Information processing device, information processing system, information processing method, and program

Abstract

To assist in determining the extent of the impact of a fire or accident that occurred in a building. [Solution] The information processing device includes: an acquisition unit that acquires detection information detected by sensors placed in a building; a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building; an estimation unit that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition unit; a display information generation unit that generates display information that displays the structure of the building and displays the area of ​​influence estimated by the estimation unit in an identifiable manner; and a transmission unit that transmits the display information to a predetermined destination.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing system, an information processing method, and a program.

Background Art

[0002] Conventionally, technologies for assisting in response to disasters have been studied.

[0003] For example, Patent Document 1 discloses a technology that can recognize the situation of a disaster and the situation of measures taken against the disaster in real time at a support base, and by ensuring the coping ability to deal with the disaster at the support base, it is possible to reduce the complicated work at the disaster site and improve the coping efficiency for the disaster.

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the prior art, when a fire or an accident (such as gas leakage) occurs in a building, it is difficult to grasp the structure of the building and the scope of influence of the fire or the accident. As a result, it is conceivable that the influence of the fire or the accident in the building may expand.

[0005] The present invention has been made in view of the above points, and an object thereof is to assist in grasping the scope of influence of a fire or an accident that has occurred in a building.

Means for Solving the Problems

[0006] To solve the above problems, the information processing device includes: an acquisition unit that acquires detection information detected by sensors placed in the building; a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building; an estimation unit that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition unit; a display information generation unit that generates display information that displays the structure of the building and displays the area of ​​impact estimated by the estimation unit in an identifiable manner; and a transmission unit that transmits the display information to a predetermined destination. [Effects of the Invention]

[0007] This can help in determining the extent of the impact of a fire or accident that occurs in a building. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an example of the configuration of the information processing system in the first embodiment. [Figure 2] This figure shows an example of the hardware configuration of the information processing device 10 in the first embodiment. [Figure 3] This figure shows an example of the functional configuration of the information processing device 10 in the first embodiment. [Figure 4] This figure shows the input and output of the machine learning model m1 in the first embodiment. [Figure 5] This is a flowchart illustrating an example of a processing procedure performed by the information processing device 10 in the first embodiment. [Figure 6] This figure shows an example of the display information in the first embodiment. [Figure 7] This is a flowchart illustrating an example of a processing procedure performed by the information processing device 10 in the second embodiment. [Figure 8] This figure shows an example of the display information in the second embodiment. [Figure 9] This figure shows the input and output of the machine learning model m1 in the second embodiment. [Figure 10]This figure shows the input and output of the machine learning model m1 in the third embodiment. [Figure 11] This figure shows an example of the functional configuration of the information processing device 10 in the fourth embodiment. [Figure 12] This is a flowchart illustrating an example of a processing procedure performed by the information processing device 10 in the fourth embodiment. [Figure 13] This figure shows an example of the display information in the fourth embodiment. [Modes for carrying out the invention]

[0009] Embodiments of the present invention will be described below with reference to the drawings. Figure 1 is a diagram showing an example of the configuration of an information processing system in the first embodiment. In Figure 1, one or more sensors 20 located in building B1 are connected to the information processing device 10 via a network such as the Internet. In addition, one or more terminals 40 can be connected to the information processing device 10 via a network such as the Internet.

[0010] Building B1 is a structure that sees a large number of people coming and going, such as a hospital, commercial facility, office building, mixed-use complex, or educational facility. Building B1 may also be a high-rise or large-scale facility.

[0011] Sensor 20 is an element or device capable of measuring the conditions inside building B1. In this embodiment, it is sufficient to detect the occurrence of an accident such as a fire or gas leak (hereinafter referred to as "fire, etc.") occurring inside building B1, so sensor 20 only needs to be capable of measuring conditions that change in response to the occurrence of a fire, etc. For example, sensor 20 can measure the air conditions inside building B1. Air conditions include, for example, temperature, smoke, gas concentration, etc. Sensor 20 does not have to be of one type, but may be of multiple types. Also, multiple sensors 20 may be distributed and placed inside building B1.

[0012] The information processing device 10 is one or more computers that generate display information indicating the situation of a fire or other incident occurring in building B1, based on information (hereinafter referred to as "detection information") including values ​​(for example, temperature, smoke, gas concentration, etc.) detected (measured) by the sensor 20.

[0013] Terminal 40 is a smartphone, tablet, or dedicated terminal used by persons who perform firefighting or rescue operations in response to a fire or other incident (for example, firefighters or rescue workers). Hereinafter, the user of Terminal 40 (a person who performs firefighting or rescue operations) will be referred to as the terminal user. Terminal 40 receives display information generated by the information processing device 10 and displays the said display information.

