Anomaly detection system and anomaly detection method

Abstract

To efficiently and appropriately detect abnormality of a sound generation device.SOLUTION: An abnormality detection system 1 moves to sound collection positions of sound generation devices SG1 to SG4 which convert a sound signal from a sound source SS into a sound and emit the sound. The system includes: an unmanned aerial vehicle 50 for acquiring sound data including sound components of the sound generators SG1 to SG4; and an abnormality determination device 10 for performing determination for abnormality of the sound generators SG1 to SG4 by analyzing the sound data acquired by the unmanned aerial vehicle 50.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present disclosure relates to an abnormality detection system that detects an abnormality in a sound generating device that emits sound, and an abnormality detection method that detects an abnormality in a sound generating device. [Background technology]

[0002] It is known that a mobile object such as a drone is used to detect whether there is an abnormality in a structure (e.g., a building, a bridge, etc.) or equipment that is located in a place that is difficult for humans to access. For example, Patent Document 1 discloses a method in which an aircraft is used to knock on the wall of a structure to be inspected that is located at a high position, and based on the sound generated by knocking on the wall, whether there is any damage inside the structure is detected. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-139929 A Summary of the Invention [Problem to be solved by the invention]

[0004] On the other hand, in facilities (called public address facilities) that amplify important public broadcasts such as disaster broadcasts using sound generating devices such as speakers, it is known to collect sounds emitted from the sound generating devices and determine whether there is an abnormality in the sound generating devices from the collected sounds. Conventionally, workers would go near the sound generating devices to collect the sounds. In the conventional method, in facilities where many sound generating devices are installed, workers must travel to each sound generating device, which takes time to collect the sounds and places a heavy burden on the workers.

[0005] One possible method for detecting an abnormality in the sound generating device is to detect whether or not a sound signal from a sound source is properly transmitted to the sound generating device. This method eliminates the need to have an operator go to the sound generating device. However, even if a sound signal is properly transmitted from the sound source to the sound generating device, there may be an abnormality in the sound generating mechanism of the sound generating device (for example, in the diaphragm (cone paper) if the sound generating device is a speaker) and sound may not be properly emitted. In this way, whether or not a sound signal is transmitted cannot be used to properly determine whether or not an abnormality in the sound generating device exists.

[0006] In view of the above-mentioned points, an object of the present disclosure is to efficiently and appropriately detect an abnormality in a sound generating device. [Means for solving the problem]

[0007] In order to solve the above problems, according to one aspect of the present disclosure, an abnormality detection system that detects an abnormality in a sound generating device that converts a sound signal from a sound source into sound and emits it includes a mobile body that moves to a sound collection position of the sound generating device and acquires sound data including sound components of the sound generating device, and an abnormality judgment device that analyzes the sound data acquired by the mobile body to make a judgment regarding an abnormality in the sound generating device.

[0008] According to another aspect of the present disclosure, an abnormality detection method for detecting an abnormality in a sound generating device that converts a sound signal from a sound source into sound and emits it includes the steps of moving a mobile object to a sound collection position of the sound generating device, having the mobile object moved to the sound collection position of the sound generating device acquire sound data including sound components of the sound generating device, and making a determination regarding the abnormality in the sound generating device by analyzing the sound data acquired by the mobile object. Effect of the Invention

[0009] The anomaly detection system and anomaly detection method disclosed herein move a moving object to a sound collection position of a sound-generating device and use the moving object to collect the sound of the sound-generating device, thereby enabling efficient acquisition of sound data and efficient and appropriate judgment regarding an abnormality in the sound-generating device. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 shows an example of the configuration of a public address system. [Diagram 2] FIG. 2 illustrates an example of the overall configuration of an anomaly detection system according to the present disclosure. [Diagram 3] FIG. 3 illustrates an example configuration of the abnormality determination device and the control terminal according to the present disclosure. [Figure 4] FIG. 4 shows a front view of an unmanned aerial vehicle according to the present disclosure. [Diagram 5] FIG. 5 shows an example configuration of an unmanned aerial vehicle according to the present disclosure. [Figure 6] FIG. 6 is a flowchart showing the operation of the anomaly detection system according to the first embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the embodiments will be described in detail with reference to the drawings as appropriate. However, more detailed explanation than necessary may be omitted. For example, detailed explanation of already well-known matters or duplicate explanation of substantially the same configuration may be omitted. This is to avoid the following explanation becoming unnecessarily redundant and to facilitate understanding by those skilled in the art. Note that the attached drawings and the following explanation are provided to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter described in the claims.

[0012] In the following description, a "sound generating device" is a device that converts a sound signal, which is an electrical signal output from a sound source, into sound and emits it to the outside. Specifically, the sound generating device is a speaker. In this case, the sound generating device mainly comprises an electromagnetic coil that converts the sound signal into vibration, and a diaphragm (cone paper) that vibrates due to the vibration generated by the electromagnetic coil to generate sound. The sound source is, for example, a device that automatically outputs pre-recorded voice as a sound signal, a device that converts sound such as a microphone into an electrical signal, or both. The sound source may also include a device that amplifies the electrical signal, such as an amplifier.

