Abnormality detection device, abnormality detection system, and abnormality detection method

Abstract

The purpose of the present invention is to provide a technology with which it is possible for an actor to specifically ascertain the content of an abnormality and / or a countermeasure therefor. This abnormality detection device comprises: a type information acquisition unit that acquires, on the basis of abnormality information that indicates an abnormality of an autonomous system and is expressed in a language other than a natural language, type information including at least one of an abnormality type of the autonomous system and a countermeasure type corresponding to the abnormality type, each of the abnormality type and the countermeasure type being expressed in the natural language; and a notification control unit that performs control to notify an actor of the type information from a notification unit.

Description

Abnormal Detection Device, Abnormal Detection System, and Abnormal Detection Method

[0001] The present disclosure relates to an abnormal detection device, an abnormal detection system, and an abnormal detection method.

[0002] Various technologies have been proposed for autonomous systems capable of autonomous operation. For example, Patent Document 1 proposes a technology for notifying a system monitor of an abnormality when an abnormality occurs in an autonomous system.

[0003] International Publication No. 2023 / 026750

[0004] However, in the technology of Patent Document 1, since an abnormality is notified by an alert sound and light emission that are not natural language, there is a problem that an actor such as a monitor cannot specifically grasp the content of the abnormality.

[0005] Therefore, the present disclosure has been made in view of the above problems, and an object thereof is to provide a technology that enables an actor to specifically grasp at least one of the content of an abnormality and a countermeasure method.

[0006] The abnormal detection device according to the present disclosure indicates an abnormality of an autonomous system, and based on abnormality information represented in a language other than natural language, a type information acquisition unit that acquires type information including at least one of an abnormality type of the autonomous system and a countermeasure type for the abnormality type, each represented in the natural language, and a notification control unit that controls to notify an actor, who is a person related to the autonomous system, of the type information from a notification unit.

[0007] According to the present disclosure, control is performed to notify an actor from a notification unit of type information including at least one of an abnormality type and a countermeasure type, each represented in natural language. With such a configuration, an actor can specifically grasp at least one of the content of an abnormality and a countermeasure method. The object, features, aspects, and advantages of the present disclosure will become clearer from the following detailed description and the accompanying drawings.

[0008] Figure 1 is a block diagram showing the configuration of the anomaly detection system according to Embodiment 1. Figure 2 is a diagram showing actor information according to Embodiment 1. Figure 3 is a diagram showing a table used by the type information acquisition unit according to Embodiment 1. Figure 4 is a block diagram showing the configuration of the anomaly detection system according to Embodiment 1. Figure 5 is a diagram showing a table used by the identification unit according to Embodiment 1. Figure 6 is a diagram showing an example of type information notification according to Embodiment 1. Figure 7 is a flowchart showing the main operations of the anomaly detection device according to Embodiment 1. Figure 8 is a block diagram showing the configuration of the anomaly detection system according to Modification 1. Figure 9 is a block diagram showing the configuration of the anomaly detection system according to Modification 2. Figure 10 is a block diagram showing the configuration of the anomaly detection system according to Modification 2. Figure 11 is a block diagram showing the configuration of the anomaly detection system according to Modification 3. Figure 12 is a block diagram showing the hardware configuration of the anomaly detection device according to another modification. Figure 13 is a block diagram showing the hardware configuration of the anomaly detection device according to another modification.

[0009] <Embodiment 1> Figure 1 is a block diagram showing the configuration of an anomaly detection system according to this embodiment 1. The anomaly detection system comprises an input device 1, an anomaly detection device 2, and a notification unit 3. The anomaly detection device 2 is connected to the input device 1 and the notification unit 3 in a communicative manner, and is also connected to the autonomous system 7 and the peripheral system 8 in a communicative manner. The communication referred to here may be wired communication or wireless communication.

[0010] Autonomous system 7 generates state information for autonomous system 7. Autonomous system 7 includes, for example, robots and machine tools, and all machines present around the actor described later. Robots include, for example, autonomous robots capable of autonomous operation. If autonomous system 7 includes an autonomous robot capable of autonomous operation, the state information for autonomous system 7 includes, for example, the behavior of the autonomous robot indicating whether it is standing still or moving, and the position of the autonomous robot. If autonomous system 7 includes a machine tool, the state information for autonomous system 7 includes, for example, the behavior of the machine tool. The following description will focus on the case where autonomous system 7 includes an autonomous robot, but the same applies to the case where autonomous system 7 includes a machine tool.

