial state prediction system 1A does not review the position of the sensor 200 and returns to step S114.

[0121] On the other hand, if the difference between the time-evolved state quantity f13 after assimilation and the measurement value r3 of the sensor 200 is greater than the predetermined control value (No in S116), the spatial state prediction system 1A reviews the position of the sensor 200 and executes the processing after step S101 again. In the spatial state prediction method of the second embodiment, these steps S101 to S116 are repeatedly executed.

[0122] (Summary) An example of the spatial state prediction method and the like according to one aspect of the present disclosure is illustrated.

[0123] The spatial state prediction method of Example 1 is a method for performing data assimilation processing using the measurement values of the sensors 200 that detect the state quantities in the space 10 and predicting the state quantities in the space 10. The method includes a measurement value acquisition step of acquiring the measurement values of the plurality of sensors 200 arranged at a plurality of positions in the space 10, and a step of performing data assimilation processing on the simulated state quantities f11m at the plurality of positions in the space 10 at the first time t1 using the measurement values r1 of the plurality of sensors 200 at the first time t1 to obtain assimilation result data a11, and inputting the assimilation result data a11 into the prediction model of the state quantity in the space 10 and time-developing it to predict the post-assimilation time-developed state quantity f12, which is the state quantity at a predetermined position in the space 10 at the second time t2, a post-assimilation prediction step; a non-assimilation prediction step of predicting the non-assimilated time-developed state quantity f22, which is the state quantity at a predetermined position at the second time t2, by inputting the simulated state quantity f11 at a predetermined position at the first time t1 into the above prediction model and time-developing it without performing data assimilation processing; and a determination step of determining the appropriateness of the position of the sensor 200 arranged at the predetermined position based on the post-assimilation time-developed state quantity f12, the non-assimilated time-developed state quantity f22, and the measurement value r2 of the sensor 200 at the second time t2.

[0124] In this way, by determining the appropriateness of the position of the sensor 200 based on the post-assimilation time-developed state quantity f12 obtained by performing data assimilation processing, the non-assimilated time-developed state quantity f22 obtained without performing data assimilation processing, and the measurement value r2 of the sensor 200 at the second time t2, it becomes possible to move the sensor 200 with an inappropriate position to another position different from the current one, for example, a position where the state of the space 10 can be appropriately grasped. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately grasped.

[0125] The spatial state prediction method of Example 2 is the spatial state prediction method described in Example 1. In the determination step, when comparing at the second time t2, if the non-assimilated time evolution state quantity f22 is closer to the measurement value r2 of the sensor 200 than the assimilated time evolution state quantity f12, it may be determined that the position of the sensor 200 arranged at the predetermined position is inappropriate.

[0126] In this way, by determining the suitability of the position of the sensor 200 based on the assimilated time evolution state quantity f12, the non-assimilated time evolution state quantity f22, and the measurement value r2 at the second time t2, it becomes possible to move the sensor 200 with an inappropriate position to another position, for example, a position where the state of the space 10 can be appropriately captured. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured.

[0127] The spatial state prediction method of Example 3 is the spatial state prediction method described in Example 1 or 2, and may further include a step of moving the sensor 200 arranged at the predetermined position when it is determined that the position of the sensor 200 is inappropriate.

[0128] In this way, by moving the sensor 200 with an inappropriate position, it becomes possible to move the sensor 200 to a position where the state of the space 10 can be appropriately captured. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured. Also, by moving and using the sensor 200, it becomes possible to reduce the total number of sensors 200 arranged in the space 10.

[0129] The spatial state prediction method of Example 4 is the spatial state prediction method described in Example 1 or 2. In the determination step, when comparing at the second time t2, if the assimilated time evolution state quantity f12 is closer to the measurement value r2 of the sensor 200 than the non-assimilated time evolution state quantity f22, it is determined that the position of the sensor 200 arranged at the predetermined position is appropriate, and it may not be necessary to move the sensor 200 arranged at the predetermined position.

[0130] By not moving the sensor 200 with an appropriate position in this way, it is possible to predict the state quantity in the space 10 using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured.

[0131] The space state prediction method of Example 5 is the space state prediction method described in Example 1 or 2. In the measurement value acquisition step, the measurement values of a plurality of sensors 200 arranged at a plurality of positions in the space 10 are acquired. In the post-assimilation prediction step, a plurality of time-evolving state quantities after assimilation are predicted. In the non-assimilated prediction step, a plurality of time-evolving state quantities after non-assimilation are predicted. In the determination step, the suitability of the positions of the plurality of sensors 200 is determined. When it is determined that the position of a specific sensor 200 among the plurality of sensors 200 is inappropriate, data assimilation processing may be performed using the measurement values of the other sensors 200 excluding the specific sensor 200.

[0132] In this way, by excluding a specific sensor 200 with an inappropriate position and performing data assimilation processing using other sensors 200 with appropriate positions, it is possible to predict the state quantity in the space 10 using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured.