[0013] "Abnormality of the sound generating device" refers to, for example, a case where a sound signal is output but no sound is emitted from the sound generating device, or a case where a sound signal is output and a sound is emitted from the sound generating device but the emitted sound does not correspond to the sound signal. "A case where the emitted sound does not correspond to the sound signal" includes, for example, a case where the emitted sound includes a loud noise sound other than the sound corresponding to the sound signal, or a case where the emitted sound is higher or lower than the sound signal (i.e., the frequency of the emitted sound is higher or lower than the frequency of the sound corresponding to the sound signal).

[0014] Each embodiment of the present invention will be described below.

[0015] 1. First embodiment The abnormality detection system 1 described below is, for example, a system for detecting abnormalities in sound generating devices SG1 to SG4 provided in a loudspeaker system 100 as shown in Fig. 1. The loudspeaker system 100 is a system that transmits various types of information by voice to a large number of people, such as a system that amplifies broadcasts important to the public, such as disaster broadcasts, in urban areas, and a system that amplifies advertisements and store guides in commercial facilities.

[0016] The sound generating devices SG1 to SG4 of the loudspeaker equipment 100 are devices that convert sound signals from a sound source SS into sound and emit the sound, and are devices for amplifying voices to people in the installation area A of the loudspeaker equipment 100. The sound generating devices SG1 to SG4 are, for example, speakers. The sound source SS is, for example, owned by the manager of the loudspeaker equipment 100 or the installation area A, and is a device that converts pre-recorded sounds into sound signals and outputs them to the sound generating devices SG1 to SG4. The sound source SS can also convert voices input from a microphone or the like into sound signals and output them to the sound generating devices SG1 to SG4.

[0017] In the loudspeaker system 100, predetermined sounds are emitted from the sound generating devices SG1 to SG4 at predetermined times. The abnormality detection system 1 detects abnormalities in the sound generating devices SG1 to SG4 using the sounds emitted from the sound generating devices SG1 to SG4 at predetermined times.

[0018] The number of sound generating devices SG1 to SG4 provided in the loudspeaker system 100 is not limited to four, and may be any number depending on the size of the installation area A, etc.

[0019] In the loudspeaker system 100, a large number of sound generating devices SG1 to SG4 are generally arranged. For this reason, the abnormality detection system 1 described below aims to efficiently collect sounds from the large number of sound generating devices SG1 to SG4. Specifically, a moving object is moved to the sound collection positions of the sound generating devices SG1 to SG4 to efficiently collect sounds from the moving object. In addition, the abnormality detection system 1 aims to appropriately judge abnormalities in the sound generating devices SG1 to SG4 using collected sound data.

[0020] The sound generating devices SG1 to SG4 of the loudspeaker system 100 are placed in high positions such as on the walls of buildings or on the tops of poles. In order to collect sounds of sufficient volume from the sound generating devices SG1 to SG4, it is preferable that the sound collection positions are in the vicinity of the installation positions of the sound generating devices SG1 to SG4. For this reason, in the anomaly detection system 1 described below, an unmanned flying object 50 is used as a moving object that collects the sounds from the sound generating devices SG1 to SG4.

[0021] 1-1.Configuration 1-1-1. Configuration of anomaly detection system 2 shows a schematic diagram of an overall configuration of an anomaly detection system 1 according to the present embodiment. The anomaly detection system 1 includes an anomaly judgment device 10, a control terminal 30 capable of communicating with the anomaly judgment device 10 via a network N, and an unmanned aerial vehicle 50 (an example of a moving body) capable of wireless communication with the control terminal 30 and whose flight is controlled by the control terminal 30. The network N includes a wired LAN (Local Area Network), a wireless LAN, a WAN (Wide Area Network), and / or the Internet, etc.

[0022] The abnormality judgment device 10 moves the unmanned aerial vehicle 50 to the sound collection position of the sound generating devices SG1-SG4 for which an abnormality is to be detected, and transmits a command (called a sound collection command) to the control terminal 30 to cause the unmanned aerial vehicle 50 to collect the sounds of the sound generating devices SG1-SG4. The abnormality judgment device 10 also analyzes the sounds collected by the unmanned aerial vehicle 50 and makes a judgment regarding abnormalities in each of the sound generating devices SG1-SG4.

[0023] The control terminal 30 receives a sound collection command from the abnormality judgment device 10 and causes the unmanned aerial vehicle 50 to perform an operation in accordance with the sound collection command. Specifically, for example, the unmanned aerial vehicle 50 is automatically or manually flown close to the sound generating devices SG1 to SG4 that are targets for detecting abnormalities, and then the unmanned aerial vehicle 50 is caused to collect sounds from the sound generating devices SG1 to SG4 near the unmanned aerial vehicle 50 and transmit the collected sound data to the abnormality judgment device 10 via the control terminal 30.