[0014] Figure 2 shows an example of the hardware configuration of the information processing device 10 in the first embodiment. As shown in Figure 2, the information processing device 10 is built by a computer and includes a CPU 101, ROM 102, RAM 103, HD 104, HDD (Hard Disk Drive) controller 105, display 106, external device connection I / F (Interface) 108, network I / F 109, data bus 110, keyboard 111, pointing device 112, DVD-RW (Digital Versatile Disk Rewritable) drive 114, and media I / F 116.

[0015] Among these, the CPU 101 controls the operation of the entire information processing apparatus 10. The ROM 102 stores programs used for driving the CPU 101 such as the IPL. The RAM 103 is used as a work area for the CPU 101. The HD 104 stores various data such as programs. The HDD controller 105 controls the reading or writing of various data to and from the HD 104 according to the control of the CPU 101. The display 106 displays various information such as a cursor, menu, window, characters, or images. The external device connection I / F 108 is an interface for connecting various external devices. External devices in this case are, for example, a USB (Universal Serial Bus) memory, a printer, and the like. The network I / F 109 is an interface for data communication using a communication network. The data bus 110 is an address bus, a data bus, etc. for electrically connecting each component such as the CPU 101 shown in FIG. 2.

[0016] Also, the keyboard 111 is a type of input means having a plurality of keys for inputting characters, numerical values, various instructions, and the like. The pointing device 112 is a type of input means for selecting and executing various instructions, selecting a processing target, moving a cursor, and the like. The DVD-RW drive 114 controls the reading or writing of various data to and from the DVD-RW 113 as an example of a removable recording medium. Note that it is not limited to DVD-RW, and it may be DVD-R or the like. The media I / F 116 controls the reading or writing (storage) of data to and from the recording medium 115 such as a flash memory.

[0017] FIG. 3 is a diagram showing a functional configuration example of the information processing apparatus 10 in the first embodiment. In FIG. 3, the information processing apparatus 10 includes a machine learning model m1, a learning unit 11, an acquisition unit 12, a determination unit 13, an estimation unit 14, a display information generation unit 15, and a transmission unit 16. Each of these units is realized by a process executed by the CPU 101 for one or more programs installed in the information processing apparatus 10. The information processing apparatus 10 also uses storage units such as a learning data storage unit 121 and a building information storage unit 122. Each of these storage units can be realized, for example, using an HD 104 or a storage device that can be connected to the information processing apparatus 10 via a network.

[0018] The building information storage unit 122 stores map information of the building B1. The map information is a map (floor plan) inside the building showing the internal structure of the building B1 and the layout of stores and the like included in the building. The map information may be represented in three dimensions or may not be represented in three dimensions (for example, it may be represented in two dimensions). When the map information is not represented in three dimensions, the following "three dimensions" may be replaced with the dimension corresponding to the map information. For example, the map information is point cloud data obtained by scanning the inside of the building B1 with a three-dimensional scanner. Point cloud data refers to a collection of points each having three-dimensional coordinate values. The map information may further be data capable of synthesizing an image inside the building B1 with respect to the point cloud data. By using the map information, the inside of the building B1 where a fire or the like has occurred can be reproduced on the virtual space of the information processing apparatus 10 by digital twin technology. In the present embodiment, the three-dimensional coordinate system inside the building B1 for expressing the map information is referred to as a "map coordinate system". That is, each point of the point cloud data as the map information has three-dimensional coordinate values in the map coordinate system. Hereinafter, "position information" refers to information represented by coordinate values in the map coordinate system unless otherwise specified.

[0019] Figure 4 shows the input and output of the machine learning model m1 in the first embodiment. As shown in Figure 4, the machine learning model m1 is a machine learning model (e.g., a neural network) that takes a sequence of detected information as input and outputs the area (estimated) affected by the fire, etc., in building B1. Machine learning is a technology that enables computers to acquire human-like learning abilities. It is a technology in which a computer autonomously generates algorithms necessary for judgments such as data identification from pre-acquired training data, and applies these to new data to make predictions.

[0020] The area affected by a fire or other disaster (hereinafter referred to as the "affected area") refers to the location of the fire or disaster, or the area affected by fire, smoke, or gas. The affected area is represented by a set of unit spaces that divide the space represented by the map information into multiple parts, for example, in a mesh-like manner. Therefore, each unit space has positional information based on the map coordinate system. If the map information is 3-dimensional, the unit space is 3-dimensional, and if the map information is 2-dimensional, the unit space is 2-dimensional. For example, the output from the machine learning model m1 may be a vector containing the dimensions corresponding to each of the unit spaces in total, and the presence or absence of influence in each unit space may be indicated by 1 or 0. Alternatively, the value of each dimension of the vector may be the degree of influence of the fire or other disaster in the unit space corresponding to that dimension. The degree of influence is a real value in which a larger value indicates a greater influence, and 0 indicates no influence. Hereinafter, a vector containing the dimensions corresponding to each of the unit spaces, in which the presence or absence of influence or the degree of influence in each unit space is indicated by the value of each dimension, will be called a "spatial vector". The set of unit spaces corresponding to dimensions with values ​​greater than 0 in the spatial vector will be the affected area. The size of the unit space can be appropriately determined depending on the desired resolution for representing the area of ​​influence. For example, if the map information is point cloud data, one point may represent one unit space, or a set of multiple points may represent one unit space.

