Spatial event identification system, spatial event identification method, and program

Abstract

To identify as many spatial events as possible with a small number of sensors SOLUTION: A spatial event identification system includes a sensor provided to be able to detect vibration acceleration generated in a plurality of identification target spaces in a building at a certain level or higher, a generation unit that generates a spectrogram indicating vibration acceleration with respect to a time axis and a frequency axis based on a detection output of the sensor, and an identification unit that identifies an event occurring in the plurality of identification target spaces and a position in the plurality of identification target spaces where the event has occurred based on the spectrogram generated by the generation unit.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a spatial event identification system, a spatial event identification method, and a program. [Background technology]

[0002] BACKGROUND ART There is known a technique for monitoring a subject's walking characteristics and falls based on the output of a sensor that is provided so as to be able to measure acceleration, vibration, deflection, etc. of a floor (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Special Publication No. 2006-525073 Summary of the Invention [Problem to be solved by the invention]

[0004] For example, a building may have multiple spaces such as rooms. When identifying events occurring in each of these multiple spaces, it is desirable to be able to identify events in as many spaces as possible using as few sensors as possible, as this simplifies the overall system configuration.

[0005] In view of the above, an object of the present invention is to make it possible to identify as many spatial phenomena as possible using a small number of sensors. [Means for solving the problem]

[0006] One aspect of the present invention that solves the above-mentioned problems is a spatial event identification system that includes a sensor that is configured to be able to detect vibration acceleration occurring in multiple spaces to be identified in a building at a certain level or above, a generation unit that generates a spectrogram that shows vibration acceleration on a time axis and a frequency axis based on the detection output of the sensor, and an identification unit that identifies an event that has occurred in the multiple spaces to be identified and the location in the multiple spaces to be identified where the event occurred based on the spectrogram generated by the generation unit.

[0007] One aspect of the present invention is a spatial event identification method in a spatial event identification system, which includes a generation step in which a generation unit generates a spectrogram showing vibration acceleration with respect to a time axis and a frequency axis based on the detection output of a sensor that is configured to be able to detect vibration acceleration occurring in a plurality of spaces to be identified in a building at a certain level or above, and an identification step in which an identification unit identifies an event that has occurred in a plurality of spaces to be identified and the position in the plurality of spaces to be identified where the event occurred, based on the spectrogram generated by the generation step.

[0008] One aspect of the present invention is a program for causing a computer in a spatial event identification system to function as a generation unit that generates a spectrogram showing vibration acceleration on a time axis and a frequency axis based on the detection output of a sensor that is configured to be able to detect vibration acceleration occurring in multiple spaces to be identified in a building at a certain level or above, and an identification unit that identifies an event that has occurred in multiple spaces to be identified and the location in the multiple spaces to be identified where the event occurred based on the spectrogram generated by the generation unit. [Effects of the Invention]

[0009] According to the present invention, it is possible to obtain an effect that it becomes possible to identify as many spatial phenomena as possible using one sensor. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram illustrating an example of the overall configuration of a spatial event identification system according to an embodiment of the present invention. [Figure 2] FIG. 10 is a diagram showing another example of the overall configuration of the spatial event identification system according to the present embodiment. [Figure 3] 10A and 10B are diagrams illustrating an example of vibration acceleration detected in each room for each event in the present embodiment. [Figure 4] FIG. 2 is a diagram illustrating an example of the functional configuration of a spatial event identification device according to the present embodiment. [Figure 5] FIG. 10 is a diagram showing an example of a processing procedure executed by the spatial event identification device in this embodiment in response to spatial event identification. [Figure 6] FIG. 1 is a diagram illustrating an example of the configuration of a trained model in this embodiment. [Figure 7] FIG. 10 is a diagram illustrating an example of the configuration of a trained model in a first modified example of this embodiment. [Figure 8] FIG. 10 is a diagram illustrating an example of the configuration of a trained model in a second modified example of this embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] <Embodiment> Fig. 1 shows an example of the overall configuration of a spatial event identification system according to this embodiment. The figure shows three rooms RM (RM-1, RM-2, and RM-3: examples of spaces to be identified) adjacent to each other on the same floor of a building BL (an example of a structure). A corresponding resident RS lives in each room RM and uses the facilities within. The spatial event identification system of this embodiment has the configuration described below and performs spatial event identification with three rooms RM as identification targets. Here, spatial event identification refers to identifying the type of event that has occurred and identifying the room RM in which the event occurred. In spatial event identification, the room RM in which the event has occurred may be identified down to its position within the room RM.