[0024] 1-1-2. Configuration of the anomaly detection device The abnormality determination device 10 shown in FIG. 3 is a computer device that functions as a server or the like, and includes a control unit 11, a storage unit 12, and a communication unit 19. The control unit 11 is an electronic circuit such as a CPU (Central Processing Unit). The control unit 11 operates as an arithmetic processing device and a control device, and controls the abnormality determination device 10 according to various programs to execute functions described below. The storage unit 12 is a ROM (Read Only Memory), a RAM (Random Access Memory), a HDD (Hard Disk Drive), an SSD (Solid State Drive), a FeRAM (Ferroelectric Random Access Memory), or the like. The ROM, HDD, SSD, FeRAM, or the like stores programs, calculation parameters, calculation results, and the like used by the control unit 11. The RAM temporarily stores programs used in executing the functions of the control unit 11, parameters that change appropriately during the execution, and the like. The communication unit 19 is, for example, a network interface card for connecting to a wired LAN or the Internet. The communication unit 19 may be a wireless communication interface for connecting to a base station or a communication interface compatible with a wireless LAN.

[0025] The control unit 11 executes the functions of a command unit 111 , a receiving unit 112 , and a determining unit 113 by executing a program read from the storage unit 12 .

[0026] The command unit 111 transmits a command (sound collection command) for causing the unmanned air vehicle 50 to collect the sounds of the sound generating devices SG1 to SG4 to the control terminal 30 via the communication unit 19. The sound collection command includes, for example, a flight route to the sound collection positions of the sound generating devices SG1 to SG4 and a command for collecting the sounds of each of the sound generating devices SG1 to SG4.

[0027] The flight route included in the sound collection command may include information on the order in which the sound generating devices SG1 to SG4 are moved. For example, when moving in the order of sound generating device SG1, sound generating device SG2, sound generating device SG3, and sound generating device SG4, the flight route includes, for example, a flight route from the current position of the unmanned aerial vehicle 50 to sound generating device SG1, a flight route from sound generating device SG1 to sound generating device SG2, a flight route from sound generating device SG2 to sound generating device SG3, a flight route from sound generating device SG3 to sound generating device SG4, and a flight route from sound generating device SG4 to the parking area of ​​the unmanned aerial vehicle 50.

[0028] Furthermore, in order to properly detect an abnormality in the sound generating devices SG1 to SG4, it is preferable that the sounds collected from the sound generating devices SG1 to SG4 are always the same or similar. The same sound means, for example, sounds contained in voices and / or music with the same content. On the other hand, similar sounds means, for example, sounds whose voices and / or music have different contents but whose characteristics (for example, frequency patterns) contained in the sounds are the same or similar.

[0029] Therefore, the command to collect a sound included in the sound collection command may include, for example, information (for example, time) indicating the timing at which the same or similar sound is emitted from the sound generating devices SG1 to SG4. This makes it possible to control the unmanned aerial vehicle 50 to wait until a specific sound is emitted from the sound generating devices SG1 to SG4 and then start collecting sound.

[0030] Furthermore, the command to collect the sound may include information regarding the length of the collected sound. This makes it possible to obtain sound data that includes enough information to properly detect an abnormality, while not becoming excessively large in data size.

[0031] The receiving unit 112 receives collected sound data from the unmanned aerial vehicle 50 via the control terminal 30 and the wireless communication unit 39. The receiving unit 112 stores the received sound data in the memory unit 12. The sound data may be deleted from the memory unit 12 after being used to determine whether or not there is an abnormality in the sound generating devices SG1 to SG4, or may remain stored in the memory unit 12 after the determination. By storing the sound data even after the determination regarding an abnormality, the sound data can be used, for example, as teacher data for the judgment unit 113.

[0032] The judgment unit 113 analyzes the sound data received by the receiving unit 112 and makes a judgment regarding an abnormality in the sound generating devices SG1 to SG4. In this embodiment, the judgment unit 113 is configured by a neural network having an input layer to which information related to the sound data is input, and an output layer that outputs a judgment result regarding an abnormality in the sound generating devices SG1 to SG4 from which the sound data input to the input layer has been obtained. Note that an intermediate layer may be included between the input layer and the output layer of the neural network. Furthermore, the intermediate layer may further include multiple layers (so-called deep neural network).

[0033] The information on the sound data input to the input layer can be, for example, the sound data itself without any processing, data obtained by extracting specific frequency components from the sound data, or data obtained by Fourier transforming the sound data. The data obtained by Fourier transforming the sound data can be, for example, data representing the frequency included in the sound data and the magnitude of the frequency components.