[0021] The detection information sequence refers to data in which the detection information of each sensor 20 is arranged in a predetermined order (hereinafter referred to as the "sensor order"). In other words, the Nth detection information in the detection information sequence is the detection information of the Nth sensor 20 in the sensor order. Each detection information may also include the position information of each sensor 20.

[0022] The learning unit 11 trains the machine learning model m1 using multiple training data stored in the training data storage unit 121. Each training data is a pair of detection information sequences and the influence range as the correct answer for those detection information sequences. The influence range is represented by the spatial vector described above. The learning unit 11 takes the detection information sequences contained in the training data as input and updates the parameters of the machine learning model m1 so that the output from the machine learning model m1 approaches the influence range (spatial vector) contained in that training data. By performing this learning process (parameter update) for multiple training data, the machine learning model m1 becomes a machine learning model that has learned the correspondence between detection information sequences and influence ranges.

[0023] For example, the creator of the training data may create the training data using the following procedure:

[0024] (1) Identify one or more locations in building B1 that could be the source of a fire or other incident (hereinafter referred to as "potential source locations"). For example, a location that uses gas, such as a restaurant, may be identified as a potential source location.

[0025] (2) For each potential source location, estimate the area affected if a fire or other incident occurs at that location. Multiple affected areas may be estimated for a single potential source location. If the degree of impact is to be considered, estimate the degree of impact for each affected area, including each unit space contained within that area.

[0026] (3) For each candidate source location and for each estimated result (affected area) in (2), the detection information of each sensor 20 is estimated.

[0027] (4) For each area of ​​influence in (2), training data is created by combining the estimated result of (3) for that area of ​​influence (estimated value of detection information from each sensor 20) and the area of ​​influence.

[0028] Furthermore, in the estimations of (2) and (3), it is desirable, from the standpoint of accuracy of the machine learning model m1, to consider the following information (hereinafter referred to as "building information") that may affect the fire situation, etc., with respect to building B1.

[0029] • Layout of Building B1 • Materials, width, ceiling height, etc. of building B1 • Information included in the building specifications, such as interior and exterior finishes and foundation work specifications. • The type and location of specified items stored or placed in Building B1 (for example, items containing highly flammable substances (chemicals), or equipment that could exacerbate the effects of a fire, such as gas pipes) (for example, ledger information of items stored or placed in Building B1) • Information regarding the electrical wiring in building B1 In other words, the way a fire or other disaster spreads can vary depending on this building information. By considering this building information, the machine learning model m1 can learn a correspondence between the detected information sequence and the affected area, taking into account the differences in how fires and other disasters spread according to the building information.

[0030] Furthermore, the results of computer simulations may be used for the estimations in (2) and (3).

[0031] The acquisition unit 12 acquires (receives) detection information from each sensor 20 at multiple timings (for example, periodically). The acquisition unit 12 may acquire detection information under normal circumstances, or it may start acquiring detection information from each sensor 20 in response to the activation (sounding) of a fire alarm or the like (i.e., after a fire or the like has occurred).

[0032] The determination unit 13 determines whether or not a fire or the like has occurred based on the detection information from each sensor 20 acquired by the acquisition unit 12. However, if the acquisition unit 12 starts acquiring detection information from each sensor 20 in response to the activation (sounding) of a fire alarm or the like (i.e., after a fire or the like has occurred), the information processing device 10 does not need to include the determination unit 13.

[0033] The estimation unit 14 uses a machine learning model m1, which is a machine learning model that has learned the correspondence between the detection information sequence and the area affected by the fire or accident in building B1 (affected area), and the detection information acquired by the acquisition unit 12 from each sensor 20 to estimate the area affected by the fire or accident in building B1 (affected area). If the machine learning model m1 outputs the degree of impact for each unit space, the estimation unit 14 will also estimate the geographical distribution of the degree of impact within the affected area. The geographical distribution of the degree of impact (hereinafter simply referred to as "distribution") refers to the degree of dispersion of the degree of impact for each unit space included in the affected area.

[0034] The display information generation unit 15 generates display information that shows the structure of building B1 and displays the estimated range (influence range) estimated by the estimation unit 14 in an identifiable manner. Display information is information generated based on map information and influence range, for display on each terminal 40. For example, the display information generation unit 15 generates display information by superimposing information indicating the influence range (e.g., an image or figure) onto the map information stored in the building information storage unit 122. If the machine learning model m1 outputs the degree of influence for each unit space, and the estimation unit 14 also estimates the distribution of the degree of influence within the influence range, the display information generation unit 15 generates display information that also shows the said distribution within the influence range.

