Apparatus, method, and program for processing acoustic data detected in facility

Abstract

To provide an apparatus for identifying an abnormality of each device on the basis of acoustic data.SOLUTION: An apparatus includes: a device database connection section connected to a device database for recording priority data indicating the priority of acoustic sensors used for abnormality detection in each of a plurality of devices in a facility; an acoustic data acquisition section for acquiring acoustic data detected by each of the plurality of acoustic sensors in the facility; an acoustic sensor selection section for selecting an acoustic sensor used for the abnormality detection of each device among the plurality of acoustic sensors on the basis of the priority data recorded in the device database; and an abnormal device identification section for identifying an abnormality of each device on the basis of the acoustic data detected by the selected acoustic sensor.SELECTED DRAWING: Figure 10

Description

Technical Field

[0001] The present invention relates to an apparatus, a method, and a program for processing acoustic data detected within a facility.

Background Art

[0002] Patent Document 1 describes that "the inspection support system 1 shown in FIG. 1 enables an inspector 4 at a monitoring location 3 remote from an inspection site 2 to inspect the equipment at the inspection site 2" (paragraph 0010), that "the sound information from the unmanned mobile body 10 includes the sound information respectively acquired by a plurality of microphones 131 to 13m, and the sound determination unit 72 can determine the sound source direction, the sound source position, and the sound volume based on this sound information" (paragraph 0037), and that "the graphic addition unit 73 adds at least one of a graphic representing the position of the sound source and a graphic representing the sound volume to the captured image acquired from the unmanned mobile body 10 based on the determination result by the sound determination unit 72" (paragraph 0038). [Prior Art Document] [Patent Document] [Patent Document 1] Japanese Patent Application Laid-Open No. 2020-149349

Summary of the Invention

[0003] In a first aspect of the present invention, there is provided an apparatus including: a device database connection unit connected to a device database that records priority data indicating the priority of acoustic sensors used for abnormality detection for each of a plurality of devices within a facility; an acoustic data acquisition unit that acquires acoustic data detected by each of a plurality of acoustic sensors within the facility; an acoustic sensor selection unit that selects, from among the plurality of acoustic sensors, an acoustic sensor used for abnormality detection of each device based on the priority data recorded in the device database; and an abnormal device identification unit that identifies an abnormality of each device based on the acoustic data detected by the selected acoustic sensor.

[0004] In the above-described apparatus, at least one of the plurality of acoustic sensors may have different sound collection characteristics from at least one other acoustic sensor.

[0005] In the above-described apparatus, each of the plurality of acoustic sensors may be mounted on each of a plurality of robots that can move within the facility.

[0006] The above-described apparatus may include a schedule determination unit that determines a schedule for each robot equipped with each acoustic sensor to collect the sound of each device based on the priority of each acoustic sensor indicated in the priority data for each device.

[0007] Any of the above-described apparatuses may further include a priority update unit that updates the priority data according to the accuracy of abnormality identification based on the acoustic data detected by each of the plurality of acoustic sensors for each device.

[0008] Any of the above-described apparatuses may include a device control unit that performs control for dealing with the abnormality of one device on at least one of the one device or other devices in the facility in response to the identification of the abnormality of the one device.

[0009] In a second aspect of the present invention, there is provided a method including: an acoustic data acquisition step of acquiring acoustic data detected by each of a plurality of acoustic sensors in a facility; an acoustic sensor selection step of selecting, from among the plurality of acoustic sensors, an acoustic sensor to be used for abnormality detection of each device based on priority data indicating the priority of the acoustic sensors to be used for abnormality detection of each of the plurality of devices in the facility; and an abnormal device identification step of identifying the abnormality of each device based on the acoustic data detected by the selected acoustic sensor.

[0010] In a third aspect of the present invention, there is provided a program that causes a computer to function as a device database connection unit connected to a device database that records priority data indicating the priority of acoustic sensors used for abnormality detection for each of a plurality of devices in a facility, an acoustic data acquisition unit that acquires acoustic data detected by each of the plurality of acoustic sensors in the facility, an acoustic sensor selection unit that selects an acoustic sensor used for abnormality detection of each device from among the plurality of acoustic sensors based on the priority data recorded in the device database, and an abnormal device identification unit that identifies an abnormality of each device based on the acoustic data detected by the selected acoustic sensor.

[0011] Note that the above summary of the invention does not enumerate all the features of the present invention. Also, sub-combinations of these feature groups can also be inventions.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Embodiments for Carrying Out the Invention

[0013] Hereinafter, the present invention will be described through embodiments of the invention. However, the following embodiments do not limit the invention according to the claims. Also, not all combinations of features described in the embodiments are essential for the solution means of the invention.

[0014] FIG. 1 shows a schematic configuration of a facility 1 according to this embodiment. The facility 1 may be an entire or partial segment such as a factory or a plant, and a plurality of devices 10 are arranged therein. Such a factory or plant may be, for example, a factory for producing various industrial products, an industrial plant such as a chemical or metal plant, a plant for managing and controlling wells and their surroundings in a gas field or an oil field, a plant for managing and controlling power generation such as hydraulic, thermal, or nuclear power, a plant for managing and controlling environmental power generation such as solar or wind power, a plant for managing and controlling water supply and sewerage or a dam, etc. Also, the facility 1 may be an entire or partial part of a building or a transportation vehicle that includes a plurality of devices 10 to be controlled and monitored. The facility 1 includes a plurality of devices 10, a control system 20, one or more robots 30, a monitoring device 40, a collected information database 50, a device database 60, and a robot database 70.

[0015] A plurality of devices 10 are provided at various locations within the facility 1. Each of the plurality of devices 10 may be installed either indoors or outdoors within the area of the facility 1. At least some of the plurality of devices 10 may be a process device, a power generation device, or any other arbitrary device (or equipment) that is under the control of the control system 20, or may be a part of such a device. At least some of the devices 10 may be provided with field devices that operate under the control of the control system 20, other devices, or an operator. Also, at least some of the devices 10 may be the field devices themselves.

[0016] Such field devices may be, for example, sensor devices such as pressure gauges, flow meters, temperature sensors, valve devices such as flow control valves and on-off valves, actuator devices such as fans and motors, imaging devices such as cameras or videos that capture objects such as the status of plants or facilities, acoustic devices such as microphones or speakers that collect abnormal sounds or emit alarm sounds in plants or facilities, position detection devices that output the position information of the devices possessed by the facility 1, or other devices. Also, the other devices 10 among the plurality of devices 10 may be structures such as pipes, storage tanks, supports, partition walls, or others that are not under the control of the control system 20.

[0017] The control system 20 is connected to at least one of the plurality of devices 10 that is to be controlled. The control system 20 may be, for example, a distributed control system (DCS). The control system 20 controls each device 10 according to the state of each device 10 measured by sensors or the like provided in each device 10 to be controlled.

[0018] Each of one or more robots 30 is used to monitor at least some of the plurality of devices 10 within the facility 1. Each robot 30 can move within the facility 1 to collect sounds and capture images and so on.

[0019] The monitoring device 40 is communicably connected to each of one or more robots 30. The monitoring device 40 may be connected to each robot 30 via a wireless network such as a mobile phone network, a wireless WAN, a wireless LAN, or Bluetooth (registered trademark), or may be connected to each robot 30 via a wired network such as wired Ethernet (registered trademark). The monitoring device 40 according to the present embodiment is installed inside the facility 1. Alternatively, the monitoring device 40 may be provided outside the facility 1 and remotely monitor each device 10 inside the facility 1. For example, the monitoring device 40 may be installed in another facility, or may be realized by a cloud server on the Internet, for example.

[0020] The monitoring device 40 receives, from each robot 30, the collected data such as acoustic data and image data collected by each robot 30 inside the facility 1, and monitors the state of each device 10 using the collected data. When the monitoring device 40 identifies an abnormality in any of the devices 10, it may instruct the control system 20 to perform control for dealing with the abnormality.

[0021] The monitoring device 40 uses a collected information database 50, a device database 60, and a robot database 70 to monitor each device 10 inside the facility 1 using one or more robots 30. The collected information database 50 records the collected data collected from each robot 30. The device database 60 records device data about each device 10 inside the facility 1. The robot database 70 stores data related to each robot 30. These databases may be storage devices such as a hard disk connected to the monitoring device 40 by wire or wirelessly, or cloud storage on the Internet, or may be temporarily stored in the memory of the monitoring device 40 or the like.

[0022] FIG. 2 shows the configuration of the control system 20 according to the present embodiment together with the display device 240 and the input device 250. The display device 240 displays a display screen output by the control system 20. The input device 250 inputs an instruction to the control system 20 from a user such as an operator, a worker, or a maintenance staff of the facility 1 and supplies it to the control system 20. The display device 240 and the input device 250 may be provided in a monitoring console or a user terminal connected to the control system 20.

[0023] The control system 20 may be a computer such as a workstation, a server computer, a general-purpose computer, or other computers, or may be a computer system to which a plurality of computers are connected. Such a computer system is also a computer in a broad sense. Further, the control system 20 may be implemented by one or more executable virtual computer environments in the computer. Instead of this, the control system 20 may be a dedicated computer designed for controlling each device 10, or may be dedicated hardware realized by a dedicated circuit. In the present embodiment, the control system 20 is installed in the facility 1, but the control system 20 may be provided outside the facility 1 by a cloud computing system or the like on the Internet.

[0024] The control system 20 includes a state acquisition unit 210, a state determination unit 220, a display processing unit 230, an instruction input unit 260, and a control unit 270. The state acquisition unit 210 acquires device state data indicating the state of each device 10, such as the internal state value of each device 10 and the measurement value measured by a sensor provided in each device 10, from each device 10. The state acquisition unit 210 may receive the device state data of each device 10 by using communication according to a communication protocol such as HART (registered trademark), BRAIN, Foundation Fieldbus (registered trademark), ISA100.11a. The state acquisition unit 210 may receive the device state data of the device 10 detected by the monitoring device 40 from the monitoring device 40 by using the collection data collected by the robot 30.

[0025] The state determination unit 220 is connected to the state acquisition unit 210. The state determination unit 220 determines whether each device 10 is normal or abnormal by using the device state data acquired by the state acquisition unit 210. The state determination unit 220 may calculate a health index indicating the soundness of the operation in the entire facility 1 or a specific range within the facility 1 from at least one of the internal state values or measured values of two or more devices 10, and determine whether the operation is normal or not. The state determination unit 220 may determine the normality or abnormality of at least some of the devices 10 by receiving from the monitoring device 40 the determination result of the normality or abnormality of the device 10 determined by the monitoring device 40 using the collection data collected by the robot 30.

[0026] The display processing unit 230 is connected to the state acquisition unit 210 and the state determination unit 220. The display processing unit 230 performs display control to cause the display device 240 to display a display screen including device state data such as the internal state value and measured value of each device 10, a soundness index for the entire facility 1 or a specific range within the facility 1, and the occurrence status of abnormalities of each device 10.

[0027] The instruction input unit 260 receives an instruction from a user for the control system 20 from the input device 250. Such an instruction may be, for example, an instruction from a user who has seen the display screen displayed on the display device 240 to change the operation of at least one device 10. The instruction input unit 260 may be connected to the monitoring device 40 and may be communicable with the monitoring device 40 by wire or wirelessly. The instruction input unit 260 may receive from the monitoring device 40 an instruction to perform control for dealing with an abnormality in response to the detection of an abnormality in at least one device 10 by the monitoring device 40.

[0028] The control unit 270 is connected to the state acquisition unit 210 and the instruction input unit 260. The control unit 270 controls each device 10 according to the device state data acquired by the state acquisition unit 210. The control unit 270 may control each device 10 according to the instruction in response to receiving an instruction from a user or an instruction from the monitoring device 40.

[0029] Figure 3 shows the processing flow of the control system 20 according to this embodiment. In step 300 (S300), the state acquisition unit 210 acquires device state data indicating the state of each device 10 from each device 10. The state acquisition unit 210 may receive device data indicating the state of at least one device 10 from the monitoring device 40.

[0030] In S310, the state determination unit 220 determines whether each device 10 is normal or abnormal. The state determination unit 220 may determine whether the device 10 is normal (not abnormal) according to whether at least one of the internal state value or the measured value of the device 10 is within a predetermined normal range corresponding to the value. Further, the state determination unit 220 calculates a health index indicating the soundness of the operation in the entire facility 1 or a specific range within the facility 1 from at least one of the internal state values or the measured values of two or more devices 10 using a pre-defined calculation formula or the like, and determines whether the operation of the entire facility 1 or a specific range within the facility 1 is normal according to whether the value of the health index is within the normal range.