[0011] The peripheral system 8 is a system surrounding the autonomous system 7. The peripheral system 8 includes, for example, at least one of the following: a system that cooperates with the autonomous system 7, a system that can objectively acquire the state of the autonomous system 7, and a system that can acquire information about people and objects around the autonomous robot. In this specification, for example, at least one of A, B, C, ..., and Z means any one of all combinations that can be obtained by selecting one or more types from the multiple types of A, B, C, ..., and Z.

[0012] Systems that cooperate with the autonomous system 7 include, for example, elevator systems and automatic door systems used by the autonomous robot. Systems that can objectively acquire the status of the autonomous system 7 include, for example, surveillance cameras. Systems that can acquire information about people and objects around the autonomous robot include, for example, terminals carried by people around the autonomous robot (e.g., smartphones, smartwatches) and access control systems.

[0013] The peripheral system 8 generates peripheral environment information, which is information about the environment surrounding the autonomous system 7. This peripheral environment information includes, for example, the open / closed status of the elevator system, the results of video analysis from surveillance cameras, and the positions of people around the autonomous robot.

[0014] Input device 1 accepts input of actor information, which includes at least one of actor state information and attribute information. An actor is a person related to the autonomous system 7, and includes, for example, a pedestrian in the vicinity of the autonomous robot, a user of the autonomous system 7, and a maintenance person for the autonomous system 7. For example, if the autonomous system 7 includes an autonomous robot for cleaning, the user of the autonomous system 7 is the cleaning staff, and the maintenance person for the autonomous system 7 is a person in charge of maintenance among the cleaning staff, or a system support person from the manufacturer, etc.

[0015] The actor's state information includes, for example, the actor's behavior, such as whether the actor is standing still or moving, and the actor's location. The actor's attribute information indicates, for example, whether the actor is a nearby pedestrian, a user of the autonomous system 7, or a maintenance worker for the autonomous system 7.

[0016] Actor information may include not only actor status information and attribute information, but also, as shown in Figure 2, notification content notified to the actor by the anomaly detection system, target notification unit, notification restrictions, whether or not an urgency notification is issued, and whether or not a work instruction notification is issued. In Figure 2, the notification content and other information are set for the actor's attribute information, but they may also be set for the actor's status information.

[0017] For example, if an actor located very close to the autonomous system 7 is busy, and the actor's status information indicates that the actor has moved more than a certain amount, a simple notification may be issued. In this case, only the type of action to be taken may be set as the notification content, the speaker of the autonomous robot may be set as the target notification unit, and another actor located near the autonomous system may be set as the notification restriction. For example, if an actor's attribute information indicates that they are a customer service staff member, and the actor's status information indicates that the actor has stopped moving, the contents of Figure 2 may be set to notify an actor located near the autonomous system 7 and that is moving. Actor information may be input to the input device 1 by the actor's operation, or it may be automatically input to the input device 1 by a terminal held by the actor.

[0018] The notification unit 3 in Figure 1 is a notification device capable of notifying the actor of various types of information expressed in natural language, and includes, for example, at least one of a speaker and a display device. The natural language is a language that is generally understandable to the actor, such as the actor's native language. The speaker and display device may be an accessory device of the anomaly detection device 2 or the autonomous robot, or they may be a tablet located away from the anomaly detection device 2 or the autonomous robot.

[0019] The anomaly detection device 2 controls the notification unit 3 based on the status information of the autonomous system 7, the surrounding environment information of the peripheral system 8, and the actor information received by the input device 1. The configuration of the anomaly detection device 2 will be mainly described below.

[0020] The anomaly detection device 2 comprises an anomaly detection unit 2a which is an anomaly acquisition unit, an actor information acquisition unit 2b, a type information acquisition unit 2c, an urgency calculation unit 2d, a identification unit 2e, a notification control unit 2f, and a learning information provision unit 2g.