[0035] The transmission unit 16 transmits the display information generated by the display information generation unit 15 to a predetermined destination.

[0036] The following describes the processing procedures performed by the information processing device 10. Figure 5 is a flowchart illustrating an example of the processing procedures performed by the information processing device 10 in the first embodiment. Steps S110 to S130 are repeatedly performed even under normal circumstances (when no fire or other incident occurs).

[0037] In step S110, the acquisition unit 12 acquires detection information from each sensor 20.

[0038] Next, the determination unit 13 determines whether or not a fire or other incident has occurred based on the detection information from each sensor 20 acquired by the acquisition unit 12 (S120). For example, the determination unit 13 compares each detection information with a threshold value, and if any of the detection information exceeds the threshold value, it determines that a fire or other incident has occurred.

[0039] If it is determined that no fire or other incident has occurred (No in S130), the process from step S110 onwards is repeated after a certain waiting period.

[0040] If it is determined that a fire or other incident has occurred (Yes in S130), the estimation unit 14 estimates the affected area of ​​the fire or incident based on the output from the machine learning model m1, which is input to a sequence of detection information obtained in step S110 arranged in the order of the sensors (S140). For example, the estimation unit 14 estimates the affected area as the range formed by the set of unit spaces in which the output value for each unit space from the machine learning model m1 is greater than 0. If the machine learning model m1 outputs an impact level for each unit space, the estimation unit 14 will also estimate the distribution of impact levels within the affected area.

[0041] Next, the display information generation unit 15 generates display information (S150) by superimposing information indicating the area of ​​influence estimated by the estimation unit 14 (e.g., an image or a figure) onto the map information stored in the building information storage unit 122. The display information may be in a general format, such as a web page, or in a format specific to the application installed on the terminal 40. If the machine learning model m1 outputs an influence level for each unit space, and the estimation unit 14 also estimates the distribution of influence levels within the area of ​​influence, the display information generation unit 15 generates display information that also shows this distribution within the area of ​​influence. For example, the display information generation unit 15 changes the display mode of each unit space within the area of ​​influence using a color or the like according to the influence level. In this case, the display information generation unit 15 may quantize each unit space into N levels based on an N-level threshold for the influence level. In this case, there can be up to N types of display modes according to the influence level.

[0042] Next, the transmission unit 16 transmits the display information generated by the display information generation unit 15 to a predetermined destination (S160). The predetermined destination is, for example, a terminal 40. Each terminal 40 that receives the display information displays the display information. The destination information (address information) of each terminal 40 to which the display information is to be transmitted may be stored in advance in the information processing device 10, but the terminals 40 may be unspecified. Therefore, the transmission unit 16 may transmit the display information to a predetermined URL. The terminals 40 may download the display information by accessing the URL.

[0043] Figure 6 shows an example of display information in the first embodiment. Figure 6 shows an example in which the affected area is represented by a frame line L1. The area within frame line L1 may be filled with a semi-transparent color. Furthermore, if the estimation unit 14 also estimates the distribution of the degree of influence within the affected area, the area within frame line L1 may be color-coded according to the degree of influence. For convenience, an example of the position of sensor 20 is also shown by a rectangle in Figure 6, but the figure indicating the position of sensor 20 does not have to be included in the display information.

[0044] According to this display information, terminal users can see at a glance the extent of the impact of a fire or other incident occurring in building B1. Therefore, according to the first embodiment, it is possible to help understand the extent of the impact of a fire or accident that has occurred in a building.

[0045] As a result, for example, firefighting measures can be quickly planned based on the situation as assessed. Furthermore, firefighting operations can be carried out efficiently while ensuring the safety of firefighters and people inside the building. This is expected to reduce both human casualties and property damage.

[0046] Next, a second embodiment will be described. The differences between the second embodiment and the first embodiment will be described. Therefore, points not specifically mentioned may be the same as in the first embodiment.

[0047] In the second embodiment, the building information storage unit 122 further stores location information (hereinafter referred to as "hazardous materials location information") for each of the specified items stored or placed in building B1 (for example, items containing highly flammable substances (chemicals), or equipment that may exacerbate the effects of fire, such as gas pipes).

[0048] Furthermore, in the second embodiment, the estimation unit 14 estimates the locations where a predetermined item is located within a predetermined range outside the affected range, based on the affected range estimated based on the detection information acquired by the acquisition unit 12 and the hazardous material placement information stored in the building information storage unit 122. The predetermined range outside the affected range is, for example, an area adjacent to the affected range and having a predetermined extent. Hereinafter, the locations estimated by the estimation unit 14 will be referred to as "hazardous locations".

[0049] A hazardous area is a location that is not currently affected by a fire or other disaster, but contains items that could easily exacerbate the fire's impact, and is close to the affected area, thus posing a potential danger to human health if approached. Hazardous areas are represented, for example, by their coordinate values ​​in the map coordinate system.