[0031] In S320, the display processing unit 230 performs display control to cause the display device 240 to display a display screen including the internal state value and the measured value of each device 10, the health index for the entire facility 1 or a specific range within the facility 1, and the abnormal occurrence status of each device 10. The display processing unit 230 may generate a video output of the display screen and supply it to the display device 240, or may generate html or a script for generating the display screen and transmit it to the display device 240.

[0032] In S330, the instruction input unit 260 receives an instruction from the user for the control system 20 from the input device 250. The instruction input unit 260 may receive an instruction for performing control for dealing with an abnormality transmitted by the monitoring device 40 that has detected an abnormality in at least one device 10.

[0033] In S340, the control unit 270 controls each device 10 according to a predetermined control algorithm, control model, or the like, using the device state data acquired by the state acquisition unit 210. The control unit 270 may calculate a control value for each device 10 by PI control, PID control, or the like. The control unit 270 may also calculate a control value for each device 10 according to the device state data acquired by the state acquisition unit 210, using various machine learning models or the like. The control unit 270 may change the control content for each device 10 in response to an instruction from the instruction input unit 260.

[0034] The control system 20 repeats the processes from S300 to S340. Thereby, the control system 20 can adaptively control each device 10 according to the state of each device 10.

[0035] FIG. 4 shows the configuration of the robot 30 according to the present embodiment. The robot 30 includes one or more sensors 400, one or more sensors 410, one or more actuators 420, a state acquisition unit 430, a communication unit 440, and a control unit 450.

[0036] Each of the one or more sensors 400 detects or measures the state within the facility 1. In the example of this figure, the sensor 400a is an acoustic sensor that detects the sound within the facility 1 and outputs it as acoustic data. The sensor 400b is an image sensor that captures an image within the facility 1 and outputs it as image data. The sensor 400c is a LiDAR sensor. The sensor 400c measures at least one of the shape of each object within the irradiation range of light or laser or the distance to each irradiation point by irradiating light or laser to the outside and detecting the reflected light, and outputs the measurement result as LiDAR data. The robot 30 may not include at least one of the sensors 400a to 400c. The robot 30 may also include at least one of a temperature sensor, a humidity sensor, a gas sensor, or other various sensors.

[0037] Each of the one or more sensors 410 detects or measures the state of the robot 30. In the example of this figure, the sensor 410a is a position sensor that detects the position of the robot 30 and outputs it as position data. The sensor 410a may be a GPS, specifically, a GPS (Global Positioning System) receiver that receives signals from GPS satellites to identify the position of the robot 30. Alternatively, the sensor 410a may be a sensor for identifying the position of the robot 30 using any position detection system capable of detecting the position of the robot 30 within the facility 1.

[0038] The sensor 410b is an azimuth sensor such as a geomagnetic sensor, for example. The sensor 410b measures the azimuth of the robot 30 or the azimuth of the detection direction of each sensor 400. The sensor 410c is an angle sensor or an elevation angle sensor, etc., for measuring the detection direction of each sensor 400. The sensor 410b and the sensor 410c output direction data indicating at least one of the azimuth of the robot 30 or the detection direction of each sensor 400 to the state acquisition unit 430. Note that the robot 30 may not include at least one of the sensors 410a to 410c. The robot 30 may include at least one of a speed sensor, an acceleration sensor, or other various sensors.

[0039] Each of the one or more actuators 420 is a motor or the like for driving each part of the robot 30. The robot 30 may include various actuators 420 such as an actuator 420 for movement by traveling or flying, etc., an actuator 420 for changing the orientation of the entire robot 30 or a part thereof, and an actuator 420 for operating equipment such as an arm.

[0040] The state acquisition unit 430 is connected to one or more sensors 400 and one or more sensors 410. The state acquisition unit 430 acquires the state inside the facility 1 measured by the one or more sensors 400 and the state of the robot 30 measured by the one or more sensors 410. The state acquisition unit 430 may acquire the state inside the facility 1 by receiving various measurement data such as acoustic data, image data, and LiDAR data from the one or more sensors 400. The state acquisition unit 430 may acquire the state of the robot 30 by receiving various measurement data such as position data and direction data from the one or more sensors 410.

[0041] The communication unit 440 is connected to the state acquisition unit 430. The communication unit 440 performs wireless or wired communication with the monitoring device 40. The communication unit 440 transmits the various measurement data acquired by the state acquisition unit 430 to the monitoring device 40. Also, the communication unit 440 receives various instructions for the robot 30 from the monitoring device 40.

[0042] The control unit 450 controls each actuator 420 according to the various measurement data acquired by the state acquisition unit 430 and the instructions from the monitoring device 40. Thereby, the robot 30 can perform operations corresponding to the state inside the facility 1, the state of the robot 30, and the instructions from the monitoring device 40.

[0043] Note that some robots 30 may be fixedly installed inside the facility 1, and such robots 30 do not need to be provided with actuators for movement. Also, inside the facility 1, monitoring devices (such as surveillance cameras and monitoring microphones) that are equipped with one or more sensors 400 but are not classified as robots may be provided. Such monitoring devices will have some functions and configurations of the robot 30. Therefore, in this specification, for the sake of convenience of explanation, at least one of the robots 30 may be such a monitoring device, and such a monitoring device shall perform the following processes, etc. within the scope of the implemented functions.

[0044] FIG. 5 shows the processing flow of the robot 30 according to the present embodiment. In S500, one or more sensors 400 observe the state of the facility 1 in the vicinity of the robot 30. One or more sensors 410 observe the state of the robot 30. The state acquisition unit 430 acquires various measurement data indicating the state of the facility 1 from one or more sensors 400. The state acquisition unit 430 acquires various measurement data indicating the state of the robot 30 from one or more sensors 410.

[0045] In S510, the communication unit 440 transmits various measurement data indicating the state in the facility 1 and the state of the robot 30 observed by one or more sensors 400 and one or more sensors 410 to the monitoring device 40. Here, when the central direction of measurement of each sensor 400 coincides with the direction of the robot 30, that is, for example, when the sensor unit or the like mounting each sensor 400 faces the front of the robot 30, the measurement direction (central direction of measurement) of each sensor 400 coincides with the direction of the robot 30. In such a case, if the robot 30 transmits its own direction to the monitoring device 40, the monitoring device 40 can obtain the measurement direction of each sensor 400.

[0046] When the direction of the sensor unit mounting each sensor 400 is variable with respect to the robot 30 main body, the central direction of measurement of each sensor 400 and the direction of the robot 30 do not necessarily coincide. In this case, the robot 30 may transmit direction data including its own direction and the measurement direction of each sensor 400 to the monitoring device 40. Alternatively, the robot 30 may transmit direction data including its own direction and the difference or offset of the measurement direction of each sensor 400 with respect to its own direction to the monitoring device 40. In this case, the monitoring device 40 can obtain the measurement direction of each sensor 400 by adding the difference or offset of the measurement direction of each sensor 400 to its own direction.

[0047] In S520, the communication unit 440 receives an instruction from the monitoring device 40. In S530, the control unit 450 performs an operation according to the instruction from the monitoring device 40. The control unit 450 drives at least one actuator 420 according to an instruction for movement, an instruction for changing direction, an instruction for changing the direction of an acoustic sensor, an image sensor, etc. received from the monitoring device 40, thereby operating the robot 30 as instructed. Further, the control unit 450 may set various parameters within the robot 30 according to an instruction from the monitoring device 40.

[0048] The robot 30 repeats the processes from S500 to S530. Thereby, the robot 30 can move within the facility 1 according to an instruction from the monitoring device 40, and perform sound collection by the sensor 400a, image capturing by the sensor 400b, acquisition of LiDAR data by the sensor 400c, etc.

[0049] FIG. 6 shows the configuration of the monitoring device 40 according to the present embodiment together with the collected information database 50, the device database 60, the robot database 70, the input device 680, and the display device 690. The input device 680 inputs an instruction for the monitoring device 40 from a user such as an operator, a worker, or a maintenance worker of the facility 1 and supplies it to the monitoring device 40. The display device 690 displays a display screen output by the monitoring device 40. The input device 680 and the display device 690 may be provided in a monitoring console or a user terminal etc. connected to the monitoring device 40. The display device 690 and the input device 680 may be shared with the display device 240 and the input device 250 shown in FIG. 2.

[0050] The monitoring device 40 may be a facility management device that manages each device 10 in the facility 1, or may be a device that realizes some of the functions included in the facility management device. The monitoring device 40 may be a computer such as a workstation, a server computer, a general-purpose computer, or other computers, or may be a computer system to which a plurality of computers are connected. Such a computer system is also a computer in a broad sense. Further, the monitoring device 40 may be implemented by one or more executable virtual computer environments in the computer. Such a computer functions as the monitoring device 40 by executing a program for monitoring each device 10 using the robot 30. Alternatively, the monitoring device 40 may be a dedicated computer designed for monitoring each device 10, or may be dedicated hardware realized by a dedicated circuit.

[0051] In this embodiment, the monitoring device 40 is installed in the facility 1. Alternatively, the monitoring device 40 may be provided outside the facility 1 by a cloud computing system or the like on the Internet. Also, in this embodiment, the monitoring device 40 remotely operates each robot 30. Alternatively, the monitoring device 40 may be mounted on at least one robot 30 and directly operate each robot 30 or a group of robots including two or more robots 30 on site.

[0052] The monitoring device 40 may be an example of a device, and includes a communication unit 600, a collected information database connection unit 620, a device database connection unit 630, a robot database connection unit 640, a monitoring processing unit 650, an input processing unit 660, a display processing unit 665, and a communication unit 670. The communication unit 600 performs wireless or wired communication with the robot 30. The communication unit 600 includes an acoustic data acquisition unit 602, an image data acquisition unit 604, a LiDAR data acquisition unit 606, a position data acquisition unit 608, a direction data acquisition unit 610, and an instruction transmission unit 612.

[0053] The acoustic data acquisition unit 602 acquires acoustic data detected by each of a plurality of acoustic sensors within Facility 1. Each of the plurality of acoustic sensors may be an acoustic sensor 400a mounted on each of a plurality of robots 30 that can move within the facility. Each acoustic sensor 400a may be mounted on a separate robot 30. Alternatively, two or more acoustic sensors 400a may be mounted on a common robot 30. The acoustic data acquisition unit 602 may acquire acoustic data by receiving the acoustic data transmitted by the robot 30.

[0054] Here, each acoustic sensor 400a may be a sensor that detects sound and may output acoustic data corresponding to the detected sound. Each acoustic sensor may be a piezoelectric sensor such as a microphone, or may be another sensor. At least one of the plurality of acoustic sensors 400a may have different sound collection characteristics from at least one other acoustic sensor 400a. The sound collection characteristics may be characteristics such as how much sound can be collected, and may be at least one of, for example, directivity, sensitivity, frequency characteristics, or dynamic range. The directivity may indicate the relationship between the orientation of the acoustic sensor 400a with respect to the sound source and the sound collection performance. The sensitivity may indicate the ratio of the sound pressure to the output value. The sensitivity may also indicate the area of the collectable sound. The frequency characteristics may indicate the relationship between the frequency and the sound collection performance. The dynamic range may indicate the range of the magnitude of the detectable sound. Among the plurality of acoustic sensors 400a, at least one acoustic sensor 400a may have different other characteristics (for example, at least one of durability, accuracy, temperature characteristics, and noise characteristics) from at least one other acoustic sensor 400a.

[0055] The image data acquisition unit 604 acquires the image data captured within Facility 1. The image data acquisition unit 604 may acquire the image data by receiving the image data transmitted by the robot 30. The LiDAR data acquisition unit 606 acquires the LiDAR data detected by the sensor 400c of the robot 30 within Facility 1. The LiDAR data acquisition unit 606 may acquire the LiDAR data by receiving the LiDAR data transmitted by the robot 30.

[0056] The position data acquisition unit 608 acquires the position data indicating the position of the robot 30. The position data acquisition unit 608 may acquire the position data by receiving the position data transmitted by the robot 30. The direction data acquisition unit 610 acquires the direction data indicating the directions of the robot 30 and each sensor 400. The direction data acquisition unit 610 may acquire the direction data by receiving the direction data transmitted by the robot 30. The instruction transmission unit 612 transmits the instruction for the robot 30 determined within the monitoring device 40 to the robot 30.