[0012] In the figure, the facilities provided in each room RM are a toilet 10, a sink 20, and a bathtub 30. For example, in each room RM, the toilet 10, the sink 20, and the bathtub 30 may be separated into different rooms, but the illustration of the room divisions and the like is omitted. In the following description, when no distinction is made between the toilet 10, the sink 20, and the bathtub 30 or when they are referred to collectively, they will be referred to as "equipment."

[0013] In the example shown in the figure, the drainage water from each piece of equipment in room RM-1 and room RM-2 is arranged to flow into drainage system 40-1, and the drainage water from each piece of equipment in room RM-3 is arranged to flow into drainage system 40-2, which is different from drainage system 40-1. The drainage systems 40-1 and 40-2 are each formed by installing drainage pipes within building BL.

[0014] The figure shows an example in which a sensor 100 is attached to a predetermined position of a drainage system 40-1. The sensor 100 is, for example, an acceleration sensor, and detects acceleration (vibration acceleration) corresponding to vibrations transmitted to the attachment position of the sensor 100.

[0015] In this embodiment, the sensor 100 is provided at a position where it can significantly detect vibrations corresponding to events occurring in each of the rooms RM-1, RM-2, and RM-3 that are the classification targets of the spatial event classification device 200. In other words, the sensor 100 is provided at a position where it can detect vibrations occurring in the rooms RM (RM-1, RM-2, RM-3). As an example of such a position, the figure shows an example where the sensor 100 is set to a predetermined position of the drainage system 40-1. In this case, the drainage system 40-2 to which the drainage of the equipment in the room RM-3 is connected is a separate drainage system from the drainage system 40-1, but vibrations caused by an event occurring in the room RM-3 are transmitted to the installation position of the sensor 100 via, for example, a wall or a pillar in the building BL. In this manner, in this embodiment, spatial event discrimination can be performed using one sensor with a predetermined number of rooms RM as discrimination targets.

[0016] As can be understood from the above description, the installation position of the sensor 100 may be any position that can detect vibrations in each room RM to be identified. Therefore, the detectable position may be, for example, a predetermined position in the drainage system 40-2, a floor or wall of the room RM, or a predetermined position on a structure other than the room RM, depending on the structure of the building BL and the relative positions of the rooms RM.

[0017] The spatial event discrimination device 200 uses the vibration acceleration detected by the sensor 100 to perform spatial event discrimination. The spatial event identification device 200 may be provided in the building BL, or may be provided in a predetermined facility other than the building BL. Furthermore, the spatial event identification device 200 may be provided as a server on a network, or as a cloud server, for example.

[0018] The notification terminal 300 is a terminal that outputs notification information based on spatial event identification by the spatial event identification device 200. For example, the spatial event identification system of this embodiment can be used to monitor residents in each room in a nursing home or the like. When such monitoring is performed, for example, in response to determining that a resident RS has fallen in a certain room RM through spatial event identification by the spatial event identification device 200, the notification terminal 300 may notify the monitor that the resident RS has fallen and the room RM of the resident RS.

[0019] Furthermore, in this embodiment, the positional relationship of rooms RM to be identified by the spatial event identification system is not limited to being on the same floor as illustrated in Fig. 1. For example, multiple rooms RM to be identified by the spatial event identification system may be on different floors.

[0020] Fig. 2 shows an example in which multiple rooms RM to be identified are located on different floors. In Fig. 2, the same parts as in Fig. 1 are given the same reference numerals and their description will be omitted. In Fig. 2 as well, three rooms RM (RM-1, RM-2, RM-3) are to be identified. In addition, Fig. 2 shows an example in which rooms RM-1 and RM-2 are located next to each other on the same floor, while room RM-3 is located on the floor below rooms RM-1 and RM-2 and directly below room RM-2.