[0133] The space state prediction method of Example 6 is the space state prediction method described in Example 1 or 2. In the measurement value acquisition step, the measurement values of a plurality of sensors 200 arranged at a plurality of positions in the space 10 are acquired. In the post-assimilation prediction step, a plurality of time-evolving state quantities after assimilation are predicted. In the non-assimilated prediction step, a plurality of time-evolving state quantities after non-assimilation are predicted. In the determination step, the suitability of the positions of the plurality of sensors 200 is determined. When it is determined that the position of a specific sensor 200 among the plurality of sensors 200 is inappropriate, the RMSE is calculated using the measurement values of the other sensors 200 excluding the specific sensor 200. When the value of the RMSE is larger than a predetermined value, the specific sensor 200 may be moved.

[0134] In this way, when the value of RMSE is large, by moving a specific sensor 200 whose position is determined to be inappropriate, it becomes possible to move the sensor 200 to a position where the state of the space 10 can be appropriately captured. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured. Also, by moving and using the sensor 200, it becomes possible to reduce the total number of sensors 200 arranged in the space 10.

[0135] The space state prediction method of Example 7 is the space state prediction method described in any one of Examples 1 to 6, and the period from the first time t1 to the second time t2 may be a period of 10 minutes or more and 50 minutes or less.

[0136] According to this configuration, it can be determined whether the change in the state of the space 10 could be captured within an appropriate measurement time. Thereby, the state quantity in the space 10 can be predicted using the sensor 200 arranged at a position where the state of the space 10 can be appropriately captured.

[0137] The space state prediction method of Example 8 is the space state prediction method described in any one of Examples 1 to 6, and the sensor 200 may measure the temperature, flow rate, or gas concentration in the space 10.

[0138] According to this, it is possible to suppress a decrease in the prediction accuracy of the temperature, flow rate, or gas concentration, which are the state quantities of the space 10.

[0139] The spatial state prediction system 1A of Example 9 is a spatial state prediction system that performs data assimilation processing using the measurement values of sensors 200 that detect state quantities in space 10 and predicts the state quantities in space 10. It includes an information acquisition unit 520 that acquires the measurement values of a plurality of sensors 200 arranged at a plurality of positions in space 10, and assimilates the simulated state quantities f11m at a plurality of positions in space 10 at the first time t1 using the measurement values r1 of the plurality of sensors 200 at the first time t1 to obtain assimilation result data a11. By inputting the assimilation result data a11 into the prediction model of the state quantity in space 10 and evolving it over time, an assimilated time-evolved state quantity f12, which is the state quantity at a predetermined position in space 10 at the second time t2, is predicted. An assimilation unit 510A, a simulation unit 530 that predicts an unassimilated time-evolved state quantity f22, which is the state quantity at a predetermined position at the second time t2, by inputting the simulated state quantity f11 at a predetermined position at the first time t1 into the above prediction model and evolving it over time without performing data assimilation processing, and a determination unit 540 that determines the appropriateness of the position of the sensor 200 arranged at a predetermined position based on the assimilated time-evolved state quantity f12, the unassimilated time-evolved state quantity f22, and the measurement value r2 of the sensor 200 at the second time t2.

[0140] In this way, by determining the appropriateness of the position of the sensor 200 based on the assimilated time-evolved state quantity obtained by performing data assimilation processing, the unassimilated time-evolved state quantity without performing data assimilation processing, and the measurement value of the sensor 200 at the second time t2, it becomes possible to move the sensor 200 with an inappropriate position to another position different from the current one, for example, a position where the state of space 10 can be appropriately grasped. As a result, the state quantity in space 10 can be predicted using the sensor 200 arranged at a position where the state of space 10 can be appropriately grasped.

[0141] (Other embodiments) As described above, the spatial state prediction method and the like in the present disclosure have been described based on each example and the like. However, the present disclosure is not limited to these examples. As long as the gist of the present disclosure is not deviated from, various modifications conceived by those skilled in the art applied to each example, and other forms constructed by combining some components in each example are also included within the scope of the present disclosure.

[0142] For example, in the above, an example of moving one sensor 200 out of a plurality of sensors 200 has been shown. However, the number of sensors 200 to be moved is not limited to one. For example, when there are two or more sensors 200 with inappropriate positions, each sensor mobile body 250 equipped with each sensor 200 may be moved simultaneously.

Industrial Applicability

[0143] The spatial state prediction method and the like of the present disclosure can be applied to the use of predicting state quantities of a space such as the temperature of a space in a building.