[0034] The judgment result output from the output layer is, for example, whether the sound generation devices SG1 to SG4 to be detected are abnormal or normal. Specifically, the output layer outputs, for example, "1" if the sound generation devices SG1 to SG4 to be detected are abnormal, and outputs "0" if they are normal. Alternatively, conversely, "0" may be output from the output layer if the sound generation devices SG1 to SG4 to be detected are abnormal, and "1" may be output from the output layer if they are normal.

[0035] The output layer of the neural network can also output a numerical value other than "0" or "1". Using this characteristic, the output layer of the determination unit 113 may output, for example, a numerical value (score) related to an abnormality in the sound generation devices SG1 to SG4. Specifically, for example, the greater the degree of abnormality in the sound generation devices SG1 to SG4, the greater the numerical value that can be output from the output layer. In this case, for example, when the numerical value output from the output layer becomes equal to or greater than a predetermined threshold value, it can be determined that an abnormality that requires investigation exists in the sound generation devices SG1 to SG4.

[0036] The output layer can also output a number of different values. For example, the output layer can output the presence or absence of a number of abnormalities that may occur in the sound generation devices SG1 to SG4, or a score indicating the degree of a number of abnormalities.

[0037] The determination unit 113, which is configured by a neural network, is trained in advance before use. In other words, the determination unit 113 is a trained model of the neural network.

[0038] As described later, the sound data acquired by the unmanned aerial vehicle 50 includes sound components of the sound generating devices SG1 to SG4 and components of the operating sound of the unmanned aerial vehicle 50 (mainly the blade sounds of the rotor blades 55). Therefore, in this embodiment, the learning of the judgment unit 113 is performed using, for example, sound data including normal sound components of the sound generating devices SG1 to SG4 and components of the operating sound of the unmanned aerial vehicle 50 as teacher data. As a result, even if the acquired sound data is input directly to the input layer, an appropriate judgment result regarding an abnormality is output from the output layer.

[0039] The teacher data may be sound data including sound components of the sound generating devices SG1-SG4 having an abnormality and operation sound components of the unmanned aerial vehicle 50. The teacher data may be sound data actually acquired by the unmanned aerial vehicle 50, or may be theoretical sound data generated by modeling the sound components from the sound generating devices SG1-SG4 and the operation sound components of the unmanned aerial vehicle 50.

[0040] Furthermore, the teacher data may contain only sound components from the sound generating devices SG1 to SG4. In this case, sound data containing the operation sound components of the unmanned aerial vehicle 50 may be input to the input layer, or data from which the operation sound has been removed may be input.

[0041] 1-1-3. Configuration of the control terminal The control terminal 30 shown in FIG. 3 is a computer device for controlling the unmanned aerial vehicle 50 by wireless communication, and includes a control unit 31, a memory unit 32, an input unit 36, a display unit 37, and a wireless communication unit 39. The control unit 31 is an electronic circuit such as a CPU, and operates as an arithmetic processing device and a control device. The control unit 31 controls the control terminal 30 according to various programs to execute the functions described below. The memory unit 32 is a semiconductor memory such as a ROM, RAM, or flash memory. The ROM and flash memory store programs used by the control unit 31, calculation parameters, calculation results, etc. The RAM temporarily stores programs used in executing the functions of the control unit 31, and parameters that change appropriately during the execution. The input unit 36 ​​includes a joystick, buttons, or a touch panel, and controls the flight of the unmanned aerial vehicle 50 by input operations by the pilot. The input unit 36 ​​also accepts inputs for controlling the directions of the speaker, microphone, and camera mounted on the unmanned aerial vehicle 50. The display unit 37 is composed of a liquid crystal display device or an organic EL display device. The display unit 37 displays information related to the sound collection command received from the abnormality determination device 10 and the flight route. The display unit 37 also displays videos and images captured by the camera of the unmanned aerial vehicle 50.

[0042] The wireless communication unit 39 may be, for example, a wireless communication interface for connecting to a base station, or a communication interface compatible with a wireless LAN.

[0043] The control unit 31 executes the functions of the flight instruction unit 311 and the sound collection instruction unit 312 by executing a program read from the memory unit 32. The flight instruction unit 311 extracts a flight route for the unmanned aerial vehicle 50 from the sound collection command received from the abnormality judgment device 10, and transmits flight instructions according to the extracted flight route to the unmanned aerial vehicle 50. The sound collection instruction unit 312 extracts an instruction to collect sounds from each sound generating device SG1 to SG4 from the sound collection command received from the abnormality judgment device 10, and transmits a sound collection instruction to the unmanned aerial vehicle 50 according to the command to collect the sound.

[0044] If the command to collect sound received from the abnormality judgment device 10 includes information indicating the timing at which the same or similar sounds will be emitted from the sound generating devices SG1 to SG4, the sound collection instruction unit 312 sends a sound collection instruction to the unmanned aerial vehicle 50 at the timing indicated in this information.