[0021] The anomaly detection unit 2a detects an anomaly in the autonomous system 7 based on at least one of the state information of the autonomous system 7 and the surrounding environment information of the peripheral system 8. The anomaly detection unit 2a may also detect an anomaly using a table that pre-associates the state information of the autonomous system 7, at least one of the surrounding environment information of the peripheral system 8, and the anomaly. Alternatively, the anomaly detection unit 2a may detect an anomaly using the results of machine learning (training) on ​​the state information of the autonomous system 7, at least one of the surrounding environment information of the peripheral system 8, and the anomaly.

[0022] For example, if the status information of the autonomous system 7 indicates that the autonomous robot has stopped moving, the abnormality detection unit 2a may detect that the autonomous system 7 is abnormal. For example, even if the status information of the autonomous system 7 indicates that the autonomous robot is operating normally inside the elevator, if the surrounding environment information of the surrounding system 8 indicates that the elevator door has been open for a certain period of time or longer, or that the elevator has been occupied for a certain period of time or longer, or that the elevator has been called more frequently than a certain number of times, the abnormality detection unit 2a may detect that the autonomous system 7 is abnormal. For example, if the surrounding environment information of the surrounding system 8 indicates that people and objects around the autonomous robot have stopped moving, the abnormality detection unit 2a may detect that the autonomous system 7 is abnormal.

[0023] The anomaly detection unit 2a indicates the detected anomaly and acquires (generates) anomaly information expressed in a language other than natural language. This non-natural language is, for example, an error code, a language that is incomprehensible to the average person.

[0024] For example, the abnormal information may be an error code indicating an autonomous system error, such as a communication error (network error) or a stack error. For example, the abnormal information may be an error code indicating an elevator coordination error, such as an error in the elevator door remaining open for a certain period of time or an error in the elevator's occupancy time for a certain period of time. For example, the abnormal information may be an error code indicating a behavioral error of the autonomous system 7, such as an error in the autonomous robot, people and objects around the autonomous robot stopping moving for a certain period of time or an error in the autonomous robot stopping operating for a certain period of time. Furthermore, the abnormal information acquired by the abnormality detection unit 2a may include the date and time of abnormality detection, the position of the autonomous robot at the time of abnormality detection, and a code indicating the identification information and state of the autonomous system 7 in which the abnormality was detected.

[0025] The actor information acquisition unit 2b acquires actor information received by the input device 1 and outputs the actor information to the type information acquisition unit 2c.

[0026] The type information acquisition unit 2c acquires (generates) type information based on the abnormality information from the abnormality detection unit 2a and the actor information from the actor information acquisition unit 2b. The type information includes at least one of the abnormality type of the autonomous system 7 and the countermeasure type for that abnormality type, each expressed in natural language. The abnormality type is the specific type of abnormality indicated by the abnormality information and is expressed in natural language, for example, "The robot collided with a wall and stopped." The countermeasure type is the specific countermeasure method for the abnormality type and is expressed in natural language, for example, "The passage is too narrow, so temporarily move the robot to a safe location."

[0027] The type information acquisition unit 2c may acquire the type of action using a table that pre-associates the error code of the abnormal information with the attribute information of the actor and the type of action, as shown in Figure 3. In Figure 3, "101" is the error code for a communication error, "102" is the error code for an obstacle detection error, and "103" is the error code for a fire possibility error.

[0028] The type information acquisition unit 2c may acquire the abnormality type using a table that pre-associates a combination of the error code of the abnormality information and the attribute information of the actor with the abnormality type, similar to the acquisition of the response type described above. Extending this, the type information acquisition unit 2c may acquire the type information using a table that pre-associates a combination of the abnormality information and actor information with type information that includes at least one of the abnormality type and the response type. Furthermore, the abnormality detection unit 2a may acquire the type information using the results of machine learning (training) on ​​the abnormality information, actor information, and type information.

[0029] The type information acquisition unit 2c outputs the type information to the notification control unit 2f. However, as shown in Figure 2, if the actor information includes notification restrictions and those restrictions are met, the type information acquisition unit 2c does not need to output the type information to the notification control unit 2f, and the actor information acquisition unit 2b does not need to output the actor information to the type information acquisition unit 2c.