[0050] Figure 7 is a flowchart illustrating an example of a processing procedure performed by the information processing device 10 in the second embodiment. In Figure 7, the same step numbers are used for steps identical to those in Figure 5, and their explanations are omitted. In Figure 7, step S141 is added after step S140.

[0051] In step S141, the estimation unit 14 estimates hazardous locations based on the affected area estimated in step S140 and the hazardous material placement information stored in the building information storage unit 122. Specifically, the estimation unit 14 identifies locations within a predetermined range outside (around) the affected area that include one of the placement locations for each item indicated by the hazardous material placement information.

[0052] In step S150, the display information generation unit 15 further generates display information that allows for the identification of the hazardous locations estimated by the estimation unit 14. For example, the display information may be generated by superimposing a figure indicating the hazardous locations onto the map information.

[0053] Therefore, in the second embodiment, such display information is displayed on each terminal 40.

[0054] Figure 8 shows an example of the display information in the second embodiment. In Figure 8, the same parts as in Figure 6 are assigned the same reference numerals. In Figure 8, the hazardous area is surrounded by a dashed border line L2.

[0055] Based on this displayed information, terminal users can also identify dangerous areas. As a result, terminal users can avoid dangerous areas when carrying out firefighting or rescue operations. Consequently, the occurrence of harm to terminal users can be reduced. In particular, the presence and arrangement of hazardous materials such as chemicals at the site of a fire can sometimes lead to counterproductive results from conventional firefighting methods, but this embodiment is expected to increase the likelihood of suppressing such counterproductive results.

[0056] Furthermore, the machine learning model m1 may be trained to also estimate dangerous areas. In this case, the input and output of the machine learning model m1 could be configured as follows.

[0057] Figure 9 shows the input and output of the machine learning model m1 in the second embodiment. As shown in Figure 9, the machine learning model m1 takes a series of detection information and hazardous material placement information as input and outputs a series of affected areas and hazardous locations. The hazardous locations series is an array of hazardous locations.

[0058] One set of training data for such a machine learning model m1 is a combination of a detection information sequence, hazardous material placement information, the correct impact range for the detection information sequence, and a correct hazardous location sequence corresponding to the impact range. The correct hazardous location sequence corresponding to the impact range is an array of locations within a predetermined range outside the impact range that include any of the placement locations indicated by the hazardous material placement information.

[0059] The learning unit 11 updates the parameters of the machine learning model m1 for each set of training data so that the output from the machine learning model m1, which is input to the detection information sequence and hazardous material placement information contained in the training data, approaches the influence range (spatial vector) and hazardous location sequence contained in the training data. As a result, the machine learning model m1 becomes a machine learning model that has learned the correspondence between the detection information sequence and hazardous material placement information and the influence range and hazardous locations.

[0060] When the machine learning model m1 is trained in this manner, steps S140 and S141 are realized in a single step. That is, the estimation unit 14 estimates the affected area of ​​the fire, etc., and the hazardous locations based on the output from the machine learning model m1, which has been input with detection information indicating that a fire, etc., has occurred and information on the placement of hazardous materials.

[0061] Next, a third embodiment will be described. The differences between the third embodiment and the first embodiment will be described. Therefore, points not specifically mentioned may be the same as in the first embodiment.

[0062] In the third embodiment, the input and output of the machine learning model m1 differ from those in the first embodiment. Figure 10 shows the input and output of the machine learning model m1 in the third embodiment.

[0063] As shown in Figure 10, in the third embodiment, the machine learning model m1 takes a sequence of detection information as input and outputs information on the scope of impact and the method of digestion.

[0064] Information regarding fire extinguishing methods refers to information about fire extinguishing methods for fires corresponding to the detected information sequence. For example, information regarding fire extinguishing methods may include the type of fire extinguisher. That is, just as the appropriate fire extinguisher differs depending on the type of fire (the burning object), it is thought that the detected information sequence differs depending on the type of fire. Therefore, in the third embodiment, the machine learning model m1 also learns the correspondence between the detected information sequence and the information regarding fire extinguishing methods.

[0065] One training dataset for such a machine learning model m1 is a set of data consisting of a sequence of detection information, the correct range of influence for that sequence of detection information, and information on the correct fire extinguishing method for that sequence of detection information.

[0066] The learning unit 11 updates the parameters of the machine learning model m1 for each set of training data so that the output from the machine learning model m1, which is input to the detection information sequence contained in the training data, approaches the information regarding the affected area (spatial vector) and fire extinguishing method contained in the training data. As a result, the machine learning model m1 becomes a machine learning model that has learned the correspondence between the detection information sequence and the information regarding the affected area and fire extinguishing method.