[0045] 1-1-4. Configuration of unmanned aerial vehicles FIG. 4 shows the front appearance of the unmanned aerial vehicle 50, and FIG. 5 shows an example of the overall configuration of the unmanned aerial vehicle 50. The unmanned aerial vehicle 50 is an unmanned aircraft that can be remotely controlled or automatically piloted, and performs flight operations such as ascending, moving forward, rotating, and descending in response to flight instructions from the operation terminal 30. The unmanned aerial vehicle 50 includes a main body 50a, a plurality of arms 50b extending horizontally radially from the main body 50a, a plurality of legs 50c extending downward, and rotors 55 attached to the upper ends of each leg 50c. The unmanned aerial vehicle 50 flies using lift generated by rotating the rotors 55 due to the rotation of the motor 56. The unmanned aerial vehicle 50 also generates a reaction by changing the rotation direction of some of the rotors 55, preventing the main body 50a itself from rotating. A microphone 64 is attached to the main body 50a of the unmanned aerial vehicle 50. The microphone 64 collects sounds from the sound generating devices SG1 to SG4 that the unmanned aerial vehicle 50 approaches. The microphone 64 is capable of changing the sound collection direction by changing the angle in the vertical direction using the microphone direction control mechanism 63.

[0046] As shown in FIG. 5, the unmanned aerial vehicle 50 includes, as components for flight, a control device 51, a storage device 52, a group of sensors 53, a GPS receiver 54, a battery 58, and a wireless communication device 59. The control device 51 is an electronic circuit such as a CPU, and functions as an arithmetic processing device and a control device. The control device 51 controls the unmanned aerial vehicle 50 according to various programs to execute functions described below. The storage device 52 is a memory such as a ROM, RAM, or flash memory. The ROM and flash memory store programs, calculation parameters, calculation results, etc. used by the control device 51. The RAM temporarily stores programs used in executing the functions of the control device 51, and parameters that change as appropriate during the execution of the programs.

[0047] The sensor group 53 includes an acceleration sensor and an angular velocity sensor for detecting acceleration and angular velocity for controlling the attitude of the unmanned aerial vehicle 50, an air pressure sensor for detecting the altitude of the unmanned aerial vehicle 50, and a geomagnetic sensor for detecting the orientation of the unmanned aerial vehicle 50. The GPS receiver 54 includes an antenna and a signal processing unit, and receives signals from GPS satellites to detect position information of the unmanned aerial vehicle 50. The battery 58 stores power required for the operation of the unmanned aerial vehicle 50 and supplies it to each component. The wireless communication device 59 includes a wireless communication interface for wireless communication with the control terminal 30.

[0048] The unmanned aerial vehicle 50 further includes a camera direction control mechanism 65 and a camera 66. The camera 66 captures images of the surroundings or the direction of travel of the unmanned aerial vehicle 50. The camera 66 can change the shooting direction by changing the angle in the vertical direction using the camera direction control mechanism 65.

[0049] The control device 51 executes the functions of the flight control unit 511, the sound collection control unit 512, and the image control unit 513 by executing the program read from the storage device 52. The flight control unit 511 controls the number of rotations and the rotation speed of the motor 56 according to the flight instructions from the operation terminal 30 to execute the flight of the unmanned aerial vehicle 50. The flight control unit 511 acquires the inclination and rotation direction of the main body 50a shown in FIG. 4 based on the output data of the above-mentioned sensor group 53 and the GPS receiver 54, and acquires position information including the latitude and longitude during flight, altitude, and the azimuth angle of the main body 50a. The storage device 52 stores a program in which an algorithm is implemented to control the attitude and basic flight operations of the unmanned aerial vehicle 50 during flight. The program flies the unmanned aerial vehicle 50 while correcting the attitude and position of the main body 50a according to the flight instructions transmitted from the operation terminal 30. The unmanned aerial vehicle 50 may be manually operated by the operator using the operation terminal 30, or may be autonomously flown along a flight route transmitted from the operation terminal 30.

[0050] The sound collection control unit 512 collects the sounds of the sound generating devices SG1-SG4 as sound data using the microphone 64 in accordance with a sound collection instruction transmitted from the control terminal 30 and received via the wireless communication device 59, and transmits the collected data to the control terminal 30 via the wireless communication device 59. The sound collection control unit 512 may directly convert the data collected using the microphone 64 into sound data without compressing it, or may compress the collected data into sound data. The sound collection control unit 512 may also extract only characteristic parts used for detecting anomalies from the data collected using the microphone 64 and convert the extracted data into sound data.

[0051] The image control unit 513 temporarily stores the video and images captured by the camera 66 in the storage device 52 and transmits them to the control terminal 30 via the wireless communication device 59.