[0021] In the example shown in the figure, the drainage water from the equipment in rooms RM-1 and RM-2 is combined into drainage system 40-1, and the drainage water from the equipment in room RM is combined into drainage system 40-2. As for sensor 100, a predetermined position in drainage system 40-1 is identified as a position to which vibrations corresponding to events occurring in each of rooms RM (RM-1, RM-2, RM-3) to be identified are transmitted, and an example is shown in which sensor 100 is provided at the predetermined position.

[0022] FIG. 3 shows an example of the vibration acceleration results detected by sensor 100 when a common type of event occurs in three rooms A, B, and C. Rooms A, B, and C do not necessarily have to be arranged in the same manner as rooms RM-1, RM-2, and RM-3 in FIG. 1 or 3, but are arranged in a predetermined positional relationship within the building. Each of rooms A, B, and C is equipped with a toilet 10, a sink 20, and a bathtub 30, similar to room RM in FIG. 1 and 3. Sensor 100 is installed in a position where it can detect vibrations caused by the occurrence of an event in each of rooms A, B, and C.

[0023] Figure 3(A) shows the vibration acceleration detected by sensor 100 over time when toilet 10 is flushed for each of rooms A, B, and C. In other words, Figure 3(A) shows the vibration acceleration detected in response to the occurrence of an event, such as flushing toilet 10, in each of rooms A, B, and C. The change in the detected vibration acceleration over time can be interpreted as a vibration waveform that indicates the vibration generated in response to the occurrence of an event.

[0024] 3(B) shows the vibration acceleration detected by the sensor 100 over time when water is drained from the sink 20 for each of the rooms A, B, and C. That is, FIG. 3(B) shows the vibration acceleration detected in response to the occurrence of an event such as draining water from the sink 20 in each of the rooms A, B, and C.

[0025] Figure 3(C) shows the vibration acceleration detected by sensor 100 over time when water is drained from bathtub 30 for each of rooms A, B, and C. In other words, Figure 3(C) shows the vibration acceleration detected in response to the occurrence of an event such as draining water from bathtub 30 in each of rooms A, B, and C.

[0026] Fig. 3(D) shows the vibration acceleration detected by the sensor 100 over time when the resident RS is walking in each of the rooms A, B, and C. That is, Fig. 3(D) shows the vibration acceleration detected in response to the occurrence of an event involving the walking of the resident RS in each of the rooms A, B, and C.

[0027] As can be seen from FIGS. 3(A) to 3(D), the change pattern of the vibration acceleration detected by the sensor 100 over time differs for each event. Although not shown, for example, when water is supplied to the toilet 10, the sink 20, and the bathtub 30, a change pattern corresponding to each event is obtained. Also, although not shown, when water is supplied to the sink 20 or the bathtub 30, different vibration patterns are generated due to differences in pressure when hot water and cold water are discharged from the faucet, and therefore the change pattern of the vibration acceleration also differs. Also, although not shown, a unique change pattern different from other events is obtained for the vibration acceleration when the resident RS is turned on. Furthermore, even for a common event, different change patterns occur in rooms A, B, and C. From these results, it can be seen that the vibration acceleration detected by the sensor 100 makes it possible to identify the content of the event that has occurred and the room in which the event has occurred.

[0028] Fig. 4 shows an example of the functional configuration of spatial event identification device 200. Spatial event identification device 200 may be configured to include, as hardware, a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), and a storage device such as an HDD (Hard Disk Drive) or SSD (Solid State Drive). Spatial event identification device 200 may also include, as hardware, a GPU (Graphics Processing Unit). The functions of spatial event identification device 200 shown in the figure are realized by the CPU (and GPU) included in spatial event identification device 200 executing a program.

[0029] The spatial event identification device 200 in the figure includes a sensor-compatible communication unit 201 , a terminal-compatible communication unit 202 , a control unit 203 , and a storage unit 204 .

[0030] The sensor corresponding communication unit 201 is communicably connected to the sensor 100. Instead of the sensor corresponding communication unit 201, for example, a data interface that receives data on vibration acceleration detected by the sensor 100 may be provided.

[0031] The terminal corresponding communication unit 202 is connected to the notification terminal 300 so as to be able to communicate with it.

[0032] The control unit 203 executes various controls in the spatial event discrimination device 200. The control unit 203 includes a generation unit 231, a discrimination unit 232, and a discrimination result output unit 233.