[0067] In this case, in step S140 of Figure 5, the estimation unit 14 estimates the affected area of ​​the fire and the firefighting method based on the output from the machine learning model m1, which has been input with the detection information sequence.

[0068] In step S150, the display information generation unit 15 generates display information that includes information about the digestion method estimated by the estimation unit 14.

[0069] As a result, in the third embodiment, information regarding fire extinguishing methods is also displayed on each terminal 40. Therefore, firefighting activities can be carried out using a fire extinguishing method appropriate to the fire that is occurring.

[0070] Furthermore, information regarding fire extinguishing methods may include information other than the type of fire extinguisher, such as the method of water discharge.

[0071] Furthermore, the third embodiment may be combined with the second embodiment.

[0072] Next, a fourth embodiment will be described. The differences between the fourth embodiment and the first embodiment will be described. Therefore, points not specifically mentioned may be the same as in the first embodiment.

[0073] Figure 11 shows an example of the functional configuration of the information processing device 10 in the fourth embodiment. In Figure 11, the information processing device 10 further includes a route generation unit 17. The route generation unit 17 is realized by a process that one or more programs installed in the information processing device 10 cause the CPU 101 to execute.

[0074] The route generation unit 17 generates a route from a predetermined location in building B1 to a point in a unit space identified based on the influence level (influence level for each unit space) estimated by the estimation unit 14. The unit space identified based on the influence level is a unit space with a relatively high influence level (for example, the top N units). A relatively high influence level is considered to indicate a high probability of being the source of a fire or other incident. The predetermined location in building B1 is, for example, an entrance or exit to building B1, or the location of a terminal user around building B1 (a place where firefighters are assembled). Therefore, the route generation unit 17 generates a route from outside building B1 to a point that is highly likely to be the source of a fire or other incident.

[0075] Route generation by the route generation unit 17 can be performed using known techniques. For example, data representing all passages in building B1 as a graph structure (hereinafter referred to as "graph data") is prepared in advance and stored in the building information storage unit 122. This graph data includes a set of nodes and a set of edges connecting the nodes. Each edge represents a passage, and the distance of the section corresponding to that edge in the passage is assigned to it. Each node is assigned coordinate values ​​in the map coordinate system. The route generation unit 17 searches for a route from a predetermined position as the starting point and a unit space identified based on the degree of influence as the ending point in the graph data. The search algorithm can use known techniques. In this case, edges that pass through dangerous areas or areas of influence are assigned a higher cost than other edges. For edges that pass through areas of influence, the greater the degree of influence of the unit space they pass through, the greater the cost assigned. By searching for a route that minimizes the sum of distance and cost, the route generation unit 17 can generate the shortest route to the source of a fire or other incident while avoiding danger as much as possible.

[0076] Figure 12 is a flowchart illustrating an example of a processing procedure performed by the information processing device 10 in the fourth embodiment. In Figure 12, the same step numbers are used for steps identical to those in Figure 5, and their explanations are omitted. In Figure 12, step S142 is added after step S140.

[0077] In step S142, the route generation unit 17 generates a route from a predetermined location (such as the exit of building B1) to a point in a unit space identified based on the degree of impact (the location of the fire, etc.) using the method described above. The location information of the predetermined location may be input by the terminal user when step S142 is executed.

[0078] In step S150, the display information generation unit 15 generates display information to show the route generated by the route generation unit 17. For example, the display information generation unit 15 generates the display information by superimposing a figure showing the route onto the map information.

[0079] Therefore, in the fourth embodiment, the route is also displayed on each terminal 40.

[0080] Figure 13 shows an example of the display information in the fourth embodiment. In Figure 13, the same parts as in Figure 8 are assigned the same reference numerals.

[0081] In Figure 13, when point P1 is identified as a unit space based on the degree of impact, the route from the entrance of building B1 to point P1 is shown by arrows r1, r2, and r3. With this kind of display information, each terminal user can obtain useful information for firefighting operations.

[0082] The fourth embodiment may be combined with either or both of the second and third embodiments.

[0083] Next, a fifth embodiment will be described. The differences between the fifth embodiment and the first embodiment will be described. Therefore, points not specifically mentioned may be the same as in the first embodiment.

[0084] In the fifth embodiment, a simulation of the transition of the affected area of ​​a fire or accident occurring in building B1 is pre-executed using a computer for each of several expected distributions of the degree of influence for each unit space. In this simulation, the structure of building B1 as shown in the map information and the building information described above are given as known parameters. In this situation, a simulation of the transition of the affected area of ​​a fire or accident is executed for each of several expected distributions of the degree of influence. As for the multiple expected distributions of the degree of influence, for example, the degree of influence of each unit space included in the ground truth of the affected area of ​​each of the multiple training data sets of the machine learning model m1 may be used. Note that the transition of the affected area of ​​a fire or accident refers to how the affected area of ​​a fire or accident spreads over time. The simulation results obtained for each distribution of the degree of influence are stored in the building information storage unit 122 in association with the distribution of the degree of influence.