[0052] 1-2.Operation The operation of the anomaly detection system 1 shown in FIGS. 1 to 5 will be described with reference to FIG. First, the command unit 111 of the abnormality determination device 10 determines whether or not it is time to move the unmanned aerial vehicle 50 to the sound generating devices SG1-SG4 that are targets for abnormality determination (S101). Specifically, the command unit 111 determines whether or not it is time for the sound generating devices SG1-SG4 that are targets for abnormality determination to emit a predetermined sound to be collected. The predetermined sound to be collected is, for example, the same sound or a similar sound that is periodically emitted from the sound generating devices SG1-SG4.

[0053] If it is determined that it is not time to move the unmanned aerial vehicle 50 ("No" in S101), the unmanned aerial vehicle 50 will wait until the appropriate timing arrives.

[0054] On the other hand, if it is determined that it is time to move the unmanned aerial vehicle 50 ("Yes" in S101), the command unit 111 outputs the above-mentioned sound collection command to the control terminal 30. The control terminal 30, which has received the sound collection command, moves the unmanned aerial vehicle 50 to the sound generating device SG1-SG4 that is to be subjected to an abnormality judgment, according to the flight route included in the sound collection command (S102).

[0055] After the unmanned aerial vehicle 50 arrives at the sound collection position of the sound generating devices SG1 to SG4 that are to be judged for abnormality, the control terminal 30, in accordance with the sound collection command, causes the unmanned aerial vehicle 50 to hover at the sound collection position of the sound generating devices SG1 to SG4, and uses the microphone 64 of the unmanned aerial vehicle 50 to collect the sounds of the sound generating devices SG1 to SG4 (S103). In this way, the unmanned aerial vehicle 50 is hovering when collecting the sounds of the sound generating devices SG1 to SG4. In other words, the unmanned aerial vehicle 50 is also operating while collecting the sounds of the sound generating devices SG1 to SG4. Therefore, the sound data collected using the microphone 64 includes components of the sounds of the sound generating devices SG1 to SG4 and components of the operating sounds of the unmanned aerial vehicle 50 (mainly the blade sounds of the rotor 55).

[0056] After collecting the sounds from the sound generating devices SG1-SG4 that are to be subjected to anomaly judgment, the unmanned aerial vehicle 50 transmits the collected sounds as sound data to the anomaly judgment device 10 via the control terminal 30 (S104). After receiving the sound data transmitted from the unmanned aerial vehicle 50, the receiver 112 of the anomaly judgment device 10 stores the sound data in the memory unit 12.

[0057] Next, the control terminal 30 determines whether or not it has collected sounds from all of the sound generating devices SG1-SG4 in the loudspeaker equipment 100 (or all of the sound generating devices SG1-SG4 targeted for abnormality judgment) (S105). If it has not collected sounds from all of the sound generating devices SG1-SG4 ("No" in S105), the control terminal 30 moves the unmanned aerial vehicle 50 in accordance with the sound collection command from the current sound generating device SG1-SG4 to the next target sound generating device SG1-SG4 from which to collect sounds, and collects sounds from the sound generating devices SG1-SG4 (S102-S104).

[0058] After collecting the sounds from all the sound generating devices SG1 to SG4 ("Yes" in S105), the judgment unit 113 of the abnormality judgment device 10 sequentially inputs the sound data stored in the memory unit 12 and judges whether or not there is an abnormality in each of the sound generating devices SG1 to SG4. The judgment unit 113 stores the judgment results regarding the abnormality in each of the sound generating devices SG1 to SG4 in the memory unit 12.

[0059] If the judgment result of any of the sound generation devices SG1 to SG4 indicates the presence of an abnormality, the presence of the abnormality may be notified to a terminal (e.g., a personal computer, a smartphone, or a tablet terminal) connected to the abnormality judgment device 10. This notification may be sent to the terminal by email, may be displayed as a pop-up on the terminal, or may be notified by emitting a predetermined alarm sound on the terminal, for example.

[0060] Features The anomaly detection system 1 and anomaly detection method for detecting anomalies in the sound generating devices SG1 to SG4 using the unmanned aerial vehicle 50 according to the above embodiment move the unmanned aerial vehicle 50 to the sound collection positions of the sound generating devices SG1 to SG4 and have the unmanned aerial vehicle 50 collect the sounds of the sound generating devices SG1 to SG4. In addition, a determination is made regarding an abnormality in the sound generating devices SG1 to SG4 by analyzing the sound data acquired by the unmanned aerial vehicle 50. In this way, by having the unmanned aerial vehicle 50 collect the sounds of the sound generating devices SG1 to SG4, the sound data of the sound generating devices SG1 to SG4 can be acquired efficiently.

[0061] In the anomaly detection system 1 according to the above embodiment, the determination of an anomaly in the sound generating devices SG1 to SG4 based on the sound data is performed by a neural network (a trained model of the neural network). This makes it possible to appropriately determine an anomaly in the sound generating devices SG1 to SG4 without designing a logical model, algorithm, or the like that analyzes the sound data and calculates a determination result regarding an anomaly.