[0030] In the above explanation, the type information acquisition unit 2c acquired type information based on the abnormality information from the abnormality detection unit 2a and the actor information from the actor information acquisition unit 2b, but this is not the only way. For example, if there is no need to change the type information based on the actor information, the type information acquisition unit 2c may acquire type information based on the abnormality information from the abnormality detection unit 2a, and the input device 1 and the actor information acquisition unit 2b may not be provided as shown in Figure 4.

[0031] The urgency calculation unit 2d calculates the urgency of the anomaly detected by the anomaly detection unit 2a based on at least one of the following: the state information of the autonomous system 7, the surrounding environment information of the peripheral system 8, and the anomaly information of the anomaly detection unit 2a. In this embodiment 1, the urgency increases as the likelihood of the autonomous system 7 obstructing the movement of people and objects, the likelihood of the autonomous system 7 causing harm to people and objects, or the likelihood of danger occurring to the autonomous system 7 increases.

[0032] For example, the urgency calculation unit 2d may determine the movement status of people around the autonomous robot based on the analysis results of video footage from the surveillance cameras of the surrounding system 8, the position of the autonomous robot, and the position of terminals held by people. The urgency calculation unit 2d may then calculate a high level of urgency if the movement status indicates that people cannot move, a medium level of urgency if the situation is crowded but people can move, and a low level of urgency if the situation is not crowded and people can move.

[0033] For example, the urgency level calculation unit 2d may determine the attributes of people around the autonomous robot based on the analysis results of video footage from the surveillance cameras of the surrounding system 8, the position of the autonomous robot, and the position and settings (e.g., personal information) of terminals held by people. The urgency level calculation unit 2d may then calculate a high level of urgency if the attributes indicate people who have difficulty dealing with the abnormality on their own, such as elderly people with mobility difficulties, children, or people carrying heavy luggage, and calculate a medium or low level of urgency if the attributes indicate people who can easily deal with the abnormality on their own.

[0034] For example, the urgency level calculation unit 2d may determine the state of people around the autonomous robot based on the analysis results of video footage from the surveillance cameras of the surrounding system 8, the position of the autonomous robot, and the positions and sensors of terminals held by people around the autonomous robot. The urgency level calculation unit 2d may then calculate a high level of urgency if the person is lying down, a medium level of urgency if the person is standing still, and a low level of urgency if the person is moving.

[0035] For example, the urgency calculation unit 2d may determine, based on the analysis results of the video footage from the surveillance camera of the surrounding system 8 and the position of the autonomous robot, whether the autonomous robot is located in a place it should not occupy, a place where it may fall, or a place that obstructs human movement. The urgency calculation unit 2d may then calculate a high level of urgency if the autonomous robot is located in any of these places, and a medium or low level of urgency otherwise. Places that an autonomous robot should not occupy are, for example, places where elevators or automatic doors are located, places where an autonomous robot may fall are, for example, places near stairs or escalators, and places where an autonomous robot obstructs human movement are, for example, narrow passages.

[0036] For example, the urgency calculation unit 2d may calculate the urgency based on the abnormality information from the abnormality detection unit 2a, or the output of abnormality information by the abnormality detection unit 2a may be used as a trigger for the urgency calculation unit 2d to calculate the urgency. The urgency calculation unit 2d may also calculate the urgency using the results of machine learning (training) on ​​at least one of the state information of the autonomous system 7, the surrounding environment information of the peripheral system 8, and the abnormality information from the abnormality detection unit 2a, as well as the urgency. Furthermore, the urgency may be accompanied by reasons for calculating the urgency, such as the autonomous robot obstructing passage or the autonomous robot being located in a narrow passage. The urgency calculation unit 2d may also assign a higher urgency the higher the priority of the task of the autonomous system 7.

[0037] When an abnormality occurs, the identification unit 2e acquires the first work status of the autonomous system 7 and, based on the first work status, identifies a work instruction for the actor to take over the work of the autonomous system 7. For example, as shown in Figure 5, the identification unit 2e may use a table that pre-associates the processes included in the autonomous robot's action plan with work instructions to identify a work instruction that corresponds to the same or next process as the first work status as the work instruction for the actor to take over the work.

[0038] In this embodiment 1, the identification unit 2e uses the output of abnormal information by the abnormality detection unit 2a as a trigger for acquiring the first work status in the identification unit 2e, but it is not limited to this. For example, the identification unit 2e may also use the output of type information by the type information acquisition unit 2c as a trigger for acquiring the first work status in the identification unit 2e.