[0085] In this case, in step S150 of Figure 5, the display information generation unit 15 generates display information that shows the simulation results corresponding to the influence distribution estimated by the estimation unit 14, from among the simulation results stored in the building information storage unit 122 for each influence distribution. Here, the simulation results corresponding to the influence distribution estimated by the estimation unit 14 refer to the simulation results corresponding to the same distribution as the influence distribution estimated by the estimation unit 14, or the simulation results corresponding to the distribution that is most similar to the influence distribution estimated by the estimation unit 14.

[0086] For example, the display information generation unit 15 may generate the display information by superimposing an image showing the simulation results onto the map information. The image showing the simulation results refers to an image showing how the affected area spreads.

[0087] As a result, in the fifth embodiment, each terminal 40 displays predictive information about the process of changes in the future scope of impact, starting from the current scope of impact. Therefore, terminal users can carry out firefighting and rescue operations based on these future predictions.

[0088] The fifth embodiment may be combined with one or more of the second to fourth embodiments.

[0089] Furthermore, the information processing device 10 is not limited to a general-purpose server computer, as long as it is a device equipped with communication and computing functions. The information processing device 10 may be, for example, an output device such as a PJ (Projector), IWB (Interactive White Board: an electronic whiteboard with the ability to communicate with each other), or digital signage, a HUD (Head Up Display) device, industrial machinery, imaging devices, sound collection devices, medical equipment, networked home appliances, a notebook PC (Personal Computer), a mobile phone, a smartphone, a tablet device, a game console, a PDA (Personal Digital Assistant), a digital camera, a wearable PC, or a desktop PC.

[0090] Furthermore, each function of each embodiment can be realized by one or more processing circuits. Hereinafter, "processing circuit" as used herein includes processors programmed to execute each function by software, such as processors implemented by electronic circuits, as well as devices such as ASICs (Application Specific Integrated Circuits), DSPs (digital signal processors), FPGAs (field programmable gate arrays), and conventional circuit modules designed to execute the functions described above.

[0091] Furthermore, the apparatus described in each of the above embodiments represents only one of several computing environments for carrying out the embodiments disclosed herein.

[0092] In one embodiment, the information processing device 10 includes a plurality of computing devices, such as a server cluster. The plurality of computing devices are configured to communicate with each other via any type of communication link, including a network or shared memory, and perform the processing disclosed herein.

[0093] Although embodiments of the present invention have been described in detail above, the present invention is not limited to these specific embodiments, and various modifications and changes are possible within the scope of the gist of the present invention as described in the claims.

[0094] Examples of the present invention are as follows:

[0095] <1> An acquisition unit that acquires detection information detected by sensors placed in the building, An estimation unit that uses a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and the detection information acquired by the acquisition unit to estimate the area affected by a fire or accident in the building, A display information generation unit generates display information that displays the structure of the building in a way that allows the influence area estimated by the estimation unit to be identifiable, A transmission unit that transmits the aforementioned display information to a predetermined destination, An information processing device having

[0096] <2> The machine learning model has further learned the correspondence between the detection information and the degree of impact caused by a fire or accident in each unit space that divides the interior space of the building into multiple sections. The estimation unit further estimates the degree of influence for each unit space using the machine learning model and the detection information acquired by the acquisition unit. The display information generation unit generates the display information that shows the distribution of the degree of influence within the influence range estimated by the estimation unit. Characterized by <1> The information processing device described above.

[0097] <3> The estimation unit further identifies the location where the predetermined article is located within a predetermined area outside the said influence range, based on the influence range estimated based on the detection information acquired by the acquisition unit and the location information of the predetermined article in the building. The display information generation unit generates the display information so as to be able to identify the location identified by the estimation unit. Characterized by <1> or <2> The information processing device described above.

[0098] <4> The aforementioned machine learning model has further learned the correspondence between the detection information and the information regarding the digestion method. The estimation unit further estimates information regarding the digestion method using the machine learning model and the detection information acquired by the acquisition unit. The display information generation unit generates the display information which includes information about the digestion method estimated by the estimation unit. Characterized by <1> ~ <3> An information processing device as described in any of the following.

[0099] <5> A path generation unit generates a path from a predetermined location in the building to the unit space identified based on the degree of influence estimated by the estimation unit. It has, The display information generation unit generates the display information indicating the route. Characterized by <2> The information processing device described above.

[0100] <6> The display information generation unit generates the display information that shows the results of the simulation corresponding to the distribution of impact estimated by the estimation unit, from among the results of the simulation regarding the transition of the affected area due to fire or accident in the building, which was performed according to the expected distribution of impact. Characterized by <2> The information processing device described above.