[0033] The generation unit 231 generates a spectrogram based on the vibration acceleration output from the sensor 100. The spectrogram indicates the strength of vibration on the time axis and the frequency axis by the vibration acceleration.

[0034] The discrimination unit 232 performs spatial event discrimination based on the spectrogram generated by the generation unit 231 . In this embodiment, the identification unit 232 may use a first trained model 2321 and a second trained model 2322 for spatial event identification.

[0035] The first trained model 2321 estimates an event that has occurred in response to an input of a spectrogram. Such a first trained model 2321 may be constructed, for example, by inputting training data in which a spectrogram is labeled with a label indicating the corresponding event to a learning device and causing the learning device to learn. The first trained model 2321 may also be constructed, for example, by deep learning.

[0036] The second trained model 2322 estimates the room RM in which an event is occurring in response to an input of a spectrogram. Such a second trained model 2322 may be constructed, for example, by inputting training data in which spectrograms obtained by causing various events to occur in each room RM and labeled with the corresponding room RM into a learning device and causing the learning device to learn. The second trained model 2322 may also be constructed, for example, by deep learning.

[0037] In spatial event identification, the identification unit 232 inputs the spectrograms generated by the generation unit 231 into the first trained model 2321 and the second trained model 2322, respectively, to obtain an estimation result of the event by the first trained model 2321 and an estimation result of the room RM in which the event occurred (which may be a position within the room RM) by the second trained model 2322. Based on the obtained estimation result of the event and the estimation result of the room RM in which the event occurred, the identification unit 232 determines the event that occurred and the room RM in which the event occurred, and outputs the determined content as an identification result.

[0038] The identification result output unit 233 generates notification information based on the identification result by the identification unit 232, and causes the generated notification information to be output by the notification terminal 300. The notification information may be output by the notification terminal 300, for example, by display, sound, lighting of a lighting unit, or the like.

[0039] An example of a processing procedure executed by the spatial event identification device 200 of this embodiment in response to spatial event identification will be described with reference to the flowchart of FIG. Step S100 : In the spatial event discrimination device 200 , the generation unit 231 acquires the vibration acceleration detected by the sensor 100 .

[0040] Step S102: The generation unit 231 performs a short-time FFT on the vibration acceleration of the sensor acquired in step S100. The time (frame time) corresponding to the frame length set in the short-time FFT here may be set so as to obtain a time resolution shorter than the duration of vibrations generated by an event such as the walking of the resident RS. By performing the short-time FFT using the frame time set in this manner, it becomes possible to improve the accuracy of identifying an event such as walking in response to short vibrations corresponding to footsteps generated by the walking of the resident RS.

[0041] Step S104: The generation unit 231 performs spectrogram calculation using the frequency characteristics for each frame obtained by the short-time FFT in step S102. A spectrogram is generated by the spectrogram calculation.

[0042] Step S106: The generation unit 231 performs band limitation on the spectrogram generated in step S104 as post-processing so as to fit the event to be identified. The band limitation in step S104 may be performed by, for example, setting different bands for each type of event to be identified by the identification unit 232, or by setting one common band that encompasses all types of events to be identified by the identification unit 232. By limiting the band of the spectrogram in this way, it is possible to improve the accuracy of identifying each event to be identified.

[0043] Step S108: The generation unit 231 further performs enhancement processing and normalization as post-processing of the generated spectrogram. Regarding the enhancement processing, the generation unit 231 may select either dB or vibration acceleration (Magnitude (m / s2)) for the band level based on, for example, the vibration acceleration level or S / N ratio of the sensor 100 that was the source of the spectrogram generation, and enhance the band level.

[0044] Step S110: The identification unit 232 inputs the spectrogram that has undergone the processing of step S108 into the first trained model 2321.

[0045] Step S112: The first trained model 2321 calculates the posterior probability of the occurrence of an event according to the input spectrogram. That is, an estimation is made of the event that has occurred.

[0046] Step S114: The identification unit 232 performs an event determination process to determine what kind of event has occurred based on the time variation of the posterior probability obtained in step S112.

[0047] Step S116: The identification unit 232 also inputs the spectrogram that has undergone the processing of step S108 to the second trained model 2322.