[0039] The notification control unit 2f controls the notification unit 3 to notify the actor of the type information from the type information acquisition unit 2c, the urgency from the urgency calculation unit 2d, and the work instructions from the identification unit 2e. If the presence or absence of notification of urgency and the presence or absence of notification of work instructions in Figure 2 are input to the notification control unit 2f via the type information acquisition unit 2c, the notification control unit 2f may control the notification and non-notification of urgency and the notification and non-notification of work instructions based on their presence or absence. Furthermore, if the target notification unit in Figure 2 is input to the notification control unit 2f via the type information acquisition unit 2c, the notification control unit 2f may change the notification unit 3 that notifies the actor based on the target notification unit.

[0040] Figures 6(a) and 6(b) show examples of type information notifications by the notification unit 3. In Figure 6(a), the type information, which includes the abnormality type but does not include the action type, is changed based on the actor's attribute information, and in Figure 6(b), the type information, which includes both the abnormality type and the action type, is changed based on the actor's attribute information.

[0041] Although not shown in the drawings, the type information including the countermeasure type but not including the abnormality type may be changed based on the attribute information of the actor, or the type information may be changed based on the state information of the actor. The type information notified to the user is more complex than the type information notified to the surrounding pedestrians, and the type information notified to the maintenance staff is more complex than the type information notified to the user.

[0042] The learning information providing unit 2g transmits and provides the abnormality information of the abnormality detection unit 2a to the learning unit 9 as learning information for the learning unit 9 to learn the actions of the autonomous system 7 that cause the abnormality. However, the abnormality information referred to here is the abnormality information obtained based on the state information of the autonomous system 7 and the surrounding environment information. The learning information providing unit 2g may transmit the abnormality information to the learning unit 9 at any time or periodically.

[0043] The learning unit 9 performs machine learning (training) based on at least one of the action information (action command) of the autonomous system 7 and the state information of the autonomous system 7 and the abnormality information. In FIG. 1, the learning unit 9 is provided in the autonomous system 7, and the learning information providing unit 2g provides the abnormality information to the autonomous system 7, but this is not the only case. For example, the learning unit 9 having a communication function with the autonomous system 7 may be provided outside the autonomous system 7, and the learning information providing unit 2g may provide the abnormality information outside the autonomous system 7.

[0044] <Operation> FIG. 7 is a flowchart showing the main operations of the abnormality detection device 2 according to the first embodiment. The operations in FIG. 7 may be performed at any time or periodically.

[0045] In step S1, the abnormality detection unit 2a acquires the state information of the autonomous system 7 and the surrounding environment information of the surrounding system 8.

[0046] In step S2, the abnormality detection unit 2a determines whether there is an abnormality in the autonomous system 7 based on the state information of the autonomous system 7 and the peripheral environment information of the peripheral system 8. Here, the abnormality detection unit 2a acquires and uses the state information of the autonomous system 7 and the peripheral environment information of the peripheral system 8, but at least either one of the state information of the autonomous system 7 and the peripheral environment information of the peripheral system 8 may be acquired and used. When it is determined that there is an abnormality in the autonomous system 7, the process proceeds to step S3, and when it is determined that there is no abnormality in the autonomous system 7, the process returns to step S1.

[0047] In step S3, the type information acquisition unit 2c acquires type information based on the abnormality information of the abnormality detection unit 2a and the actor information of the actor information acquisition unit 2b.

[0048] In step S4, the notification control unit 2f controls to notify the actor of the type information from the notification unit 3. Then, the operation of FIG. 7 ends.

[0049] <Summary of Embodiment 1> According to the abnormality detection device 2 according to the first embodiment as described above, the type information acquisition unit 2c acquires type information including at least either one of an abnormality type and a countermeasure type, each of which is represented in a natural language, based on the abnormality information represented in a language other than the natural language, and the notification control unit 2f controls to notify the actor of the type information from the notification unit 3. According to such a configuration, the actor can specifically grasp at least either one of the content of the abnormality and the countermeasure method.

[0050] Also in the first embodiment, the type information acquisition unit 2c acquires type information based on the abnormality information and the actor information. According to such a configuration, it is possible to notify the actor of type information suitable for at least either one of the state and attributes of the actor.