[0057] The collected information database connection unit 620 is connected to the collected information database 50, the communication unit 600, and the monitoring processing unit 650. The collected information database connection unit 620 records the various measurement data acquired by the acoustic data acquisition unit 602, the image data acquisition unit 604, the LiDAR data acquisition unit 606, the position data acquisition unit 608, and the direction data acquisition unit 610 within the communication unit 600 as collected data in the collected information database 50. Also, the collected information database connection unit 620 accesses the collected information database 50 in response to a request from the monitoring processing unit 650. Note that if the monitoring device 40 does not need to record the history of the collected data, it may not include the collected information database connection unit 620, and the measurement data acquired by the communication unit 600 may be supplied to the monitoring processing unit 650 without passing through the collected information database connection unit 620. In this case, the collected information database 50 is unnecessary.

[0058] The device database connection unit 630 is connected to the device database 60 and the monitoring processing unit 650. The device database connection unit 630 accesses the device database 60 in response to a request from the monitoring processing unit 650. The robot database connection unit 640 is connected to the robot database 70 and the monitoring processing unit 650. The robot database connection unit 640 accesses the robot database 70 in response to a request from the monitoring processing unit 650.

[0059] The monitoring processing unit 650 is connected to the collected information database connection unit 620, the device database connection unit 630, the robot database connection unit 640, the input processing unit 660, and the communication unit 670. The monitoring processing unit 650 reads out the robot data of each robot 30 via the robot database connection unit 640, and determines the monitoring actions in the facility 1 that each robot 30 should perform using the robot data. The monitoring processing unit 650 instructs each robot 30 to perform each operation included in the determined monitoring actions via the instruction transmission unit 612 in the communication unit 600.

[0060] Also, the monitoring processing unit 650 reads out the collected data collected by one or more robots 30 and recorded in the collected information database 50 via the collected information database connection unit 620. Then, the monitoring processing unit 650 detects the state of each device 10 monitored by each robot 30 using the collected data. The monitoring processing unit 650 determines whether each device 10 is normal or not using the collected data. The monitoring processing unit 650 may read out the device data of each device 10 via the device database connection unit 630 and determine whether each device 10 is normal or abnormal based on the information registered in the device data.

[0061] The monitoring processing unit 650 may cause the display processing unit 665 to generate a display screen for displaying the state of each device 10 and the determination result of normal or abnormal, and display it on the display device 690. The monitoring processing unit 650 may transmit the state of each device 10 and the determination result of normal or abnormal to the control system 20 via the communication unit 670.

[0062] When the monitoring processing unit 650 monitors each device 10 using one or a plurality of robots 30 and detects an abnormality in at least one device 10, it may generate an instruction to perform control for dealing with the abnormality. The monitoring processing unit 650 may transmit an instruction to perform control for dealing with the abnormality of the device 10 to the control system 20 via the communication unit 670.

[0063] The input processing unit 660 is connected to the input device 680. The input processing unit 660 receives input of information from the user to the input device 680 and supplies it to the monitoring processing unit 650. The display processing unit 665 is connected to the monitoring processing unit 650. The display processing unit 665 performs display processing to generate a display screen according to an instruction from the monitoring processing unit 650 and display it on the display device 690.

[0064] The communication unit 670 is connected to the monitoring processing unit 650. The communication unit 670 may be connected to the control system 20 and may be able to communicate with the control system 20 wirelessly or by wire. The communication unit 670 may transmit the device status data of each device 10 detected by the monitoring processing unit 650 to the control system 20. The communication unit 670 may transmit the determination result of normal or abnormal of each device 10 determined by the monitoring processing unit 650 to the control system 20. When an abnormality is detected in at least one device 10, the communication unit 670 may transmit an instruction to perform control for dealing with the abnormality to the control system 20. Note that the communication unit 670 may not have a function of transmitting at least one of the device status data of the device 10, the determination result of normal or abnormal of the device 10, or the instruction to perform control for dealing with the abnormality to the control system 20. When the monitoring device 40 does not have any of these functions, the communication unit 670 may not be provided.

[0065] According to the monitoring device 40 shown above, by monitoring the inside of the facility 1 using one or a plurality of robots 30 that can move inside the facility 1, a large number of devices 10 can be monitored using a relatively small number of robots 30. Further, since the monitoring device 40 can control the robot 30 according to the situation at each observation point and collect necessary information, it is possible to detect the state of each device 10 that cannot be collected by sensors or the like pre-installed in each device 10.

[0066] FIG. 7 shows an example of the data structure of the collected information database 50 according to the present embodiment. The collected information database 50 records the measurement data collected from each robot 30 as collected data. The collected information database 50 may record one or more records, each of which is a unit of collected data. In the data structure of the collected information database 50 shown in this figure, each record of the collected data is arranged in the row direction, and the collected data of each record includes fields of "date and time", "position and direction", "robot identification information", and "measurement data".

[0067] "Date and time" is a field for recording the date and time when the measurement data of the corresponding record was acquired by the robot 30. "Position and direction" is a field for recording at least one of the position and direction of the robot 30 or each sensor 400 at the timing when the measurement data of the corresponding record was acquired. The collected information database 50 may record, as "position and direction", at least one of the position detected by the sensor 410a of the robot 30, the orientation of the robot 30 detected by the sensor 410b, or the direction of the robot 30 or each sensor 400 detected by the sensor 410c at the timing indicated by "date and time".

[0068] "Robot identification information" is a field that records data values such as an identification number, serial number, or other unique number or character string that can identify the robot 30 that acquired the measurement data of the corresponding record. "Measurement data" is a field that records the measurement data acquired by the robot 30 at the timing indicated by "date and time". The data recorded as "measurement data" may be various types of measurement data acquired by the robot 30, such as at least one of acoustic data, image data, or LiDAR data.

[0069] The communication unit 600 in the monitoring device 40 may receive, as separate packets or packet groups, etc., data including the acquisition date and time of the measurement data by the robot 30, the robot identification information of the robot 30, and the measurement data regarding the position and orientation of the robot 30, and data including the acquisition date and time of the measurement data by the robot 30, the robot identification information of the robot 30, and the measurement data regarding the state within the facility 1. In this case, each time the communication unit 600 receives the measurement data regarding the position and orientation of the robot 30, it updates the position and orientation of the robot 30 managed by the monitoring device 40. When the communication unit 600 receives the measurement data regarding the state within the facility 1, it generates a record of the collected data by associating the measurement data with the position and orientation of the robot 30 at the acquisition date and time of the measurement data. Alternatively, when the communication unit 440 in the robot 30 transmits various measurement data indicating the state within the facility 1 and the state of the robot 30 observed by each sensor 400 and each sensor 410 to the monitoring device 40 (see S510 in FIG. 5), it may assemble the collected data for one record from these measurement data, encode it into one or two or more packets, and transmit it to the monitoring device 40.

[0070] FIG. 8 shows an example of the data structure of the device database 60 according to the present embodiment. The device database 60 records device data for each of the plurality of devices 10 in the facility 1. The device database 60 may record records, each of which is device data for one device, for the number of devices 10. In the data structure of the device database 60 shown in this figure, each record of the device data is arranged in the row direction, and the device data of each record includes fields of "device identification information", "device name", "device information", "location", "template", "acoustic data", and "priority data".

[0071] "Device identification information" is a field that records data values such as an identification number, a serial number, or other unique number or character string that can identify the corresponding device 10 within the facility 1. "Device name" is a field that records the device name assigned to the corresponding device 10 by a user or the like.

[0072] "Device information" is a field that records various information about the corresponding device 10. "Location" is a field that records the location of the device 10 within the facility 1. The location of the device 10 recorded in the "location" field may further indicate the area to which the device 10 belongs among the plurality of areas included in the facility 1 (for example, floors, rooms, and sections within a building). "Template" is a field that records image data (template image data) of the appearance of the corresponding device 10 for use in identifying the device 10 by image matching. "Acoustic data" is a field that records at least one of the acoustic data collected from the corresponding device 10 or the acoustic data emitted by the corresponding device 10 when the device 10 is normal.

[0073] "Priority data" is a field that records priority data indicating the priority of the acoustic sensor used for anomaly detection for the corresponding device 10. As an example, the priority data may associate the identification information of each of the plurality of acoustic sensors 400a (or the identification information of the robot 30 equipped with the acoustic sensor 400a) with the priority of the acoustic sensor 400a. The priority may indicate the priority order among the plurality of acoustic sensors 400a. As an example, the priority may be a value within a reference range such as 0 to 10, and it may indicate that the higher the value, the more preferably it should be used. The priorities of the respective acoustic sensors 400a may be different from each other, or the priorities may be the same among at least some of the acoustic sensors 400a.

[0074] FIG. 9 shows an example of the data structure of the robot database 70 according to the present embodiment. The robot database 70 records robot data for each of one or more robots 30 in the facility 1. The robot database 70 may record records, each of which is robot data for one robot, for the number of robots 30. In the data structure of the robot database 70 shown in this figure, the robot data of each record is arranged in the row direction, and the robot data of each record includes fields of "robot identification information", "robot name", "robot information", "position / direction", and "schedule information".

[0075] "Robot identification information" is a field that records data values such as an identification number, a serial number, or other unique numbers or character strings that can identify the corresponding robot 30 within the facility 1. "Robot name" is a field that records the robot name assigned to the corresponding robot 30 by a user or the like.

[0076] "Robot Information" is a field that records various information about the corresponding robot 30. "Position and Direction" is a field that records the position and direction of the corresponding robot 30 within Facility 1, and, if necessary, the direction of each sensor 400 (absolute direction or offset with respect to the direction of robot 30). "Schedule Information" is a field that records the schedule for the corresponding robot 30 to patrol within Facility 1 and collect sounds within Facility 1 or sounds from each device 10, etc.

[0077] FIG. 10 shows the configuration of the monitoring processing unit 650 of the monitoring device 40 according to the present embodiment. The monitoring processing unit 650 includes a sound source separation unit 1000, a device detection unit 1010, a sound source identification unit 1020, a sound anomaly detection unit 1030, an abnormal device identification unit 1040, a device control unit 1050, a robot command unit 1060, a schedule determination unit 1065, a priority update unit 1070, and an acoustic sensor selection unit 1080.

[0078] The sound source separation unit 1000 is connected to the collection information database connection unit 620. The sound source separation unit 1000 reads out, via the collection information database connection unit 620, the collection data in which acoustic data is recorded as measurement data among the collection data registered in the collection information database 50. Thereby, the sound source separation unit 1000 acquires the acoustic data, the date and time when the acoustic data was acquired, and the position and direction where the acoustic data was observed within Facility 1. The sound source separation unit 1000 extracts at least one sound component by performing sound source separation and the like on the acquired acoustic data.

[0079] The device detection unit 1010 is connected to the collection information database connection unit 620. The device detection unit 1010 reads out, via the collection information database connection unit 620, the collection data in which image data is recorded as measurement data among the collection data registered in the collection information database 50. As a result, the device detection unit 1010 acquires the image data, the date and time when the image data was acquired, and the position and direction where the image data was observed within Facility 1. The device detection unit 1010 detects at least one device 10 imaged in the image data by performing image recognition or the like on the image data. The device detection unit 1010 may search for the device 10 within Facility 1 corresponding to the device imaged in the image data by accessing the device database 60 via the device database connection unit 630 and referring to each record of the device data.

[0080] The sound source identification unit 1020 is connected to the sound source separation unit 1000 and the device detection unit 1010. The sound source identification unit 1020 identifies which of the at least one device 10 imaged in the image data is the sound source for each sound component among the at least one sound component included in the acoustic data. The sound source identification unit 1020 according to the present embodiment identifies which of the devices 10 detected from the image data by the device detection unit 1010 is the sound source for each sound component extracted from the acoustic data by the sound source separation unit 1000. The sound source identification unit 1020 instructs the device database connection unit 630 to record the sound component data regarding each sound component in association with the device 10 that is the sound source among the at least one device 10 imaged in the image data in the device database 60. When the device 10 corresponding to any sound component cannot be identified, the sound source identification unit 1020 may receive the designation of the corresponding device 10 or the like by receiving the input of information from the user via the input processing unit 660.

[0081] The abnormal sound detection unit 1030 is connected to the sound source identification unit 1020. The abnormal sound detection unit 1030 detects the degree of abnormality of the sound components using the sound component data associated with each device 10. The abnormal sound detection unit 1030 may detect the degree of abnormality of the sound components of the acoustic data detected by the selected acoustic sensor 440a among a plurality of acoustic sensors (in this embodiment, a plurality of acoustic sensors 400a as an example). The abnormal sound detection unit 1030 may detect the degree of abnormality of the sound components of the acoustic data detected by the acoustic sensor 400a selected by the acoustic sensor selection unit 1080 described later, or may detect the degree of abnormality of the sound components of the acoustic data detected by the acoustic sensor 400a pre-selected in the schedule determined by the robot command unit 1060 described later.