[0048] Step S118: The second trained model 2322 calculates the posterior probability of the location where the event occurred according to the input spectrogram. That is, the location where the event occurred is estimated. The estimated location may be in units of rooms RM, or may be a specific location (coordinates) in the room.

[0049] Step S120: The identification unit 232 performs an event determination process to determine the location where the event occurred, based on the time variation of the posterior probability obtained in step S112.

[0050] The processing corresponding to event identification using the first trained model 2321 in steps S110 to S114 and the processing corresponding to position identification using the second trained model 2322 in steps S116 to S120 may be executed in parallel.

[0051] Step S122: The identification unit 232 outputs the event determined in step S114 and the position determined in step S116 as a spatial event identification result.

[0052] The identification result output unit 233 may transmit notification information generated based on the spatial event identification result output in step S122 to the notification terminal 300, so that the notification terminal 300 issues a notification according to the content of the spatial event identification result.

[0053] <Modification> A modification of this embodiment will now be described. [First Modification] 5 in the above embodiment, event estimation by the first trained model 2321 and position estimation by the second trained model 2322 were performed in parallel. In other words, as shown in Fig. 6, a common spectrogram was input to the first trained model 2321 and the second trained model 2322, and the first trained model 2321 and the second trained model 2322 were configured to individually perform event estimation and position estimation, respectively.

[0054] In this modified example, as shown in Figure 7, first the first trained model 2321 estimates an event based on the input of a spectrogram, and then the second trained model 2322 estimates the location where the event occurred based on the input of the event estimated by the first trained model 2321 and the spectrogram. The second trained model 2322 of this modified example may be constructed, for example, by inputting training data in which position labels are attached to combinations of spectrograms and events into a learner and causing the learner to train the model. Depending on the content of the event, the location where the event occurs may be limited. Therefore, by inputting the estimated event data in addition to the spectrogram into the second trained model 2322 constructed in this modification, it becomes possible to improve the accuracy of the estimated location.

[0055] [Second Modification] Furthermore, when identifying an event and a position, the identification unit 232 may use a single trained model (integrated trained model) that integrates the functions of the first trained model 2321 and the second trained model 2322. That is, as illustrated in FIG. 8, the identification unit 232 inputs a spectrogram to the integrated trained model 2323. The integrated trained model 2323 comprehensively estimates an event that has occurred and the position at which the event has occurred, in accordance with the input spectrogram. The identification unit 232 determines the event that has occurred and the position using the estimation result of the event and the position by the integrated trained model 2323, and outputs a spatial event identification result according to the determined result.

[0056] For example, when the spatial event identification system of this embodiment is applied to buildings with different structures, it is preferable to configure it with a first trained model 2321 and a second trained model 2322, as in the above embodiment or first variant example. When the structure of a building differs, the positional relationships between multiple rooms that can be detected by one sensor 100 also differ for each building. For this reason, by training the second trained model 2322 for each building, it is possible to estimate the position with high accuracy in accordance with the building. On the other hand, the change pattern of vibration acceleration detected by the sensor 100 in response to the occurrence of an event is common to a certain extent even for different buildings. Therefore, the first trained model 2321 can be commonly used for all buildings.

[0057] 1 and 2 show components that detect vibrations in three rooms RM (RM-1, RM-2, and RM-3) in a building BL using a single sensor 100. If a larger number of rooms RM are arranged in the building BL and spatial event identification is to be performed on the other rooms RM as well, the sensor 100 may be provided at a position where vibrations can be detected for each of a predetermined number of other rooms RM. In this case, in the spatial event identification device 200, the generation unit 231 may generate a spectrogram for each vibration acceleration output from each of the multiple sensors 100. The identification unit 232 may estimate an event that has occurred and the position where the event occurred by performing spatial position estimation based on each of the generated spectrograms.