[0062] 2. Other embodiments As described above, each embodiment has been described as an example of the technology disclosed in this application. However, the technology in this disclosure is not limited to this, and can be applied to embodiments in which modifications, substitutions, additions, omissions, etc. are appropriately performed. In addition, it is also possible to combine the components described in the above embodiments to create a new embodiment. For example, the following embodiments are possible.

[0063] (1) In the above embodiment, the determination as to whether or not there is an abnormality is performed after acquiring sound data from all of the sound generating devices SG1 to SG4. However, this is not limited to the above, and the determination as to whether or not there is an abnormality may be performed each time sound data from one of the sound generating devices SG1 to SG4 is acquired.

[0064] (2) If there is an appropriate model and / or algorithm for making a judgment regarding an abnormality from sound data, the judgment unit 113 may be realized as software or hardware that realizes this model and / or algorithm.

[0065] For example, the judgment unit 113 can be realized as an algorithm that compares reference value data related to the sound of normal sound generating devices SG1 to SG4 with sound data acquired by the unmanned aerial vehicle 50, and judges that the sound generating devices SG1 to SG4 are abnormal if the sound data is outside the range of the reference value data. The reference value data is data that represents the characteristics of the sound emitted from the normal sound generating devices SG1 to SG4. The reference value data includes, for example, frequency components contained in the sound emitted from the normal sound generating devices SG1 to SG4, setting information related to the sound generating devices SG1 to SG4 (for example, the volume of the sound emitted, equalizer settings, etc.), and the like.

[0066] For example, if the sound components other than those of the sound generating devices SG1-SG4 contained in sound data such as the operating sound of the unmanned aerial vehicle 50 are constant, the sound components other than those of the sound generating devices SG1-SG4 can be removed by frequency filtering, etc. In this case, the judgment unit 113 can appropriately detect abnormalities in the sound generating devices SG1-SG4 based on the sound data (from which sound components other than those of the sound generating devices SG1-SG4 have been removed) using the above algorithm.

[0067] (3) For example, when it is necessary for the sound generating devices SG1 to SG4 to amplify important content such as an emergency broadcast while collecting sound, the unmanned aerial vehicle 50 may be moved to a position away from the sound generating devices SG1 to SG4 (for example, far above the sound generating devices SG1 to SG4). This prevents the important content from being drowned out by the operating sounds of the unmanned aerial vehicle 50 and becoming inaudible to people.

[0068] (4) Determination unit 113 configured with a neural network may be configured by implementing the algorithm of the neural network in software, or by implementing the neural network in hardware.

[0069] (5) The unmanned aerial vehicle 50 that collects sound is not limited to one, and the anomaly detection system 1 may have multiple units.

[0070] (6) In the above embodiment, the unmanned aerial vehicle 50 collects the sound of one sound generating device SG1 to SG4. However, this is not limited to the above, and one unmanned aerial vehicle 50 may hover at the midpoint between the multiple sound generating devices SG1 to SG4 and collect the sounds from the multiple sound generating devices SG1 to SG4 at once. In other words, the sound collection position may be the midpoint between the multiple sound generating devices SG1 to SG4. In this case, the unmanned aerial vehicle 50 may have a sound collection device equipped with multiple microphones, such as an array microphone, as the microphone 64 that collects the sound.

[0071] This allows the unmanned aerial vehicle 50 to hover at the midpoint between the two sound generating devices SG1 to SG4 and simultaneously collect the sounds from the two sound generating devices SG1 to SG4, making it possible to efficiently collect sounds. Also, for example, it is possible to simultaneously determine abnormalities, such as any of the multiple sound generating devices SG1 to SG4 not emitting sound.

[0072] (7) The anomaly detection system 1 according to the above embodiment includes a control terminal 30, but is not limited to this. The anomaly detection system 1 may have a configuration in which a sound collection command is sent directly from the anomaly judgment device 10 to the unmanned aerial vehicle 50. Specifically, the anomaly judgment device 10 may be disposed as part of a control system for the unmanned aerial vehicle 50, and the unmanned aerial vehicle 50 may perform automatic flight and sound collection by remote communication via the communication unit 19.

[0073] (8) The unmanned aerial vehicle 50 may have a speaker for amplifying sound from the unmanned aerial vehicle 50 and a speaker direction adjustment mechanism for adjusting the direction of the speaker.

[0074] (9) In the above-described embodiments, each device or system may have a cloud computing configuration in which a single function is shared and processed jointly by multiple devices via a network.

[0075] (10) In cases where the sound-generating device is located at a relatively low position, instead of the unmanned aerial vehicle 50, an autonomous mobile robot that is equipped with a microphone for collecting sound and moves autonomously on the ground may be used to collect the sound from the sound-generating device.

[0076] (11) In the above embodiments, the term "device" or "system" includes cases where it means a collection of multiple components (devices, modules (parts), etc.), regardless of whether all the components are in the same housing. In addition, multiple devices housed in separate housings and connected via a network, and a single device in which multiple modules are housed in a single housing, may both be called a system.