[0082] The abnormal device identification unit 1040 is connected to the abnormal sound detection unit 1030. The abnormal device identification unit 1040 identifies the abnormality of each device 10 based on the acoustic data detected by the selected acoustic sensor 440a among a plurality of acoustic sensors (in this embodiment, a plurality of acoustic sensors 400a as an example), and may identify the abnormality of the device 10 that is the sound source of the sound component based on the degree of abnormality of the sound component detected by the abnormal sound detection unit 1030. The abnormal device identification unit 1040 may identify the abnormality based on the acoustic data detected by the acoustic sensor 400a selected by the acoustic sensor selection unit 1080 described later, or may identify the abnormality based on the acoustic data detected by the acoustic sensor 400a pre-selected in the schedule determined by the robot command unit 1060 described later. Identifying the abnormality of the device 10 may mean identifying whether the device 10 is abnormal, or may mean identifying the abnormal device 10 as abnormal. The abnormal device identification unit 1040 may cause the display processing unit 665 to generate a display screen for displaying the state of each device 10 detected by the robot 30 on the display device 690 and display it on the display device 690. Further, the abnormal device identification unit 1040 may cause the display processing unit 665 to generate a display screen for displaying the determination result of normal or abnormal of each device 10 on the display device 690 and display it on the display device 690.

[0083] The device control unit 1050 is connected to the abnormal device identification unit 1040. In response to the identification of an abnormality in any one of the devices 10 that is the sound component generation source, the device control unit 1050 performs control to address the abnormality of the one device 10 on at least one of the one device 10 or other devices 10 within the facility 1. The device control unit 1050 transmits an instruction to perform control to address the abnormality of the device 10 to the instruction input unit 260 within the control system 20 via the communication unit 670.

[0084] The robot command unit 1060 is connected to the collected information database connection unit 620, the robot database connection unit 640, the acoustic sensor selection unit 1080, and the generation source identification unit 1020. The robot command unit 1060 determines the operations to be performed by each robot 30 and instructs each robot 30 to perform the determined operations.

[0085] In response to the generation source identification unit 1020 being unable to identify the device 10 that is the sound generation source, the robot command unit 1060 instructs the robot 30 to perform at least one of changing the imaging direction of the image, changing the imaging magnification of the image, or changing the position of the robot 30. The robot command unit 1060 may refer to the data recorded in the collected information database 50 and instruct the robot 30 to change the imaging direction or the like. The robot command unit 1060 may read the schedule from the robot database 70 and instruct each robot 30 to perform the content of the schedule corresponding to the current time. The robot command unit 1060 may also instruct the robot 30 to perform an operation in response to an instruction from the acoustic sensor selection unit 1080.

[0086] The schedule determination unit 1065 is connected to the device database connection unit 630 and the robot database connection unit 640. Based on the priority of each acoustic sensor 440a indicated in the priority data for each device 10, the schedule determination unit 1065 determines a schedule for each robot 30 equipped with each acoustic sensor 440a to observe (for example, collect the sound of each device 10). The schedule determination unit 1065 may determine a schedule for each robot 30 (also referred to as a robot-specific schedule), and by determining the robot-specific schedule for each robot 30, may determine a schedule for the entire robots 30 in the facility 1 (also referred to as an overall schedule). The robot-specific schedule may include information necessary to determine the operation of the robot 30, such as the position of each observation point where the target robot 30 should observe the state in the facility 1 (for example, collect sound), the time to arrive at each observation point, the movement path between the observation points, or the observation direction at each observation point, etc. The schedule determination unit 1065 determines an overall schedule for collecting the sound of each device 10 based on the priority of each acoustic sensor 410a for each device 10 read from the device database 60 via the device database connection unit 640, and may record it in the robot database 70 via the robot database connection unit 640.

[0087] The priority update unit 1070 is connected to the abnormal device identification unit 1040 and the input processing unit 660. The priority update unit 1070 updates the priority data in the device database 60 according to the accuracy of abnormal identification based on the acoustic data detected by each of the plurality of acoustic sensors 440a for each device 10. The accuracy may be the accuracy rate (Accuracy, the ratio of how many of the detection results were correct out of the number of times of detecting normal and abnormal), or the precision rate (Precision, the ratio of how many of the times detected as abnormal were actually abnormal), or the recall rate (Recall, the ratio of how many of the times that were actually abnormal were correctly detected as abnormal), or a value combined from these.

[0088] The priority update unit 1070 may acquire from the user whether an abnormality has actually occurred via the input processing unit 660. The priority update unit 1070 may update the priority data based on the accuracy within the period each time the reference period elapses, or may update the priority data based on the accuracy of the previous period in response to a user operation.

[0089] In response to the low accuracy of abnormality identification based on the acoustic data detected for one device 10 by one acoustic sensor 440a, the priority update unit 1070 may update the priority data so that the priority of the one acoustic sensor 440a in the priority data recorded in the device database 60 for the one device 10 is set to a priority lower than the current priority. Note that the initial value of the priority recorded in the device database 60 may be arbitrarily set by the user, may be set uniformly regardless of the combination of the acoustic sensor 440a and the device 10, or may be set according to the combination of the acoustic sensor 440a and the device 10. The initial value of the priority may be set according to the degree of compatibility between the characteristics of the sound generated from the device 10 (for example, the pitch of the sound) and the sound collection characteristics of the acoustic sensor 440a (for example, the frequency characteristics). As an example, it may be set according to how excellent the sound collection performance of the acoustic sensor is within the frequency band of the sound generated from the device 10.

[0090] The acoustic sensor selection unit 1080 is connected to the device database connection unit 630, the sound source identification unit 1020, and the input processing unit 660. The acoustic sensor selection unit 1080 selects an acoustic sensor 440a to be used for abnormality detection of each device 10 from among a plurality of acoustic sensors (in this embodiment, a plurality of acoustic sensors 440a as an example) based on the priority data recorded in the device database 60.

[0091] The sound sensor selection unit 1080 may use the device 10 identified as the sound source of the sound data by the sound source identification unit 1020 as the device 10 (also referred to as the target device 10) whose status is to be checked. The sound sensor selection unit 1080 may use at least one device 10 specified via the input processing unit 660 as the target device 10. The sound sensor selection unit 1080 may extract at least one device 10 located in the area within the facility 1 specified via the input processing unit 660 from the device database 60 and use it as the target device 10.

[0092] The sound sensor selection unit 1080 may select any sound sensor 440a to be used for detecting an abnormality of the target device 10 based on the priority data recorded in the device database 60 for the target device 10. As an example, the sound sensor selection unit 1080 may select the sound sensor 440a with the highest priority for the target device 10. The sound sensor selection unit 1080 may instruct the robot command unit 1060 to move the robot 30 equipped with the selected sound sensor 440a to the position of the target device 10 to collect sound.

[0093] Among the sound data collected by each of the plurality of sound sensors 400a, the sound sensor selection unit 1080 may cause the sound abnormality detection unit 1030 to detect the abnormality degree of the sound data collected by the selected sound sensor 440a, and may cause the abnormal device identification unit 1040 to identify the abnormality of the target device 10 based on the sound data collected by the sound sensor 440a.

[0094] The monitoring processing unit 650 shown above has each component shown in FIG. 10 in order to realize each function such as identification of the device 10 that is the sound source of each sound component, identification of the abnormality of the device 10 of the sound source based on the abnormality degree of the sound component, instruction of control for dealing with the abnormality of the device 10, and instruction of the operation for each robot 30. Alternatively, the monitoring processing unit 650 may adopt a configuration that does not have some of the components shown in FIG. 10.

[0095] For example, the monitoring processing unit 650 may not have the device control unit 1050 and may not issue an instruction for control to handle an abnormality of the device 10. Further, the monitoring processing unit 650 may not have the robot command unit 1060 and may not issue an instruction for an operation to each robot 30. Further, the monitoring processing unit 650 may not have the abnormal sound detection unit 1030 and the abnormal device identification unit 1040 and may not identify an abnormality of the device 10 that is the sound source based on the degree of abnormality of the sound component. Even when the monitoring device 40 does not have some of the functions exemplified herein, the monitoring device 40 allows the user to receive information output or displayed by the monitoring device 40, discover an abnormality of the device 10, and cause at least one device 10 in the facility 1 to take a technical measure to be taken in response to the abnormality of the device 10.

[0096] According to the monitoring device 40 described above, it is possible to collect acoustic data detected in the facility 1 and image data captured in the facility 1, and identify the device 10 that is the sound source of each sound component included in the acoustic data. Thereby, the monitoring device 40 can analyze the potential state of each device 10 that may not be detected from the state of each device 10 collected by the control system 20 from the sound component having the device 10 as the sound source. Further, the monitoring device 40 can promptly collect abnormal sounds emitted by the devices 10 installed in various locations in the facility 1, and can promptly handle abnormalities of the devices 10.

[0097] Also, based on the priority data of the acoustic sensor 440a recorded for each of the plurality of devices 10, the acoustic sensor 440a used for detecting an abnormality of each device 10 is selected from the plurality of acoustic sensors 440a, and an abnormality of each device 10 is identified based on the acoustic data from the selected acoustic sensor 440a. Therefore, even when the acoustic sensor 440a suitable for identifying an abnormality is different for each device 10 due to differences in characteristics among the plurality of acoustic sensors 440a, the abnormality of the device 10 can be identified based on the acoustic data detected by the appropriate acoustic sensor 440a. Thus, the accuracy of abnormality identification can be improved.

[0098] Also, since at least one of the plurality of acoustic sensors 400a has different sound collection characteristics from at least one other acoustic sensor 400a, the priority can be determined based on the sound collection characteristics. Therefore, by selecting the acoustic sensor 400a according to the priority, the abnormality of the device 10 can be specified based on the acoustic data detected by the appropriate acoustic sensor 400a.

[0099] Also, each of the plurality of acoustic sensors 400a is mounted on each of the plurality of robots 30 that can move within the facility 1. Therefore, by moving the robot 30 equipped with the selected acoustic sensor 400a within the facility 1 to collect sound, the abnormality can be specified based on the acoustic data detected by the appropriate acoustic sensor 400a.

[0100] Also, based on the priority of each acoustic sensor 440a indicated in the priority data for each device 10, the overall schedule for each robot 30 equipped with each acoustic sensor 440a to collect the sound of each device 10 is determined. Therefore, by moving each robot 30 within the facility according to the determined overall schedule, the abnormality can be specified based on the acoustic data detected by the appropriate acoustic sensor 440a at the appropriate timing.

[0101] Also, since the priority data is updated according to the accuracy of abnormality specification by the acoustic data detected by each of the plurality of acoustic sensors 440a for each device 10, the priority of the acoustic sensor 440a that can perform abnormality specification more accurately becomes higher. Therefore, the accuracy of specification can be improved by specifying the abnormality based on the acoustic data detected by the acoustic sensor 440a selected based on the priority data.

[0102] Also, in response to the specification of an abnormality in one device 10, control for dealing with the abnormality of the one device 10 is performed on at least one of the one device 10 or other devices 10, so that the abnormality occurring in the one device 10 can be promptly dealt with.

[0103] FIG. 11 shows a processing flow for determining an overall schedule. Note that the schedule determination unit 1065 may start the operation of this processing flow according to a user operation, may perform it for each reference interval, or may perform it each time the priority data is updated by the priority update unit 1070 by a reference number of times.

[0104] In S1000, the schedule determination unit 1065 extracts one or more robots 30 in facility 1 as the robots 30 to be used in the schedule. Note that when the processing after S1000 is repeated and the processing of S1000 is performed multiple times, the schedule determination unit 1065 may select different robots 30 in each S1000, or may select a common robot 30 in two or more S1000s. The schedule determination unit 1065 may select the robots 30 randomly.

[0105] In S1005, the schedule determination unit 1065 selects any one of the extracted robots 30 as the robot 30 for sound collection. Note that when the processing after S1005 is repeated and the processing of S1005 is performed multiple times, the schedule determination unit 1065 may select different robots 30 in each S1005.