[0058] Note that a program for implementing the functions of the spatial event identification device 200, the notification terminal 300, etc. may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be loaded into a computer system and executed to perform the processing of the spatial event identification device 200, the notification terminal 300, etc. Here, "loading a program recorded on a recording medium into a computer system and executing it" includes installing the program on a computer system. The term "computer system" here includes hardware such as an OS and peripheral devices. The term "computer system" may also include multiple computer devices connected via a network, including communication lines such as the Internet, WAN, LAN, and dedicated lines. The term "computer-readable recording medium" refers to portable media such as floppy disks, optical magnetic disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into a computer system. Thus, the recording medium storing the program may be a non-transitory recording medium such as a CD-ROM. The recording medium may also include internal or external recording media accessible from a distribution server to distribute the program. The program code stored on the distribution server's recording medium may be different from the program code in a format executable by a terminal device. In other words, the format in which the program is stored on the distribution server does not matter as long as it can be downloaded from the distribution server and installed in a form that is executable on the terminal device. The program may be divided into multiple parts, each of which may be downloaded at different times and then combined on the terminal device, or each of the divided programs may be distributed by a different distribution server. Furthermore, the term "computer-readable recording medium" also includes a storage medium that stores a program for a certain period of time, such as volatile memory (RAM) within a computer system that serves as a server or client when a program is transmitted over a network. The program may also be one that realizes part of the functions described above.Furthermore, the above-mentioned functions may be realized in combination with a program already recorded in the computer system, that is, a so-called differential file (differential program).

[0059] <Additional Notes> (1) One aspect of this embodiment is a spatial event identification system including a sensor (100) that is configured to be able to detect vibration acceleration occurring in a plurality of spaces to be identified in a building (e.g., building BL) at a certain level or higher; a generation unit (231) that generates a spectrogram showing vibration acceleration on a time axis and a frequency axis based on the detection output of the sensor; and an identification unit (232) that identifies an event that has occurred in a plurality of spaces to be identified and the position in the plurality of spaces to be identified where the event occurred based on the spectrogram generated by the generation unit.

[0060] (2) One aspect of this embodiment is the spatial event identification system described in (1), wherein the generation unit may generate a spectrogram using the results of a fast Fourier transform using a time frame shorter than the vibration duration corresponding to the occurrence of a specified event to be identified.

[0061] (3) One aspect of this embodiment is a spatial event identification system described in (1) or (2), in which the generation unit may generate a spectrogram limited to a predetermined frequency band according to the type of event to be identified.

[0062] (4) One aspect of this embodiment is a spatial event identification system according to any one of (1) to (3), wherein the identification unit includes a trained model that has learned the relationship between spectrograms, events occurring in the plurality of spaces to be identified, and positions in the plurality of spaces to be identified where the events occur, and may identify the events that have occurred and their positions in the plurality of spaces to be identified based on the output of the trained model to which the spectrograms generated by the generation unit are input.

[0063] (5) One aspect of this embodiment is the spatial event identification system described in (4), wherein the trained models include a first trained model that has learned the relationship between spectrograms and events occurring in the multiple spaces to be identified, and a second trained model that has learned the relationship between spectrograms and positions in the multiple spaces to be identified where the events occur, and the identification unit may identify the event that has occurred based on the output of the first trained model that has input the spectrogram generated by the generation unit, and identify the positions in the multiple spaces to be identified based on the output of the second trained model that has input the spectrogram generated by the generation unit.

[0064] (6) One aspect of this embodiment is the spatial event identification system described in (5), wherein the identification unit may identify the event that has occurred based on the output of the first trained model to which the spectrogram generated by the generation unit is input, and identify positions in the multiple spaces to be identified based on the output of the second trained model to which the event identified based on the output of the first trained model and the spectrogram generated by the generation unit are input.

[0065] (7) One aspect of this embodiment is a spatial event identification system described in any one of (1) to (6), wherein the sensor may be provided in a piping installed in the building at a position where vibration acceleration occurring in the multiple spaces to be identified can be detected at a certain level or above.

[0066] (8) One aspect of this embodiment is the spatial event identification system described in (7), wherein the piping may correspond to drainage.

[0067] (9) One aspect of this embodiment is a spatial event identification method in a spatial event identification system, which includes a generation step in which a generation unit generates a spectrogram showing vibration acceleration on a time axis and a frequency axis based on the detection output of a sensor that is configured to be able to detect vibration acceleration occurring in multiple spaces to be identified in a building at a certain level or higher, and an identification step in which an identification unit identifies events that have occurred in multiple spaces to be identified and the positions in the multiple spaces to be identified where the events occurred, based on the spectrogram generated by the generation step.