[0077] Each step described in the above flow chart can be executed by one device or can be shared among multiple devices. Furthermore, when one step includes multiple processes, the multiple processes included in the step can be executed by one device or can be shared among multiple devices.

[0078] In the above embodiment, the control unit or control device of each device or equipment may include a processor configured with a dedicated electronic circuit designed to realize a predetermined function according to the function required for each device or equipment. The control unit or control device may be realized by various processors or electronic circuits such as a CPU, an MPU, a GPU, a DSP, an FPGA, an ASIC, etc. The control unit or control device may be configured with one or more processors.

[0079] Part or all of the memory unit or storage device of each device or equipment may be composed of any computer-readable recording medium, such as an optical disk, magnetic disk, magneto-optical disk, magnetic tape, HDD, SD card, SSD, etc., depending on the functions required for each device or equipment.

[0080] The above computer program is not limited to one recorded on a recording medium, but may be one acquired via a telecommunications line, a wireless or wired communication line, a network such as the Internet, or the like.

[0081] The communication unit or communication device of each device or equipment may be any communication interface, such as a wireless LAN, a wired LAN, 3G (3rd Generation), LTE (Long Term Evolution), 4G (4th Generation), 5G (5th Generation), or a millimeter wave wireless communication interface, depending on the functions required for each device or equipment.

[0082] Each process in the above-described embodiment may be realized by hardware or software (including the case where it is realized together with an OS (operating system), middleware, or a predetermined library). Furthermore, each process may be realized by a combination of software and hardware.

[0083] The order of operations of the system, device or equipment in the above embodiments is not necessarily limited to that described in the above embodiments, and the order of operations can be changed or multiple operations can be executed simultaneously without departing from the spirit of the invention. [Industrial Applicability]

[0084] The present disclosure is broadly applicable to sound generating device route guidance systems, route selection devices, and route guidance methods. [Explanation of symbols]

[0085] 1: Anomaly detection system 10: Abnormality judgment device 11: Control section 111: Command department 112: Receiving unit 113:Judgment Department 12: Storage section 19: Communications Department 30: Control terminal 31: Control section 311:Flight instruction department 312: Sound collection instruction section 32: Storage section 36: Input section 37:Display section 39: Wireless communication section 50: Unmanned aerial vehicle 50a: Main unit 50b: Arm 50c: Legs 51: Control device 511: Flight control unit 512: Sound collection control unit 513: Image control unit 52: Storage device 53: Sensor group 54: GPS receiver 55: Rotor 56: Motor 58: Battery 59: Wireless communication equipment 63: Microphone directional control mechanism 64: Mike 65: Camera direction control mechanism 66: Camera N: Network 100: Public address system A: Installation area SG1~SG4: Sound generating device SS:Sound source

Claims

1. An abnormality detection system that detects an abnormality in a speaker that converts a sound signal from a sound source into sound and broadcasts the sound, a mobile object that moves to a sound collection position of a speaker to be determined whether an abnormality occurs and acquires sound data including components of the sound of the speaker when the same or similar sound is broadcast from the speaker; an abnormality determination device that determines whether or not the speaker is abnormal by analyzing sound data acquired by the moving object; An anomaly detection system comprising:

2. The abnormality determination device includes: an input layer to which information about the sound data is input; an output layer that outputs a judgment result regarding an abnormality of the speaker from which the sound data input to the input layer is acquired; The anomaly detection system according to claim 1 , further comprising a determination unit configured with a neural network having:

3. The anomaly detection system according to claim 2 , wherein the output layer outputs, as the determination result, whether the speaker is abnormal or normal.

4. The anomaly detection system according to claim 2 , wherein the output layer outputs a numerical value related to the abnormality of the speaker as the judgment result.

5. The abnormality detection system according to claim 1 , wherein the abnormality determination device compares the sound data with reference value data relating to normal sound of the speaker, and determines that the speaker is abnormal if the sound data is outside a range of the reference value data.

6. The anomaly detection system according to any one of claims 1 to 5, wherein the sound data includes a component of the sound from the speaker and a component of an operating sound of the moving object.

7. The input layer of the neural network receives removal data obtained by removing an operation sound component of the moving object from the sound data, The anomaly detection system according to claim 2 , wherein the output layer of the neural network outputs a determination result regarding an anomaly of the speaker from which the removal data input to the input layer was obtained.

8. An abnormality detection method for detecting an abnormality in a speaker that converts a sound signal from a sound source into sound and broadcasts the sound, comprising: a step of moving a moving object to a sound collection position of the speaker at a timing when the same or similar sound is broadcast from the speaker that is to be determined as being abnormal; A step of causing the moving object that has moved to a sound collection position of the speaker to acquire sound data including a component of the sound of the speaker; A step of determining whether an abnormality exists in the speaker by analyzing sound data acquired by the moving object; The anomaly detection method includes:

Images