[0106] In S1010, the schedule determination unit 1065 selects at least one of the plurality of devices 10 in facility 1 as a sound collection target. The schedule determination unit 1065 refers to the priority data for each device 10 recorded in the device database 60, and excludes from the plurality of devices 10 those devices 10 whose priority of the acoustic sensor 400a mounted on the sound collection robot 30 (also referred to as the acoustic sensor 400a for sound collection) is lower than the reference priority, and may select at least one other device 10 as the sound collection target. When a plurality of acoustic sensors 400a are mounted on the sound collection robot 30, the schedule determination unit 1065 may exclude from the plurality of devices 10 those devices 10 whose priority of each acoustic sensor 400a for sound collection is lower than the reference priority respectively, and select at least one other device 10 as the sound collection target. The reference priority may be arbitrarily set. The schedule determination unit 1065 may randomly select at least one device 10. Note that when the processes after S1005 are repeated and the process of S1010 is performed multiple times, different devices 10 may be selected in each process of S1010, or at least one device 10 may be commonly selected in the processes of S1010 two or more times.

[0107] In S1020, the schedule determination unit 1065 determines the patrol route of the sound collection robot 30. The schedule determination unit 1065 may determine the patrol route so that the sound collection robot 30 collects the sounds of each device 10 that is the sound collection target. The schedule determination unit 1065 determines the sound collection position (also referred to as the observation point) for each device 10 based on the position of each device 10 recorded in the device database 60, and may determine the movement route between the sound collection positions so as to connect the sound collection positions. The schedule determination unit 1065 may determine the movement route between the sound collection positions so that the patrol route is the shortest as a whole. The schedule determination unit 1065 may determine the patrol route including these sound collection positions and movement routes. The schedule determination unit 1065 may make the patrol route an annular route as a whole, and may set the starting point and the return point of the sound collection robot 30 as the charging position of the robot 30.

[0108] In S1030, the schedule determination unit 1065 determines the sound collection time for each device 10 to be sound-collected, that is, the stay time of the robot 30 at the sound collection position, and the moving speed between the sound collection positions. Thereby, the time required for the sound collection robot 30 to circulate along the circulation route (also referred to as the circulation required time) is determined, and a temporary schedule for sound collection by the sound collection robot 30 (also referred to as a temporary robot-specific schedule) is determined. The temporary robot-specific schedule may include the identification information of the acoustic sensor 400a for sound collection, the identification information of the device 10 to be sound-collected, the circulation route, the stay time at each sound collection position, the moving speed between the sound collection positions, the circulation required time, and the charging position of the sound collection robot 30. The temporary robot-specific schedule may not include time information.

[0109] The schedule determination unit 1065 may make the sound collection time for each device 10 to be sound-collected the same, or may make the sound collection times of at least two devices 10 different. The schedule determination unit 1065 may calculate the number or density of the devices 10 to be sound-collected in each area within the facility 1 based on the positions of the respective devices 10 to be sound-collected recorded in the device database 60, and may determine the sound collection time within the area by the sound collection robot 30 based on this calculation result. For example, the schedule determination unit 1065 may determine that the sound collection time for each device 10 is longer in an area where the number of devices 10 is large or the density is high, and may determine that the sound collection time for each device 10 is shorter in an area where the number of devices 10 is small or the density is low. Thereby, for each area within the facility 1, based on the number or density of the devices 10 that are the targets of abnormality detection by the acoustic sensor 400a for sound collection among at least one device 10 installed in the area, the temporary schedule for sound collection within the area by the acoustic sensor 400a is determined.

[0110] The schedule determination unit 1065 may make the moving speed the same for each section between the sound collection positions, or may make the moving speeds of at least two sections different. The schedule determination unit 1065 may determine the moving speed between the sound collection positions to be equal to or lower than the maximum speed of the sound collection robot 30.

[0111] The schedule determination unit 1065 may determine the stay time and the moving speed so that the round-trip time satisfies formula (1) or formula (2). Thereby, the single sound collection robot 30 for sound collection can periodically collect sound from the device 10 to be sound-collected.

[0112] Round-trip time ≤ Allowable confirmation period (1) Round-trip time ≤ Allowable confirmation period - Charging time required (2)

[0113] The allowable confirmation period in formulas (1) and (2) may be the longest period allowed to confirm the state of the device 10, and may be the same among the devices 10, but may also be different between at least two devices 10. When the allowable confirmation periods are different among the devices 10, the shortest allowable confirmation period among the allowable confirmation periods of each device 10 to be sound-collected selected in S1010 may be used in formulas (1) and (2). The charging time required in formula (2) may be the time required for charging the sound collection robot 30, and may be the same among the robots 30, or may be different between at least two robots 30. When the charging times required are different among the robots 30, the charging time required for the sound collection robot 30 selected in S1005 may be used in formula (2).

[0114] When the process after S1005 is repeated and the process of S1030 is performed multiple times, the stay time, the moving speed, and the round-trip time determined in each process of S1030 may be the same as each other, or at least one of the stay time, the moving speed, or the round-trip time determined in at least two processes of S1030 may be different from each other.

[0115] Depending on the device 10 to be sound-collected selected in S1010, the schedule determination unit 1065 may not be able to determine the stay time and the moving speed such that the moving speed between the sound collection positions is equal to or lower than the maximum speed of the sound collection robot 30 and the round-trip time satisfies formula (1) or formula (2). In this case, the schedule determination unit 1065 may determine a temporary robot-specific schedule by associating an error code.

[0116] In S1040, the schedule determination unit 1065 determines whether all the robots 30 extracted in S1000 have been selected. If there are still unselected robots 30 (\"N\" in S1040), the schedule determination unit 1065 returns the process to S1005. As a result, in S1005, the unselected robots 30 are selected as the sound collection robots 30, and a temporary schedule for each robot 30 is determined. Consequently, a temporary schedule (also referred to as a temporary overall schedule) for all the extracted robots 30 is determined. If not all the robots 30 in facility 1 have been extracted in the above-mentioned S1000, the temporary overall schedule may not include the temporary schedules for each robot 30 of all the robots 30. If all the robots 30 have been selected in S1040 (\"Y\" in S1040), that is, if the temporary overall schedule has been determined, the schedule determination unit 1065 advances the process to S1050.

[0117] In S1050, the schedule determination unit 1065 calculates an evaluation value of the determined temporary overall schedule based on a preset evaluation function. The schedule determination unit 1065 may associate the content of the temporary overall schedule with the calculated evaluation value and store it in a storage unit (not shown).

[0118] The schedule determination unit 1065 may extract combinations of the sound collection target device 10 and the sound collection acoustic sensor 440a from the temporary overall schedule, and calculate an evaluation value according to the priority of the sound collection acoustic sensor 440a for the sound collection target device 10. In this case, the evaluation function may be a function in which the evaluation value increases as the total priority increases. When a plurality of acoustic sensors 400a are mounted on the sound collection robot 30, the schedule determination unit 1065 may calculate the evaluation value by excluding the priorities lower than the second highest among the priorities of these plurality of acoustic sensors 400a for the sound collection target device 10. The priority may be read from the device database 60.

[0119] The schedule determination unit 1065 may calculate the total moving distance of each robot 30 when the provisional overall schedule is executed, and calculate an evaluation value according to the total moving distance. In this case, the evaluation function may be a function in which the shorter the total moving distance, the larger the evaluation value.

[0120] The schedule determination unit 1065 may calculate an evaluation value according to the number of robots 30 used in the provisional overall schedule. In this case, the evaluation function may be a function in which the smaller (or larger) the number of robots 30, the larger the evaluation value.

[0121] The schedule determination unit 1065 may extract the devices 10 that are redundantly sound-collected by a plurality of robots 30 when the provisional overall schedule is executed, and calculate an evaluation value according to the number thereof. In this case, the evaluation function may be a function in which the smaller (or larger) the number of devices 10 that are redundantly sound-collected, the larger the evaluation value.

[0122] The schedule determination unit 1065 may extract the number of devices 10 that are not sound-collected when the provisional overall schedule is executed, and calculate an evaluation value according to the number thereof. In this case, the evaluation function may be a function that sets the evaluation value to zero according to the fact that the number of devices that are not sound-collected is 1 or more.

[0123] The schedule determination unit 1065 may calculate an evaluation value according to the number of provisional robot-specific schedules associated with error codes among the provisional robot-specific schedules included in the provisional overall schedule. In this case, the evaluation function may be a function that sets the evaluation value to zero according to the fact that the number of provisional robot-specific schedules associated with error codes is 1 or more.

[0124] In S1060, the schedule determination unit 1065 determines whether the creation end condition of the temporary overall schedule is satisfied. The creation end condition may be, for example, that among all combinations of each device 10 in facility 1 and each robot 30, there is no combination that has not been selected as the device 10 to be sound-collected and the sound-collecting robot 30 in the determined temporary overall schedule. The creation end condition may further include that the number of determined temporary overall schedules with a non-zero evaluation value is equal to or greater than a reference number. The reference number may be arbitrarily set. The creation end condition may be that a preset time for creating the schedule has elapsed since the start of the operation. The creation end condition may be that as a result of repeatedly performing the processes of S1000 to S1060 over a reference time, the maximum value of the evaluation value calculated in S1050 is no longer updated. If the creation end condition is not satisfied in S1060 ("N" in S1060), the schedule determination unit 1065 returns the process to S1000. Thereby, a new temporary overall schedule is determined. The schedule determination unit 1065 may perform S1000 to S1050 so that at least part of the content is different from the determined temporary overall schedule, and determine a new temporary overall schedule. If the creation end condition is satisfied in S1060 ("Y" in S1060), the schedule determination unit 1065 advances the process to S1070.

[0125] In S1070, the schedule determination unit 1065 determines, as the target for use, a temporary overall schedule that satisfies a predetermined condition among the determined temporary overall schedules. The schedule determination unit 1065 may determine any one of the temporary overall schedules as the target for use based on the evaluation value of the temporary overall schedule. The schedule determination unit 1065 may use, as the target for use, the temporary overall schedule with the largest evaluation value, or may use, as the target for use, the temporary overall schedule whose evaluation value is equal to or greater than a reference value and whose total moving distance of each robot 30 when the temporary overall schedule is executed is the shortest.

[0126] In S1080, the schedule determination unit 1065 sets the departure time of the robot 30 at a period equal to or less than the allowable confirmation period for each per-robot schedule included in the temporarily determined overall schedule determined to be the target of use, and determines it as the per-robot schedule. The schedule determination unit 1065 may align the departure times among the per-robot schedules, or may vary the departure times between at least two per-robot schedules. When there is a robot 30 among the robots 30 in Facility 1 for which the per-robot schedule has not been determined within the temporarily determined overall schedule, the schedule determination unit 1065 may determine a per-robot schedule in which the robot 30 waits at a charging position or the like. Thereby, the overall schedule is determined by the plurality of determined per-robot schedules. The schedule determination unit 1065 may record each determined per-robot schedule in the robot database 70 via the robot database connection unit 640.

[0127] In addition, in the above operation, the case where one or a plurality of acoustic sensors 400a are mounted on each robot 30 has been described, but a single acoustic sensor 400a may be mounted on each robot 30. In this case, in S1005, the schedule determination unit 1065 may select the acoustic sensor 400a instead of selecting the robot 30.

[0128] FIGS. 12 and 13 show the processing flow of the monitoring device 40 according to the present embodiment. The processing flows of FIGS. 12 and 13 may be the processing flows when instructions are given to the robot 30 according to the schedule determined by the schedule determination unit 1065. Note that the processing flows of FIGS. 12 and 13 show the case where the monitoring device 40 performs monitoring in Facility 1 using one robot 30 for convenience of explanation. The monitoring device 40 may execute the processing flows of FIGS. 12 and 13 for each of the plurality of robots 30.

[0129] In S1100, the robot command unit 1060 instructs the robot 30 to be processed to move within the facility 1 or the like via the instruction transmission unit 612. In this operation, the robot command unit 1060 may refer to the schedule information recorded in the robot database 70 in association with the robot 30 to be processed, and acquire the schedule for the robot 30 to be processed to patrol within the facility 1. The robot command unit 1060 moves the robot 30 to the next observation point according to the designation in this schedule, and supplies an instruction to perform observation at the next observation point to the robot 30. In response to this, the robot 30 moves to the next observation point (see S520 and S530 in FIG. 5).

[0130] In S1110, the robot 30 observes the state of the facility 1 and the state of the robot 30 using each sensor 400 at the observation point (see S500 in FIG. 5). The robot 30 transmits various measurement data indicating the observed state of the facility 1 and the state of the robot 30 to the monitoring device 40 (see S510 in FIG. 5). The acoustic data acquisition unit 602 in the monitoring device 40 acquires acoustic data from among the measurement data transmitted from the robot 30. When a plurality of acoustic sensors 400a are mounted on the robot 30, the acoustic data acquisition unit 602 may acquire the acoustic data detected by each of the plurality of acoustic sensors 400a. The collected information database connection unit 620 may add the position data acquired by the position data acquisition unit 608, the direction data acquired by the direction data acquisition unit 610, etc. to the acoustic data acquired by the acoustic data acquisition unit 602, and record it in the collected information database 50 as collected data.