[0051] Also in the first embodiment, the urgency calculation unit 2d calculates the urgency of the abnormality of the autonomous system 7, and the notification control unit 2f controls to notify the actor of the urgency from the notification unit 3. According to such a configuration, the actor can grasp the urgency of the abnormality of the autonomous system 7.

[0052] In this embodiment 1, the identification unit 2e identifies the work instruction, and the notification control unit 2f controls the notification unit 3 to notify the actor of the work instruction. With this configuration, the actor can easily take over the work of the autonomous system 7.

[0053] In this embodiment 1, the learning information provision unit 2g provides the learning unit 9 with abnormal information acquired based on the state information of the autonomous system 7 and the surrounding environment information, as learning information for the learning unit 9 to learn the actions of the autonomous system 7 that cause abnormalities. With this configuration, abnormalities in the autonomous system 7 can be reduced.

[0054] <Modification 1> In Embodiment 1, the type information acquisition unit 2c acquired type information based on the abnormality information from the abnormality detection unit 2a and the actor information from the actor information acquisition unit 2b, but it is not limited to this.

[0055] For example, the type information acquisition unit 2c may acquire multiple type information corresponding to a predetermined set of actor information based on the abnormality information from the abnormality detection unit 2a and a predetermined set of actor information. The notification control unit 2f may then acquire one actor information from the actor information acquisition unit 2b, as shown in Figure 8, and perform control to notify the abnormality information corresponding to the one actor information from the predetermined set of actor information corresponding to a set of abnormality information. In other words, the notification control unit 2f may perform control to notify the actor of type information from the notification unit 3 based on the actor information. With such a configuration, type information appropriate to the actor information can be notified to the actor, similar to Embodiment 1.

[0056] <Modification 2> In Embodiment 1, an example was described in which the abnormality acquisition unit is an abnormality detection unit 2a that detects abnormalities in the autonomous system 7 based on at least one of the state information of the autonomous system 7 and the surrounding environment information of the peripheral system 8, but it is not limited to this. For example, as shown in Figure 9, the abnormality acquisition unit may include an abnormality prediction unit 2h and an abnormality database 2i.

[0057] The anomaly database 2i stores past information, which includes at least one of the state information of the autonomous system 7 and the surrounding environment information of the peripheral system 8, at the time of an anomaly that occurred in the autonomous system 7 in the past. The anomaly prediction unit 2h predicts the behavior of the autonomous system 7 based on at least one of the state information of the autonomous system 7 and the surrounding environment information, as well as the past information stored in the anomaly database 2i, and predicts an anomaly in the autonomous system 7 based on the prediction result.

[0058] For example, the anomaly prediction unit 2h may predict an anomaly in the autonomous system 7 based on the degree of wear of the equipment indicated by the status information of the autonomous system 7 and the degree of wear indicated by past information stored in the anomaly database 2i. Alternatively, for example, the anomaly prediction unit 2h may calculate distance and density based on the position of the autonomous robot indicated by the status information of the autonomous system 7 and the positions of people and objects around the autonomous robot indicated by the surrounding environment information of the surrounding system 8. Then, the anomaly prediction unit 2h may predict an anomaly in the autonomous system 7 based on the degree of agreement between the calculated distance and density and the distance and density indicated by past information stored in the anomaly database 2i.

[0059] The anomaly prediction unit 2h indicates the predicted anomaly and acquires (generates) anomaly information expressed in a language other than natural language. The anomaly information acquired by the anomaly prediction unit 2h may be the same as the anomaly information acquired by the anomaly detection unit 2a. Furthermore, the anomaly information acquired by the anomaly prediction unit 2h may include the date and time of the anomaly prediction, the position of the autonomous robot at the time of the anomaly prediction, and a code indicating the identification information and state of the autonomous system 7 in which the anomaly was predicted.

[0060] Furthermore, in the above explanation, the anomaly detection unit 2a in Figure 1 and the anomaly prediction unit 2h in Figure 9 acquire (generate) anomaly information based on at least one of the state information of the autonomous system 7 and the surrounding environment information of the peripheral system 8, but this is not the only way. For example, as shown in Figure 10, neither the anomaly detection unit 2a nor the anomaly prediction unit 2h may be provided in the anomaly detection device 2, and the autonomous system 7 may generate anomaly information, and the type information acquisition unit 2c may acquire (generate) type information based on the anomaly information generated by the autonomous system 7.