[0068] (10) One aspect of this embodiment is a program for causing a computer in a spatial event identification system to function as a generation unit that generates a spectrogram showing vibration acceleration on a time axis and a frequency axis based on the detection output of a sensor that is configured to be able to detect vibration acceleration occurring in multiple spaces to be identified in a building at a certain level or above, and an identification unit that identifies an event that has occurred in multiple spaces to be identified and the location in the multiple spaces to be identified where the event occurred based on the spectrogram generated by the generation unit. [Explanation of symbols]

[0069] 10 Toilet, 20 Sink, 30 Bathtub, 40-1, 40-2 Drainage system, 100 Sensor, 200 Spatial event recognition device, 201 Sensor-compatible communication unit, 202 Terminal-compatible communication unit, 203 Control unit, 204 Memory unit, 231 Generation unit, 232 Recognition unit, 233 Recognition result output unit, 300 Notification terminal, 2321 First trained model, 2322 Second trained model, 2323 Integrated trained model

Claims

1. a sensor that is provided to be able to detect vibration acceleration occurring in a plurality of spaces to be identified in a building at a certain level or higher; a generation unit that generates a spectrogram showing vibration acceleration with respect to a time axis and a frequency axis based on the detection output of the sensor; an identification unit that identifies events that have occurred in a plurality of identification target spaces and positions in the plurality of identification target spaces where the events have occurred, based on the spectrograms generated by the generation unit; A spatial event identification system comprising:

2. The generating unit generates a spectrogram using a result of a fast Fourier transform using a time frame shorter than a vibration duration corresponding to the occurrence of a predetermined event to be identified. The spatial event identification system of claim 1 .

3. The generating unit generates a spectrogram limited to a predetermined frequency band according to the type of event to be identified.

3. A spatial event identification system according to claim 1 or 2.

4. The identification unit includes a trained model that has learned the relationship between spectrograms, events that occur in the plurality of identification target spaces, and positions in the plurality of identification target spaces where the events occur, and identifies the events that have occurred and their positions in the plurality of identification target spaces based on an output of the trained model to which the spectrograms generated by the generation unit are input.

3. A spatial event identification system according to claim 1 or 2.

5. the trained model includes a first trained model that has learned a relationship between a spectrogram and an event occurring in the plurality of identification target spaces, and a second trained model that has learned a relationship between a spectrogram and a position in the plurality of identification target spaces where an event occurs; The identification unit identifies the occurred event based on an output of the first trained model to which the spectrogram generated by the generation unit is input, and identifies positions in the plurality of identification target spaces based on an output of the second trained model to which the spectrogram generated by the generation unit is input. The spatial event identification system of claim 4.

6. The identification unit identifies the occurred event based on an output of the first trained model to which the spectrogram generated by the generation unit is input, and identifies a position in the plurality of identification target spaces based on an output of the second trained model to which the event identified based on the output of the first trained model and the spectrogram generated by the generation unit are input. The spatial event identification system of claim 5 .

7. The sensor is provided at a position on the piping installed in the building where it can detect vibration acceleration occurring in the plurality of spaces to be identified at a certain level or higher.

3. A spatial event identification system according to claim 1 or 2.

8. The piping corresponds to drainage The spatial event identification system of claim 7.

9. A method for spatial event identification in a spatial event identification system, comprising: a generation step in which a generation unit generates a spectrogram showing vibration acceleration with respect to a time axis and a frequency axis based on detection outputs of a sensor that is provided so as to be able to detect vibration acceleration occurring in a plurality of spaces to be identified in the building at a certain level or higher; an identification step in which an identification unit identifies events that have occurred in a plurality of identification target spaces and positions in the plurality of identification target spaces where the events have occurred, based on the spectrograms generated in the generation step; A spatial event identification method comprising:

10. A computer in a spatial event identification system, a generation unit that generates a spectrogram showing vibration acceleration with respect to a time axis and a frequency axis based on detection outputs of sensors that are provided so as to be able to detect vibration acceleration occurring in a plurality of spaces to be identified in the building at a certain level or higher; an identification unit that identifies events that have occurred in a plurality of spaces to be identified and positions in the plurality of spaces to be identified where the events have occurred, based on the spectrograms generated by the generation unit; A program to function as a

Images