[0131] In S1120, the sound source separation unit 1000 extracts at least one sound component by performing sound source separation or the like on the acoustic data from the robot 30. The sound source separation unit 1000 may perform sound source separation by extracting at least one sound component having different directions of the sound source as seen from the robot 30 from the acoustic data. The sound source separation unit 1000 may specify the direction of each sound component as seen from the robot 30 (sound source localization).

[0132] Here, each sensor 400a collects sound from a predetermined range of directions other than the central direction of the measurement of the sensor 400a according to the directivity, so that it collects the sound synthesized from sounds from various sound sources. The sound source separation unit 1000 performs a process of separating or decomposing the sound collected by each sensor 400a into the sounds generated by each sound source. In this specification, for convenience of explanation, the sound obtained by extracting at least a part of the sound collected by the sensor 400a is referred to as a "sound component", but the "sound component" itself is also a sound. Therefore, the acoustic data which is the data of the sound collected by the sensor 400a and the sound component data which is the data of the sound component may be in the same data format.

[0133] The sound source separation unit 1000 may utilize various sound source separation and sound source localization techniques. For example, the robot 30 may have two or more sensors 400a, and the acoustic data received by the monitoring device 40 may include acoustic data for each channel recording the sound collected by each sensor 400a.

[0134] The sound source separation unit 1000 may calculate the direction of the sound source of each sound component with respect to the robot 30 from the time difference at which each sound component included in the acoustic data reaches each sensor 400a. Here, the sound source separation unit 1000 calculates the relative direction of the sound source of each sound component with respect to the measurement direction of the sensor unit including two or more sensors 400a, and adds it to the absolute direction (central direction of measurement) in which the sensor unit itself is facing, thereby calculating the absolute direction of the sound source of each sound component at the observation point. Here, the absolute direction means an angle represented starting from at least a common reference direction (such as north) within the facility 1, such as an angle in the horizontal direction and the vertical direction with respect to north. The relative direction means an angle represented starting from an arbitrarily selected reference direction (such as the traveling direction of the robot 30).

[0135] Further, the robot 30 may have a sensor 400a such as a directional microphone, and may acquire acoustic data while changing the orientation of the sensor 400a within a predetermined range of directions with respect to the robot 30. For example, the robot 30 may acquire acoustic data while changing the orientation of the sensor 400a within the angular range of the image data captured by the sensor 400b. The sound source separation unit 1000 may use such acoustic data to specify the orientation of the sensor 400a when the magnitude (sound pressure) of each sound component is maximized as the direction of the sound source of that sound component.

[0136] By separating the acoustic data into sound sources in this way, the monitoring device 40 can identify the sound components originating from each device 10 even when it has acquired acoustic data in which the sounds from two or more devices 10 are superimposed. As a result, the monitoring device 40 can increase the detection rate of abnormalities in the device 10 and can execute appropriate countermeasures against the abnormalities in the device 10.

[0137] In S1130, the image data acquisition unit 604 in the monitoring device 40 acquires image data from among the measurement data transmitted from the robot 30. The collection information database connection unit 620 may add the position data acquired by the position data acquisition unit 608, the direction data acquired by the direction data acquisition unit 610, etc. to the image data acquired by the image data acquisition unit 604 and record it in the collection information database 50 as collection data.

[0138] In S1140, the device detection unit 1010 detects one or two or more devices 10 imaged in the image data by performing image recognition or the like on the image data from the robot 30. The device detection unit 1010 may extract one or two or more objects imaged in the image data and search the device database 60 for the device 10 corresponding to each object to detect the device 10 corresponding to each object. Here, the device detection unit 1010 may search the device database 60 for the device 10 located in the direction of the robot 30 or the sensor 400b at this timing from the position of the robot 30 at the timing when the image data was acquired.

[0139] The device detection unit 1010 may recognize each device 10 included in the image data using vision-based AR (Augmented Reality) technology. For example, the device detection unit 1010 may recognize each device 10 in the image data using markerless AR technology. The device detection unit 1010 may detect each of at least one device 10 included in the image data by matching with a predetermined template. For example, the device detection unit 1010 extracts one or more objects imaged in the image data and searches the device database 60 for the device 10 associated with the template image data that matches or is most similar to the image of each object. The device detection unit 1010 may detect the device 10 retrieved from the device database 60 for each object as the device 10 located in the image range including that object in the image data. When the template image data associated with the device 10 retrieved from the device database 60 using the position of the robot 30 and the direction of the robot 30 or the sensor 400b at the timing when the image data is acquired matches or is similar to the image of the object in the image range corresponding to the direction of the sensor 400b with respect to the position of the robot 30 in the image data with a similarity equal to or higher than a predetermined threshold, the device detection unit 1010 may determine that the device 10 corresponds to that object.

[0140] The device detection unit 1010 may also recognize each device 10 included in the image data using AR technology that uses markers. In this case, the device data stored in the device database 60 may store the identification information of the marker attached to the corresponding device 10, for example, in the device information field. The device detection unit 1010 may search the device database 60 for the device 10 associated with the device data in which the identification information matching the identification information of the marker attached to the device 10 included in the image data is stored.

[0141] The device detection unit 1010 may detect a device 10 that is not registered in the device database 60. For example, the device detection unit 1010 may detect a device 10 newly installed in the facility 1 or a part of a device 10 that is already registered in the device database 60 as an unregistered device 10. The device detection unit 1010 may estimate the type (e.g., reactor, measuring instrument, pipe, or flange, etc.) or model of the device 10 corresponding to the object from the external shape of the object extracted from the image data. The device detection unit 1010 may detect an unknown device 10 included in the image data by matching it with a template or the like predetermined for each type or model of the device, and specify the type or model of the device. The monitoring device 40 can extract from the image data an unregistered device 10 installed in the facility 1 and a part of a device 10 that has not been recognized as a failure unit capable of generating abnormal sounds and has not been registered as an individual device 10 by performing image recognition on the image data using template matching or the like, and make it detectable as the sound source.

[0142] The device detection unit 1010 may calculate the relative direction of the device 10 with respect to the center direction of the image from the position of the device 10 in the image corresponding to the image data. The device detection unit 1010 may calculate the absolute direction of each device 10 viewed from the observation point by adding the relative direction of the device 10 with respect to the center direction of the image to the absolute direction of the sensor 400b at the time of imaging the image data.

[0143] In S1150, the sound source identification unit 1020 identifies which device 10 imaged in the image data is the sound source of each sound component included in the acoustic data. The sound source identification unit 1020 may identify, for each sound component extracted from the acoustic data, among the devices 10 detected from the image data, the device 10 located in the region corresponding to the direction of the sound source within the image data as the sound source device 10. As an example, for a certain sound component and a certain device 10 detected from the acoustic data and the image data observed by the robot 30 at the same position at the same timing or different timings, the sound source identification unit 1020 may determine that the sound source of this sound component is this device 10 when the absolute direction of the sound component and the absolute direction of the device 10 viewed from the observation point are the same, or when the difference between these absolute directions is within a predetermined error range. For the device 10 already registered in the device database 60, the sound source identification unit 1020 may determine that the sound source of this sound component is this device 10 on the condition that the device 10 is located in the absolute direction of the sound component viewed from the observation point of the acoustic data.

[0144] Also, for a certain sound component and a certain device 10 obtained from acoustic data and image data observed by the robot 30 at different positions at the same timing or different timings, the sound source identification unit 1020 determines that the sound source of this sound component is this device 10 on the condition that the absolute direction of the sound component as seen from the observation point of the acoustic data and the absolute direction of the device 10 as seen from the observation point of the image data intersect or substantially intersect within a predetermined error range. Here, the device detection unit 1010 may calculate an approximate distance range from the observation point of the image data to each device 10 based on the size etc. of each device 10 in the image corresponding to the image data. In this case, in addition to the absolute direction of the sound component as seen from the observation point of the acoustic data and the absolute direction of the device 10 as seen from the observation point of the image data intersecting or substantially intersecting within a predetermined error range, when the distance from the observation point of the image data to this intersection position is within the calculated distance range, the sound source identification unit 1020 may determine that the sound source of this sound component is this device 10. Since the monitoring device 40 can identify the device 10 that is the sound source of each sound component using the local acoustic data and image data observed by the robot 30 at the observation point, the sound information of each device 10 that could not be detected by the measurement by the control system 20 and was buried in the facility 1 can be made apparent as information usable for the management of each device 10.

[0145] In S1155, the acoustic sensor selection unit 1080 selects an acoustic sensor 440a to be used for detecting an abnormality of the specified device 10 from among a plurality of acoustic sensors 400a mounted on the robot 30 based on the priority data recorded in the device database 60. The acoustic sensor selection unit 1080 may select the acoustic sensor 440a having the highest priority for the specified device 10 among the plurality of acoustic sensors 440a mounted on the robot 30. When there are a plurality of acoustic sensors 440a having the highest priority, the acoustic sensor selection unit 1080 may select these acoustic sensors 440a in order each time the process of S1155 is performed, or may continue to select the same acoustic sensor 400a. Note that when only a single acoustic sensor 400a is mounted on the robot 30, the process of S1155 does not need to be performed.

[0146] In S1160, the sound source identification unit 1020 instructs the device database connection unit 630 to record the sound component data regarding each sound component in the device database 60 in association with the device 10 that is the sound source. In response to the instruction from the sound source identification unit 1020, the device database connection unit 630 records the sound component data regarding each sound component in the acoustic data field in the record of the device data corresponding to the device 10 that is the sound source in the device database 60. When any one of the acoustic sensors 400a within the robot 30 is selected in the above-described S1155, the device database connection unit 630 may record the sound component data detected by the selected acoustic sensor 400a. The device database connection unit 630 may record a plurality of sound component data acquired at different dates and times in the acoustic data field by adding the acquisition date and time of the sound component data and the sound component data to the acoustic data field.

[0147] If the sound source of a certain sound component is a device 10 that is included in the image data but not registered in the device database 60, the sound source identification unit 1020 may register this device 10 in the device database 60 through the input of information from the user by the input processing unit 660, and record the sound component data in the device database 60 in association with this device 10. This process will be described later in relation to S1250.

[0148] In S1200, the sound anomaly detection unit 1030 detects the anomaly degree of the sound component using the sound component data associated with each device 10. Here, the anomaly degree may be represented by a real number or an integer within a predetermined range such as from 0 to 1 or from 0 to 100%, or may be a binary value such as normal or abnormal.

[0149] When the acoustic data emitted by the device 10 when the device 10 is normal is recorded in the device database 60, the sound abnormality detection unit 1030 may calculate, as the abnormality degree, a value indicating how different the sound component data of the target having the device 10 as the sound source is from the acoustic data emitted by the device 10 during normal operation. For example, the sound abnormality detection unit 1030 may determine the abnormality degree according to the comparison result between the acoustic data during normal operation and the sound component data of the target.

[0150] The sound abnormality detection unit 1030 may perform the comparison between the acoustic data during normal operation and the sound component data of the target in either the time domain or the frequency domain. When performing the comparison in the time domain, the sound abnormality detection unit 1030 adjusts the phase difference so that the time integral of the difference (such as the absolute value of the difference) between the acoustic data during normal operation and the sound component data of the target is minimized, and may use the time integral of the difference between the acoustic data and the sound component data of the target at the adjusted phase difference as the abnormality degree. In this case, the sound abnormality detection unit 1030 may adjust the amplitude so that, for example, the average amplitudes match in order to match the loudness of the sound of the acoustic data and the sound component data of the target.

[0151] When performing the comparison between the acoustic data during normal operation and the sound component data of the target in the frequency domain, the sound abnormality detection unit 1030 may calculate the abnormality degree by integrating, in the frequency direction, the difference (such as the absolute value of the difference) for each frequency between the frequency spectrum of the acoustic data during normal operation and the frequency spectrum of the sound component data of the target. In this case, the sound abnormality detection unit 1030 may perform an adjustment to match the loudness so that the acoustic data and the sound component data of the target can be compared.

[0152] When the acoustic data collected in the past from the device 10 is recorded as one or more histories in the acoustic data field in the device database 60, the sound anomaly detection unit 1030 may calculate, as the degree of anomaly, a value indicating how different the target sound component data having the device 10 as the sound source is from the one or more acoustic data collected in the past from the device 10. The sound anomaly detection unit 1030 may calculate the degree of anomaly of the target sound component data with respect to the acoustic data collected in the past in the same way as the calculation method when calculating the degree of anomaly of the target sound component data with respect to the acoustic data emitted by the device 10 during normal times described above.