[0101] <7> An information processing system including an information processing device and a terminal, The aforementioned information processing device is An acquisition unit that acquires detection information detected by sensors placed in the building, An estimation unit that uses a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and the detection information acquired by the acquisition unit to estimate the area affected by a fire or accident in the building, A display information generation unit generates display information that displays the structure of the building in a way that allows the influence area estimated by the estimation unit to be identifiable, A transmitting unit that transmits the aforementioned display information to the terminal, An information processing device having

[0102] <8> The procedure for acquiring detection information detected by sensors placed in the building, A machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and an estimation procedure that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition procedure, A display information generation procedure that generates display information that displays the structure of the building in a manner that allows for the identification of the area of ​​influence estimated by the estimation procedure, A transmission procedure for sending the aforementioned display information to a predetermined recipient, An information processing method characterized by a computer executing the following.

[0103] <9> The procedure for acquiring detection information detected by sensors placed in the building, A machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and an estimation procedure that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition procedure, A display information generation procedure that generates display information that displays the structure of the building in a manner that allows for the identification of the area of ​​influence estimated by the estimation procedure, A transmission procedure for sending the aforementioned display information to a predetermined recipient, A program that causes a computer to execute something. [Explanation of Symbols]

[0104] 10 Information Processing Devices 20 sensors 11. Learning Department 12 Acquisition Department 13 Judgment section 14 Estimation part 15 Display information generation section 16 Transmitter 17 Route generation unit 40 devices 121 Learning Data Storage Unit 122 Building Information Storage Unit B Bus m1 machine learning model [Prior art documents] [Patent Documents]

[0105] [Patent Document 1] Japanese Patent Publication No. 2003-77075

Claims

1. An acquisition unit that acquires detection information detected by sensors placed in the building, An estimation unit that uses a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and the detection information acquired by the acquisition unit to estimate the area affected by a fire or accident in the building, A display information generation unit generates display information that displays the structure of the building in a way that allows the influence area estimated by the estimation unit to be identifiable, A transmission unit that transmits the aforementioned display information to a predetermined destination, An information processing device having

2. The machine learning model has further learned the correspondence between the detection information and the degree of impact caused by a fire or accident in each unit space that divides the interior space of the building into multiple sections. The estimation unit further estimates the degree of influence for each unit space using the machine learning model and the detection information acquired by the acquisition unit. The display information generation unit generates the display information that shows the distribution of the degree of influence within the influence range estimated by the estimation unit. The information processing apparatus according to feature 1.

3. The estimation unit further identifies the location where the predetermined article is located within a predetermined area outside the said influence range, based on the influence range estimated based on the detection information acquired by the acquisition unit and the location information of the predetermined article in the building. The display information generation unit generates the display information so as to be able to identify the location identified by the estimation unit. The information processing apparatus according to feature 1.

4. The aforementioned machine learning model has further learned the correspondence between the detection information and the information regarding the digestion method. The estimation unit further estimates information regarding the digestion method using the machine learning model and the detection information acquired by the acquisition unit. The display information generation unit generates the display information which includes information about the digestion method estimated by the estimation unit. The information processing apparatus according to feature 1.

5. A path generation unit generates a path from a predetermined location in the building to the unit space identified based on the degree of influence estimated by the estimation unit. It has, The display information generation unit generates the display information indicating the route. The information processing apparatus according to feature 2.

6. The display information generation unit generates the display information that shows the results of the simulation corresponding to the distribution of impact estimated by the estimation unit, from among the results of the simulation regarding the transition of the affected area due to fire or accident in the building, which was performed according to the expected distribution of impact. The information processing apparatus according to feature 2.

7. An information processing system including an information processing device and a terminal, The aforementioned information processing device is An acquisition unit that acquires detection information detected by sensors placed in the building, An estimation unit that uses a machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and the detection information acquired by the acquisition unit to estimate the area affected by a fire or accident in the building, A display information generation unit generates display information that displays the structure of the building in a way that allows the influence area estimated by the estimation unit to be identifiable, A transmitting unit that transmits the aforementioned display information to the terminal, An information processing device having

8. The procedure for acquiring detection information detected by sensors placed in the building, A machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and an estimation procedure that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition procedure, A display information generation procedure that generates display information that displays the structure of the building in a manner that allows for the identification of the area of ​​influence estimated by the estimation procedure, A transmission procedure for sending the aforementioned display information to a predetermined recipient, An information processing method characterized by a computer executing the following.

9. The procedure for acquiring detection information detected by sensors placed in the building, A machine learning model that has learned the correspondence between the detection information and the area affected by a fire or accident in the building, and an estimation procedure that estimates the area affected by a fire or accident in the building using the detection information acquired by the acquisition procedure, A display information generation procedure that generates display information that displays the structure of the building in a manner that allows for the identification of the area of ​​influence estimated by the estimation procedure, A transmission procedure for sending the aforementioned display information to a predetermined recipient, A program that causes a computer to execute something.

Images