Explanation of Signs

[0144] 1, 1A Spatial state prediction system 2 Network 10 Space 100 Computer 101 Input unit 102 Arithmetic circuit 103 Memory 104 Output unit 105 Storage unit 105a Program 105b Temporary data 106 Communication unit 200 Sensor 201 Positioning sensor 210 Moving device 220 Information acquisition unit 230 Control unit 240 Driving mechanism 250 Sensor mobile body 310 Information terminal 320 External information source 500, 500A Spatial state prediction device 510A Assimilation Unit 520 Information Acquisition Unit 530 Simulation Unit 540 Judgment Unit 550 Control Unit 600 Equipment a11, a12, a13 Assimilation Result Data f11, f11m State Variables f12, f12m, f13, f13m State Variables of Temporal Development after Assimilation f22, f23 State Variables of Temporal Development without Assimilation r1, r2, r3 Measurement Values t1 First Time t2 Second Time t3 Third Time

Claims

1. A method for performing data assimilation processing using measurement values of sensors that detect state quantities in a space and predicting the state quantities in the space, comprising: a measurement value acquisition step of acquiring measurement values of a plurality of the sensors arranged at a plurality of positions in the space; an assimilated prediction step of obtaining assimilation result data by performing data assimilation processing on the simulated state quantities at a plurality of positions in the space at a first time using the measurement values of the plurality of sensors at the first time, and inputting the assimilation result data into a prediction model of the state quantities in the space and evolving it over time to predict the state quantity of a predetermined position in the space at a second time after assimilation; a non-assimilated prediction step of predicting the non-assimilated time-evolved state quantity, which is the state quantity of the predetermined position at the second time, by inputting the simulated state quantity of the predetermined position at the first time into the prediction model and evolving it over time without performing the data assimilation processing; a determination step of determining the appropriateness of the position of the sensor arranged at the predetermined position based on the time-evolved state quantity after assimilation, the non-assimilated time-evolved state quantity, and the measurement value of the sensor at the second time; A space state prediction method including the above steps.

2. In the determination step, when comparing at the second time, if the non-assimilated time-evolved state quantity is closer to the measurement value of the sensor than the time-evolved state quantity after assimilation, it is determined that the position of the sensor arranged at the predetermined position is inappropriate. The space state prediction method according to Claim 1.

3. Furthermore, when it is determined that the position of the sensor is inappropriate, the method includes a step of moving the sensor arranged at the predetermined position. The space state prediction method according to Claim 1 or 2.

4. In the determination step, when comparing at the second time, if the time-evolved state quantity after assimilation is closer to the measurement value of the sensor than the non-assimilated time-evolved state quantity, it is determined that the position of the sensor arranged at the predetermined position is appropriate, and the sensor arranged at the predetermined position is not moved. The space state prediction method according to Claim 1.

5. In the measurement value acquisition step, measurement values of a plurality of the sensors arranged at a plurality of positions in the space are acquired. In the assimilated prediction step, a plurality of the time-evolved state quantities after assimilation are predicted. In the non-assimilated prediction step, a plurality of the non-assimilated time-evolving state variables are predicted. In the determination step, the suitability of the positions of the plurality of sensors is determined. When it is determined that the position of a specific sensor among the plurality of sensors is inappropriate, data assimilation processing is performed using the measurement values of the other sensors excluding the specific sensor. The spatial state prediction method according to claim 1 or 2.

6. In the measurement value acquisition step, measurement values of the plurality of sensors arranged at a plurality of positions in the space are acquired. In the post-assimilation prediction step, a plurality of the post-assimilation time-evolving state variables are predicted. In the non-assimilated prediction step, a plurality of the non-assimilated time-evolving state variables are predicted. In the determination step, the suitability of the positions of the plurality of sensors is determined. When it is determined that the position of a specific sensor among the plurality of sensors is inappropriate, RMSE (Root Mean Squared Error) is calculated using the measurement values of the other sensors excluding the specific sensor. When the value of the RMSE is greater than a predetermined value, the specific sensor is moved. The spatial state prediction method according to claim 1 or 2.

7. The period from the first time to the second time is a period of 10 minutes or more and 50 minutes or less. The spatial state prediction method according to claim 1 or 2.

8. The sensor measures the temperature, flow rate, or gas concentration in the space. The spatial state prediction method according to claim 1 or 2.

9. A spatial state prediction system that performs data assimilation processing using measurement values of sensors that detect state variables in a space and predicts the state variables in the space, comprising: An information acquisition unit that acquires measurement values of the plurality of sensors arranged at a plurality of positions in the space; An assimilation unit that acquires assimilation result data by performing data assimilation processing on the simulated state variables at a plurality of positions in the space at the first time using the measurement values of the plurality of sensors 200 at the first time, and inputs the assimilation result data into a prediction model of the state variables in the space and time-evolves the data to predict the post-assimilation time-evolving state variables, which are the state variables at a predetermined position in the space at the second time. A simulation unit that predicts a non-assimilated time-evolved state quantity, which is the state quantity at the predetermined position at the second time, by inputting the simulation state quantity at the predetermined position at the first time into the prediction model and evolving it over time without performing the data assimilation process; A determination unit that determines the suitability of the position of the sensor disposed at the predetermined position based on the assimilated time-evolved state quantity, the non-assimilated time-evolved state quantity, and the measured value of the sensor at the second time; A spatial state prediction system comprising the above.

Images