[0153] The sound anomaly detection unit 1030 may determine the degree of anomaly according to the degree of deviation of the target sound component data from the distribution of the acoustic data collected in the past from the device 10. Further, the sound anomaly detection unit 1030 may calculate, as the degree of anomaly, the sum of the degrees of difference, etc. for each of a predetermined number of acoustic data having the smallest difference from the target sound component data among the acoustic data collected in the past from the device 10 (k-nearest neighbor method). In addition to the above, the sound anomaly detection unit 1030 may calculate the degree of anomaly of the acoustic data and the target sound component data using various methods for calculating the difference, difference, similarity, or degree of deviation, etc. between two sounds.

[0154] The sound anomaly detection unit 1030 may detect the degree of anomaly of the sound components using AI technology. The sound anomaly detection unit 1030 may receive the sound components emitted by each device 10 from the sound source identification unit 1020 and cause them to be learned by a machine learning model for each device 10. The sound anomaly detection unit 1030 may store the machine learning model of each device 10 or the learned parameters representing the machine learning model in the device information field of the device database 60, etc., and update the machine learning model each time it newly receives the sound components emitted by the device 10. Further, the sound anomaly detection unit 1030 may receive the determination result of normal or abnormal for each sound component. In this case, the sound anomaly detection unit 1030 may generate or update the machine learning model by learning so that the correct determination result is output when the sound component is input to the machine learning model.

[0155] The abnormal sound detection unit 1030 may use a neural network, statistical learning, or other machine learning algorithms. The abnormal sound detection unit 1030 may use the time-series data of the sound components as the input to the machine learning model, and may use the frequency spectrum of the sound components as the input to the machine learning model. The abnormal sound detection unit 1030 updates the parameters of the machine learning model so as to reduce the error between the output of the model when the samples of each sound component serving as learning data are input to the machine learning model and the learning label indicating whether the sound component is normal or abnormal. For example, when using a neural network, the abnormal sound detection unit 1030 inputs the data value at each time or the data value at each frequency in the sound component to each input node of the input layer of the neural network. The abnormal sound detection unit 1030 uses the error between the output value output by the neural network in response to inputting each sample and the label, and adjusts the weights between the neurons of the neural network and the biases of each neuron by a method such as backpropagation.

[0156] Note that the machine learning model may output the abnormality degree of the sound component in response to inputting the acoustic data collected in the past from the device 10 or the normal acoustic data and the sound component data from the device 10. Further, the machine learning model may be prepared for each type or model of the device 10 instead of being prepared for each individual of the device 10. Further, the machine learning model may be common to all the devices 10.

[0157] For example, by using the machine learning model learned in this way, the abnormal sound detection unit 1030 reads out the machine learning model corresponding to the device 10 from the device database 60 in response to receiving the sound component having the device 10 as the sound source, inputs the sound component, and can use the determination result of normal or abnormal output by the machine learning model as the abnormality degree. The abnormal sound detection unit 1030 may calculate the abnormality degree based on the difference between the normal acoustic data and the target sound component data using any other method.

[0158] In addition, when the device 10 is not registered in the device database 60 but the type or model of the device 10 has been estimated, the abnormal sound detection unit 1030 may determine the abnormality degree of the target sound component data based on the acoustic data emitted by a device 10 of the same type or model as the estimated type or model when it is normal or the acoustic data that has occurred in the past. Further, when any one of the acoustic sensors 400a in the robot 30 is selected in S1155 described above, the abnormal sound detection unit 1030 may use the sound component data detected by the selected acoustic sensor 400a as the sound component data to be detected for the abnormality degree. Thereby, in S1210 described later, the abnormality of each device 10 is specified based on the acoustic data detected by the selected acoustic sensor 440a.

[0159] In S1210, the abnormal device specifying unit 1040 specifies the abnormality of the device 10 that is the sound source of the sound component based on the abnormality degree of the sound component associated with the device 10. For example, when the abnormality degree of the sound component exceeds a predetermined threshold value, the abnormal device specifying unit 1040 may specify that the device 10 that is the sound source of the sound component is abnormal. When the abnormality degree of the sound component is equal to or less than a predetermined threshold value, the abnormal device specifying unit 1040 may specify that the device 10 that is the sound source of the sound component is normal.

[0160] In S1220, the abnormal device specifying unit 1040 may cause the display processing unit 665 to display on the display device 690 a display screen that displays the state of each device 10 detected by the robot 30. The abnormal device specifying unit 1040 may cause the display processing unit 665 to display on the display device 690 a display screen that displays information (observation location, direction, image, etc.) regarding one or more devices 10 that have been specified as the sound source of a certain sound component data but are not registered in the device database 60. Further, the abnormal device specifying unit 1040 may cause the display processing unit 665 to display on the display device 690 a display screen that displays the observation location, the direction of the sound component data, etc. for the sound component data that is not associated with any device 10. When there is an error in the state of each device 10 displayed on the display device 690, an input indicating that the displayed state is incorrect may be made for at least one device 10 via the input processing unit 660.

[0161] In S1230, the monitoring device 40 determines whether there is sound component data that is not associated with any of the devices 10. When all the sound component data is respectively associated with any of the devices 10 (i.e., "N" in S1230), the monitoring device 40 advances the process to S1240. The monitoring device 40 may advance the process to S1240 when all the sound component data is associated with any of the devices 10 registered in the device database 60.

[0162] In S1240, when an abnormality of any of the devices 10 that is the sound source of the sound component is identified, the device control unit 1050 performs control for dealing with the abnormality of the device 10 on at least one of the device 10 in which the abnormality is identified or other devices 10. The device control unit 1050 transmits an instruction to perform control for dealing with the abnormality of the device 10 to the control system 20 via the communication unit 670.

[0163] As an example, the device control unit 1050 may instruct the control system 20 to perform control to stop the operation of the device 10 in which the abnormality is identified. The device control unit 1050 may instruct the control system 20 to perform control to stop other devices 10 related to the device 10 in which the abnormality is identified (for example, devices 10 in the upstream or downstream process of the device 10 in which the abnormality is identified, etc.). The device control unit 1050 may instruct the control system 20 to perform control to stop the operations of both the device 10 in which the abnormality is identified and other devices 10 related to the device 10 in which the abnormality is identified. The content of the control instructed to the control system 20 may be preset for each device 10, and the device control unit 1050 may instruct the control system 20 to perform the control corresponding to the content associated with the device 10 in which the abnormality is identified. The monitoring device 40 advances the process to S1100 and continues the monitoring of the facility 1 by the robot 30.

[0164] If there is sound component data that is not associated with any of the devices 10 in S1230 (the "Y" in S1230), the monitoring device 40 proceeds with the process to S1250. Here, the monitoring device 40 may also proceed with the process to S1250 even when the sound component data is associated with a device 10 that is not registered in the device database 60.

[0165] In S1250, in response to the input processing unit 660 being unable to identify the device 10 located in the image area corresponding to the direction of the sound source in the acoustic data for any of the sound components, the input processing unit 660 accepts input of information from the user for identifying the device 10. Here, if the device 10 associated with the sound component is unregistered and its presence in Facility 1 has not been confirmed, the input processing unit 660 accepts input of information regarding at least a part of the device data for the unregistered device 10, and the sound source identification unit 1020 may register the device data of the device 10 corresponding to the received information in the device database 60. In response to such input of information (the "Y" in S1260), the monitoring device 40 proceeds with the process to S1150. Thereby, the sound source identification unit 1020 can identify the sound source of each sound component including the newly registered device 10.

[0166] When the device 10 associated with the sound component is unregistered, the input processing unit 660 may accept input of information indicating whether to register the device 10 in the device database 60. The sound source identification unit 1020 may register the device data of the device 10 corresponding to the type or model, location or approximate location of the device 10 detected from the image data, or other already known information in the device database 60.

[0167] Also, the input processing unit 660 may accept input of information designating the device 10 of the sound source for the sound component data that is not associated with any of the devices 10. In response to such input of information (the "Y" in S1260), the monitoring device 40 proceeds with the process to S1150. Thereby, the sound source identification unit 1020 can identify that the sound source of the sound component data is the designated device 10.

[0168] In S1270, in response to the fact that the robot command unit 1060 cannot identify the device 10 that is the sound source among the devices 10 for any sound component, the robot command unit 1060 instructs the robot 30 to perform at least one of changing the imaging direction of the image, changing the imaging magnification of the image, or changing the position of the robot 30. In the processing flow of this figure, the robot command unit 1060 performs the processing of S1270 when there is no input of information regarding the device 10 corresponding to the sound component ( "N" in S1260). Alternatively, when the robot command unit 1060 cannot identify the device 10 that is the sound source of any sound component, it may automatically instruct the robot 30 to change the imaging direction of the image, change the imaging magnification of the image, or change the position of the robot 30, etc.

[0169] For example, when it cannot be determined which of the two devices 10 is the sound source of a certain sound component, the robot command unit 1060 may instruct the robot 30 to move closer to the two devices 10. Also, when there is a sound component from outside the angle of view of the image data, the robot command unit 1060 may instruct the robot 30 to approach the direction of the sound component in terms of the imaging direction of the image, reduce the imaging magnification of the image to image the range including the direction of the sound component, or move toward the sound component outside the angle of view, etc. The monitoring device 40 proceeds with the processing to S1110 after instructing the robot 30, and performs the processing from S1110 to S1220 on the acoustic data and image data newly acquired by the robot 30. Thereby, even in a situation where the monitoring device 40 cannot determine which device 10 is the sound source in the observation from one observation point or under one observation condition, by changing the observation point or observation condition and repeating the observation, the accuracy of the association between the sound component and the device 10 that is the sound source can be further improved.

[0170] FIG. 14 shows another processing flow of the monitoring device 40 according to the present embodiment. The processing flow in FIG. 14 may be a processing flow when a device 10 or an area in which an abnormality may have occurred in the facility 1 is specified. The processing in this figure may be started in response to an operation by the user to the input device 690 to cause the robot 30 to operate and check the state of any device 10 or any area in the facility 1. The processing in this figure may be performed in parallel with the processing in FIGS. 11 and 12, or may be performed by interrupting the processing in FIGS. 11 and 12.

[0171] In S1300, in response to the user operation, the input processing unit 660 instructs the monitoring processing unit 650 to cause the robot 30 to check the state of any device 10 or any area in the facility 1.

[0172] In S1310, the acoustic sensor selection unit 1080 of the monitoring processing unit 650 selects, from among a plurality of acoustic sensors (in this embodiment, a plurality of acoustic sensors 440a as an example), the acoustic sensors 440a used for detecting abnormalities of each target device 10 that is the target of the status check among the plurality of devices 10 in the facility 1. The acoustic sensor selection unit 1080 may use, as the target device 10, at least one device 10 specified via the input processing unit 660, or may extract, from the device database 60, at least one device 10 located in the area in the facility 1 specified via the input processing unit 660 and use it as the target device 10. The acoustic sensor selection unit 1080 may select the acoustic sensors 440a used for detecting abnormalities of the target device 10 based on the priority data recorded in the device database 60. The acoustic sensor selection unit 1080 may select the acoustic sensor 440a having the highest priority for the target device 10 from among the plurality of acoustic sensors 440a in the facility 1.

[0173] In S1320, the robot command unit 1060 instructs the robot 30 equipped with the selected acoustic sensor 440a to move within the facility 1 and so on via the instruction transmission unit 612. The robot command unit 1060 may instruct the corresponding robot 30 to head towards the area of the target device 10 and to acquire the acoustic data and image data of the target device 10 and so on. The monitoring device 40 may proceed with the process to S1110 described above following S1320. In addition, when S1155 is performed after S1110, the acoustic sensor selection unit 1080 may continue to select the acoustic sensor 440a selected in S1310. Further, when S1240 is performed after S1110, the monitoring device 40 may end its operation.

[0174] Figure 15 shows an example of the image data 1300 captured by the robot 30 according to the present embodiment. In the image data 1300, devices 10 such as a plurality of pipes 1305a to h, a valve 1310, a reactor 1320, and a flow meter 1330 are captured. In the example of this figure, there is an abnormality in the valve 1310 and it is emitting abnormal sounds.

[0175] The monitoring device 40 receives the collected data including the image data 1300 and the acoustic data obtained by collecting the sound from within the viewing angle range of the image data 1300 (corresponding to S1110 and S1130 in FIG. 12). The sound source separation unit 1000 separates the acoustic data by sound source and extracts the sound components from each direction within the viewing angle range of the image data 1300 (corresponding to S1120). The device detection unit 1010 detects each device 10 captured in the image data 1300 by performing image recognition and so on on the image data 1300 (corresponding to S1140).