[0061] <Modification 3> In Embodiment 1, the identification unit 2e identified a work instruction for the actor to take over the work of the autonomous system 7 based on the first work status of the autonomous system 7 acquired when an abnormality occurred, but it is not limited to this.

[0062] For example, the actor may include a first actor and a second actor. As shown in Figure 11, the identification unit 2e may acquire the second work status of the first actor from the first actor via the input device 1. When an abnormality occurs, the identification unit 2e may acquire the first work status of the autonomous system 7 and the second work status of the first actor, and based on the first and second work statuses, identify a work instruction for either the first actor or the second actor to take over the work of the autonomous system 7. The notification control unit 2f may then perform control to notify the one of the first and second actors that will take over the work of the work instruction from the notification unit 3.

[0063] With this configuration, if the margin of the first actor's second work status is greater than the first threshold but less than or equal to the second threshold, the first actor can be notified of a reduced work instruction. Also, if the margin of the first actor's second work status is less than or equal to the first threshold, the work instruction can be notified to the second actor without notifying the first actor.

[0064] Furthermore, if an abnormality occurs, the specific unit 2e may further acquire the third work status of the second actor and, based on the first, second, and third work statuses, identify a work instruction for the first or second actor to take over the work of the autonomous system 7. In addition, the second work status of the first actor and the second work status of the second actor may be included in the peripheral environment information of the peripheral system 8.

[0065] <Other Modifications> The type information acquisition unit 2c and notification control unit 2f in Figure 1 described above will be referred to as "type information acquisition unit 2c, etc." below. The type information acquisition unit 2c, etc. is realized by the processing circuit 81 shown in Figure 12. That is, the processing circuit 81 includes a type information acquisition unit 2c that acquires type information, which includes at least one of an abnormality type and a response type, each expressed in natural language, based on abnormality information that indicates an abnormality in the autonomous system 7 and is expressed in a language other than natural language, and a notification control unit 2f that performs control to notify the actor of the type information from the notification unit 3. Dedicated hardware may be applied to the processing circuit 81, or a processor that executes a program stored in memory may be applied. Examples of processors include central processing units, processing units, arithmetic units, microprocessors, microcomputers, and DSPs (Digital Signal Processors).

[0066] If the processing circuit 81 is dedicated hardware, it may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. Each function of the type information acquisition unit 2c, etc., may be implemented by a circuit with distributed processing circuits, or the functions of each part may be implemented together by a single processing circuit.

[0067] When the processing circuit 81 is a processor, the functions of the type information acquisition unit 2c, etc., are realized in combination with software, etc. The software, etc., may include, for example, software, firmware, or both software and firmware. The software, etc., is written as a program and stored in memory. As shown in Figure 13, the processor 82 applied to the processing circuit 81 realizes the functions of each part by reading and executing the program stored in memory 83. That is, the abnormality detection device 2, when executed by the processing circuit 81, includes memory 83 for storing a program that ultimately executes the following steps: acquiring type information, which includes at least one of an abnormality type and a response type, each expressed in natural language, based on abnormality information expressed in a language other than natural language that indicates an abnormality in the autonomous system 7; and controlling the notification unit 3 to notify the actor of the type information. In other words, this program can be said to cause the computer to execute the procedures and methods of the type information acquisition unit 2c, etc. Here, memory 83 may be, for example, non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), EEPROM (Electrically Erasable Programmable Read Only Memory), HDD (Hard Disk Drive), magnetic disk, flexible disk, optical disk, compact disk, minidisc, DVD (Digital Versatile Disc), their drive devices, or any storage medium used in the future.

[0068] The above describes a configuration in which each function of the type information acquisition unit 2c, etc., is realized by either hardware or software. However, this is not the only configuration; a part of the type information acquisition unit 2c, etc., may be realized by dedicated hardware, and another part by software. For example, the function of the type information acquisition unit 2c can be realized by a processing circuit 81 as dedicated hardware, while the functions of the others can be realized by a processing circuit 81 as a processor 82 reading and executing a program stored in memory 83.