[0176] When the direction of a certain sound component coincides with (or coincides within a predetermined error range) the direction corresponding to the area where a certain device 10 is located in the image data 1300, the sound source generation location determination unit 1020 may determine that the sound source of the sound component is the device 10 (corresponding to S1150). The abnormal sound detection unit 1030 detects the degree of abnormality of the sound component using the sound component data associated with each device 10 (corresponding to S1200). In the example of this figure, the abnormal sound detection unit 1030 detects a high degree of abnormality for the sound component from the valve 1310 that is emitting an abnormal sound. As a result, the abnormal device identification unit 1040 can identify that the valve 1310 is abnormal based on the degree of abnormality of the sound component.

[0177] Here, when the pipes 1305a to 1305c and the valve 1310 are newly installed in the facility 1 and are not registered in the device database 60, the device detection unit 1010 may estimate that the pipes 1305a to 1305c are pipes and the valve 1310 is a valve based on the external shape of the object extracted from the image data 1300. The abnormal sound detection unit 1030 may detect the degree of abnormality of the sound component based on the result of comparing the sound component data from the valve 1310 with the acoustic data emitted when the valve is normal or the acoustic data emitted by one or more other valves in the facility 1 in the past.

[0178] Note that depending on the type of the sensor 400a used by the robot 30, the sound source separation unit 1000 may be able to separate the sound source in the horizontal direction, for example, but may not be able to separate the sound source in the vertical direction. In such a case, the sound source generation location determination unit 1020 may identify the device 10 that is the sound source of each sound component while appropriately changing the imaging direction and position of the robot 30 (S1270 in FIG. 13) using the horizontal position of each device 10 in the image data 1300 and the direction of each sound component in the horizontal direction.

[0179] Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus having a role of performing operations. Specific stages and sections may be implemented by a dedicated circuit, a programmable circuit supplied with computer-readable instructions stored on a computer-readable medium, and / or a processor supplied with computer-readable instructions stored on a computer-readable medium. The dedicated circuit may include digital and / or analog hardware circuits, and may include integrated circuits (ICs) and / or discrete circuits. The programmable circuit may include a reconfigurable hardware circuit including memory elements such as logical AND, logical OR, logical XOR, logical NAND, logical NOR, and other logical operations, flip-flops, registers, field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), etc.

[0180] A computer-readable medium may include any tangible device capable of storing instructions executable by an appropriate device, and as a result, a computer-readable medium having instructions stored therein will comprise a product including instructions executable to create means for performing the operations specified in the flowchart or block diagram. Examples of computer-readable media may include electronic storage media, magnetic storage media, optical storage media, electromagnetic storage media, semiconductor storage media, etc. More specific examples of computer-readable media may include floppy (registered trademark) disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray (registered trademark) disc, memory stick, integrated circuit card, etc.

[0181] Computer-readable instructions may include source code or object code written in any combination of one or more programming languages, including assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or object-oriented programming languages such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages such as the "C" programming language or similar programming languages.

[0182] Computer-readable instructions may be provided locally or via a wide area network (WAN) such as a local area network (LAN), the Internet, etc. to a processor or programmable circuit of a programmable data processing apparatus such as a general-purpose computer, a special-purpose computer, or other computers, and the computer-readable instructions may be executed to create means for performing the operations specified in a flowchart or block diagram. Examples of processors include computer processors, processing units, microprocessors, digital signal processors, controllers, microcontrollers, etc.

[0183] FIG. 16 shows an example of a computer 2200 in which multiple aspects of the present invention may be embodied, in whole or in part. Programs installed on the computer 2200 can cause the computer 2200 to function as an operation associated with the apparatus according to an embodiment of the present invention or as one or more sections of the apparatus, or can cause the operation or the one or more sections to be executed, and / or can cause the computer 2200 to execute a process according to an embodiment of the present invention or a stage of the process. Such a program may be executed by the CPU 2212 to cause the computer 2200 to execute specific operations associated with some or all of the blocks of the flowcharts and block diagrams described herein.

[0184] The computer 2200 according to this embodiment includes a CPU 2212, a RAM 2214, a graphic controller 2216, and a display device 2218, which are interconnected by a host controller 2210. The computer 2200 also includes input / output units such as a communication interface 2222, a hard disk drive 2224, a DVD-ROM drive 2226, and an IC card drive, which are connected to the host controller 2210 via an input / output controller 2220. The computer also includes legacy input / output units such as a ROM 2230 and a keyboard 2242, which are connected to the input / output controller 2220 via an input / output chip 2240.

[0185] The CPU 2212 operates according to programs stored in the ROM 2230 and the RAM 2214, thereby controlling each unit. The graphic controller 2216 acquires image data generated by the CPU 2212 in a frame buffer provided in the RAM 2214 or the like or in itself, and causes the image data to be displayed on the display device 2218.

[0186] The communication interface 2222 communicates with other electronic devices via a network. The hard disk drive 2224 stores programs and data used by the CPU 2212 in the computer 2200. The DVD-ROM drive 2226 reads a program or data from the DVD-ROM 2201 and provides the program or data to the hard disk drive 2224 via the RAM 2214. The IC card drive reads programs and data from an IC card and / or writes programs and data to the IC card.

[0187] The ROM 2230 stores therein a boot program or the like executed by the computer 2200 upon activation, and / or a program dependent on the hardware of the computer 2200. The input / output chip 2240 may also be connected to the input / output controller 2220 via various input / output units through a parallel port, a serial port, a keyboard port, a mouse port, etc.

[0188] The program is provided by a computer-readable medium such as a DVD-ROM 2201 or an IC card. The program is read from the computer-readable medium, installed in the hard disk drive 2224, the RAM 2214, or the ROM 2230 which are also examples of computer-readable media, and executed by the CPU 2212. The information processing described in these programs is read by the computer 2200, resulting in the cooperation between the programs and the various types of hardware resources described above. The apparatus or method may be configured by realizing the operation or processing of information according to the use of the computer 2200.

[0189] For example, when communication is executed between the computer 2200 and an external device, the CPU 2212 may execute a communication program loaded in the RAM 2214 and instruct the communication interface 2222 to perform communication processing based on the processing described in the communication program. The communication interface 2222 reads the transmission data stored in the transmission buffer processing area provided in a recording medium such as the RAM 2214, the hard disk drive 2224, the DVD-ROM 2201, or the IC card under the control of the CPU 2212, transmits the read transmission data to the network, or writes the received data received from the network to the reception buffer processing area or the like provided on the recording medium.

[0190] Further, the CPU 2212 may cause all or necessary portions of files or databases stored in external recording media such as a hard disk drive 2224, a DVD-ROM drive 2226 (DVD-ROM 2201), an IC card, etc. to be read into the RAM 2214, and may execute various types of processing on the data on the RAM 2214. The CPU 2212 then writes back the processed data to the external recording media.

[0191] Various types of information such as various types of programs, data, tables, and databases may be stored in the recording media and may undergo information processing. The CPU 2212 may perform various types of processing on the data read from the RAM 2214, including various types of operations, information processing, conditional judgments, conditional branches, unconditional branches, information search / replacement, etc. described throughout this disclosure and specified by the instruction sequence of the program, and write back the results to the RAM 2214. Further, the CPU 2212 may search for information in files, databases, etc. within the recording media. For example, when a plurality of entries each having an attribute value of a first attribute associated with an attribute value of a second attribute are stored in the recording media, the CPU 2212 searches for an entry that matches the condition where the attribute value of the first attribute is specified from among the plurality of entries, reads the attribute value of the second attribute stored within the entry, and thereby may obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0192] The programs or software modules described above may be stored in a computer-readable medium on or near the computer 2200. Also, a recording medium such as a hard disk or RAM provided within a server system connected to a dedicated communication network or the Internet can be used as a computer-readable medium, thereby providing the program to the computer 2200 via the network.

[0193] As described above, the present invention has been described using embodiments, but the technical scope of the present invention is not limited to the scope described in the above embodiments. It is obvious to those skilled in the art that various changes or improvements can be made to the above embodiments. It is clear from the description of the claims that forms with such changes or improvements can also be included in the technical scope of the present invention.

[0194] It should be noted that in the claims, the specification, and the drawings, the execution order of each process such as operations, procedures, steps, and stages in the apparatus, system, program, and method shown is not explicitly stated as "earlier" or "preceding", etc., and can be realized in any order unless the output of the previous process is used in the subsequent process. Regarding the operation flow in the claims, the specification, and the drawings, even if it is described using "first," "next," etc. for convenience, it does not mean that it is essential to implement in this order.

Explanation of Reference Numerals

[0195] 1 Facility 10 Equipment 20 Control System 30 Robot 40 Monitoring Device 50 Collected Information Database 60 Equipment Database 70 Robot Database 210 State Acquisition Unit 220 State Determination Unit 230 Display Processing Unit 240 Display Device 250 Input Device 260 Instruction Input Unit 270 Control Unit 400a~c Sensor 410a~b Sensor 420 Actuator 430 State Acquisition Unit 440 Communication Unit 450 Control Unit 600 Communication Unit 602 Acoustic Data Acquisition Unit 604 Image data acquisition unit 606 LiDAR data acquisition unit 608 Position data acquisition unit 610 Direction data acquisition unit 612 Instruction transmission unit 620 Collection information database connection unit 630 Equipment database connection unit 640 Robot database connection unit 650 Monitoring processing unit 660 Input processing unit 665 Display processing unit 670 Communication unit 680 Input device 690 Display device 1000 Sound source separation unit 1010 Equipment detection unit 1020 Sound source location unit 1030 Abnormal sound detection unit 1040 Abnormal equipment identification unit 1050 Equipment control unit 1060 Robot command unit 1065 Schedule determination unit 1070 Priority update unit 1080 Acoustic sensor selection unit 1300 Image data 1305a~h Pipes 1310 Valve 1320 Reactor 1330 Flow meter 2200 Computer 2201 DVD-ROM 2210 Host controller 2212 CPU 2214 RAM 2216 Graphics controller 2218 Display device 2220 Input / output controller 2222 Communication interface 2224 Hard disk drive 2226 DVD-ROM drive 2230 ROM 2240 Input / Output Chip 2242 Keyboard

Claims

1. A device database connection unit connected to a device database that records priority data indicating the priority of acoustic sensors used for anomaly detection for each of a plurality of devices in a facility; An acoustic data acquisition unit that acquires acoustic data detected by each of a plurality of acoustic sensors in the facility; An acoustic sensor selection unit that selects an acoustic sensor used for anomaly detection of each device from among the plurality of acoustic sensors based on the priority data recorded in the device database; An abnormal device identification unit that identifies an abnormality of each device based on the acoustic data detected by the selected acoustic sensor and comprising a device.

2. The device according to claim 1, wherein at least one of the plurality of acoustic sensors has different sound collection characteristics from at least one other acoustic sensor.

3. The device according to claim 2, wherein each of the plurality of acoustic sensors is mounted on each of a plurality of robots capable of moving within the facility.

4. The device according to claim 3, further comprising a schedule determination unit that determines a schedule for each robot equipped with an acoustic sensor to collect the sound of each device based on the priority of each acoustic sensor indicated in the priority data for each device.

5. The device according to claim 1, further comprising a priority update unit that updates the priority data according to the accuracy of anomaly identification based on the acoustic data detected by each of the plurality of acoustic sensors for each device.

6. The device according to claim 1, further comprising a device control unit that performs control for dealing with an abnormality of a certain device on at least one of the certain device or other devices in the facility in response to the identification of the abnormality of the certain device.

7. An acoustic data acquisition step of acquiring acoustic data detected by each of a plurality of acoustic sensors in a facility; An acoustic sensor selection step of selecting an acoustic sensor used for anomaly detection of each device from among the plurality of acoustic sensors based on priority data indicating the priority of acoustic sensors used for anomaly detection for each of the plurality of devices in the facility; An abnormal device identification step of identifying an abnormality of each device based on the acoustic data detected by the selected acoustic sensor and comprising a method.

8. A computer, A device database connection unit connected to a device database that records priority data indicating the priority of acoustic sensors used for anomaly detection for each of a plurality of devices in a facility; An acoustic data acquisition unit that acquires acoustic data detected by each of a plurality of acoustic sensors in the facility; An acoustic sensor selection unit that selects an acoustic sensor used for anomaly detection of each device from among the plurality of acoustic sensors based on the priority data recorded in the device database; An abnormal device identification unit that identifies an abnormality of each device based on the acoustic data detected by the selected acoustic sensor A program that functions as.

Images