[0069] As described above, the processing circuit 81 can realize each of the above functions through hardware, software, or a combination thereof. The same applies to the actor information acquisition unit 2b and the like.

[0070] In this disclosure in English, the articles 'a' and 'an' mean one or more. Therefore, 'a', 'an', 'one or more', and 'at least one' can be used interchangeably.

[0071] Furthermore, it is possible to freely combine each embodiment and each variation, and to modify or omit each embodiment and each variation as appropriate. The above description is illustrative in all respects and not limiting. It is understood that countless variations not illustrated are conceivable.

[0072] 2 Anomaly detection device, 2a Anomaly detection unit, 2b Actor information acquisition unit, 2c Type information acquisition unit, 2d Urgency calculation unit, 2e Identification unit, 2f Notification control unit, 2g Learning information provision unit, 2h Anomaly prediction unit, 3 Notification unit, 7 Autonomous system, 9 Learning unit.

Claims

1. An anomaly detection device comprising: a type information acquisition unit that acquires type information, each expressed in natural language, that includes at least one of the anomaly type of the autonomous system and a type of action to take for the anomaly type, based on anomaly information that indicates an anomaly in an autonomous system and is expressed in a language other than natural language; and a notification control unit that controls the notification of the type information to an actor, which is a person related to the autonomous system, from a notification unit.

2. An anomaly detection device according to claim 1, further comprising an anomaly acquisition unit that acquires anomaly information based on at least one of the state information of the autonomous system and the surrounding environment information which is information of the surrounding environment of the autonomous system.

3. An anomaly detection device according to claim 2, wherein the anomaly acquisition unit includes an anomaly detection unit that detects the anomaly based on at least one of the state information of the autonomous system and the surrounding environment information.

4. An anomaly detection device according to claim 2, wherein the anomaly acquisition unit includes an anomaly prediction unit that predicts the behavior of the autonomous system based on at least one of the state information of the autonomous system and the surrounding environment information, and predicts the anomaly based on the prediction result.

5. An anomaly detection device according to any one of claims 1 to 4, further comprising an actor information acquisition unit that acquires actor information including at least one of actor state information and attribute information, wherein the type information acquisition unit acquires type information based on the anomaly information and the actor information.

6. An anomaly detection device according to any one of claims 1 to 4, further comprising an actor information acquisition unit that acquires actor information including at least one of actor state information and attribute information, wherein the notification control unit controls the notification unit to notify the actor of the type information based on the actor information.

7. An anomaly detection device according to claim 1, further comprising an emergency calculation unit that calculates the urgency of the anomaly based on at least one of the state information of the autonomous system, the surrounding environment information which is information of the surrounding environment of the autonomous system, and the anomaly information, wherein the notification control unit performs control to notify the actor of the urgency from the notification unit.

8. An anomaly detection device according to any one of claims 1 to 7, further comprising a specification unit that identifies a work instruction for the actor to take over the work of the autonomous system based on a first work status of the autonomous system acquired when the anomaly occurs, and the notification control unit performs control to notify the actor of the work instruction from the notification unit.

9. An anomaly detection device according to any one of claims 1 to 7, wherein the actor includes a first actor and a second actor, and further comprises a specification unit that identifies a work instruction for the first actor or the second actor to take over the work of the autonomous system based on a first work status of the autonomous system and a second work status of the first actor acquired when the anomaly occurs, and the notification control unit controls the notification unit to notify the one of the first actor and the second actor that takes over the work of the work instruction.

10. An anomaly detection device according to claim 1, further comprising: an anomaly acquisition unit that acquires anomaly information based on state information of the autonomous system and surrounding environment information which is information of the surrounding environment of the autonomous system; and a learning information provision unit that provides the anomaly information to the learning unit as learning information for the learning unit to learn the actions of the autonomous system that cause the anomaly.

11. An anomaly detection system comprising an anomaly detection device according to any one of claims 1 to 10, and the notification unit.

12. An anomaly detection method comprising: a type information acquisition unit that indicates an anomaly in the autonomous system and acquires type information, which includes at least one of the anomaly type of the autonomous system and a type of action to take for the anomaly type, each expressed in natural language, based on anomaly information expressed in a language other than natural language; and a notification control unit that controls the notification unit to notify an actor, who is a person related to the autonomous system, of the type information.

Images