Distributed behavior control device, distributed behavior control system, and distributed behavior control method

Abstract

To provide a distributed behavior control device for performing behavior control by autonomously distributing each asset in an asset group even when information on a behavior target is unknown in a region of a dispatch destination of the asset, a distributed behavior control system, and a distributed behavior control method.SOLUTION: In a distributed behavior control unit 103a in a distributed behavior control device 100, an environment setting unit 110 sets environment information including a target and a region regarding behavior of an asset group, an unknown information amount calculation unit 111 calculates an unknown information amount necessary for achieving a target of behavior of the asset group based on the environment information, a dispersion degree calculation unit 112 calculates a dispersion degree for the asset group to autonomously disperse and behave based on the unknown information amount, a control allocation unit 114 determines allocation of control relative to one or more assets to autonomously disperse and behave based on the dispersion degree, a control output unit 115 allocates the control, and a visualization unit 116 causes an output device to display behavior information of each asset.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to, for example, a distributed behavior control device, a distributed behavior control system, and a distributed behavior control method. [Background technology]

[0002] When determining how to control mobile objects such as drones to operate automatically using rule-based methods, there are cases where an asset group containing multiple mobile objects is the target of control. In such cases, it is necessary to appropriately set the destination for each mobile object, i.e., each asset, and have each asset act autonomously and cooperatively.

[0003] Therefore, as in Patent Document 1, there is a proposal to acquire information about objects present in an area that is a candidate destination for an asset, assume the movement speed of the object, and then determine the dispatch destination of the asset within the asset group. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Application Publication No. 7-244644 Summary of the Invention [Problem to be solved by the invention]

[0005] In the asset group described above, when determining the behavioral control of each asset, it is necessary to determine the destination behavior of the asset. That is, based on the amount of unknown information among the information about the asset's behavioral goal, it is necessary to determine the degree to which each asset should be dispersed, the area to which each asset should be dispatched, and the degree to which each asset should be coordinated.

[0006] However, in the above-mentioned Patent Document 1, the location information and number of objects, which are information about the behavioral goal in the area where the asset is dispatched, are treated as known, and it is difficult to apply it appropriately when the information about the behavioral goal in the area where the asset is dispatched is unknown.

[0007] The present invention has been made in view of the above background, and aims to provide a distributed behavior control device, a distributed behavior control system, and a distributed behavior control method that are capable of autonomously distributing and controlling the behavior of each asset within an asset group, even when information regarding behavioral goals is unknown in the area to which the asset is dispatched. [Means for solving the problem]

[0008] In order to solve the above-mentioned problems and achieve the above-mentioned object, one embodiment of the present invention is a distributed behavior control device that controls the behavior of one or more assets, and is characterized by comprising: an environment setting unit that sets environmental information including goals and areas regarding the behavior of the one or more assets; an unknown information amount calculation unit that calculates the amount of unknown information required to achieve the behavioral goal of the one or more assets based on the set environmental information; a dispersion degree calculation unit that calculates the degree of dispersion for the one or more assets to act autonomously and distributed based on the calculated amount of unknown information; a control allocation unit that determines a control allocation for the one or more assets to act autonomously and distributed based on the calculated degree of dispersion; and a control output unit that outputs the determined control allocation as behavior information of the one or more assets.

[0009] Another embodiment of the present invention is a distributed behavior control system comprising one or more assets and a device that communicates with the one or more assets to control the behavior of each of the assets, wherein the device sets an environment based on the environmental information, which is information on behavioral goals, asset information, and information specifying the asset's behavioral area, calculates an amount of unknown information from information on the behavioral goals to which the asset is commanded in the area in which the asset's behavior is to be performed, calculates a degree of distribution that determines the degree to which assets in an asset group including multiple assets will act autonomously and in a distributed manner based on the calculated degree of distribution, determines a control allocation for autonomously distributing the behavior of the assets in the asset group based on the calculated degree of distribution, and outputs the control allocation for each of the assets.

[0010] Furthermore, another embodiment of the present invention is a distributed behavior control method for a device that controls the behavior of one or more assets, wherein environmental information is information on behavioral goals, asset information, and information specifying the behavioral area of ​​the assets, and is characterized by including: an environment setting step for setting an environment based on the environmental information; an unknown information amount calculation step for calculating an unknown information amount from information on the behavioral goal to which the asset is commanded in the area in which the asset behavior set in the environment setting step is to be performed; a dispersion degree calculation step for calculating a dispersion degree that determines the degree to which assets in an asset group including multiple assets will act in an autonomously distributed manner based on the unknown information amount calculated in the unknown information amount calculation step; a control allocation step for determining a control allocation for autonomously distributing the behavior of the assets in the asset group based on the dispersion degree calculated in the dispersion degree calculation step; and a control output step for outputting the control allocation of each asset determined by the control allocation step. [Effects of the Invention]

[0011] According to the present invention, even if information regarding behavioral goals is unknown in the area where the assets are dispatched, it is possible to autonomously distribute the assets within the asset group and control their behavior. [Brief explanation of the drawings]

[0012] [Figure 1] 1 is a configuration diagram showing an example of a system configuration including a distributed behavior control device according to a first embodiment of the present invention. [Figure 2] 1 is a block diagram showing an example of the configuration of a distributed behavior control device according to a first embodiment of the present invention. [Figure 3] FIG. 2 is an explanatory diagram illustrating an example of a database according to the first embodiment of the present invention. [Figure 4] FIG. 2 is an explanatory diagram illustrating data according to the first embodiment of the present invention. [Figure 5] FIG. 2 is an explanatory diagram illustrating parameters according to the first embodiment of the present invention. [Figure 6] FIG. 2 is an explanatory diagram for explaining calculation of an unknown information amount according to the first embodiment of the present invention. [Figure 7] FIG. 2 is an explanatory diagram illustrating grid integration and rules according to the first embodiment of the present invention. [Figure 8] FIG. 1 is an explanatory diagram illustrating a rescue method according to a first embodiment of the present invention. [Figure 9] FIG. 1 is an explanatory diagram illustrating a simulation environment and a behavioral goal according to a first embodiment of the present invention. [Figure 10] 1 is a flowchart illustrating a procedure for decentralized behavior control according to the first embodiment of the present invention. [Figure 11] 1 is a flowchart illustrating an environment setting procedure according to the first embodiment of the present invention. [Figure 12] 3 is a flowchart illustrating a procedure for calculating an unknown information amount according to the first embodiment of the present invention. [Figure 13] 1 is a flowchart illustrating a main process of calculating an amount of unknown information according to the first embodiment of the present invention. [Figure 14]10 is a flowchart illustrating a modified example of the main process of calculating an amount of unknown information according to the first embodiment of the present invention. [Figure 15] 1 is a flowchart illustrating a procedure for calculating a dispersity according to the first embodiment of the present invention. [Figure 16] 1 is a flowchart illustrating a main process of calculating a dispersity according to the first embodiment of the present invention. [Figure 17] 10 is a flowchart illustrating a modified example of the main process of calculating the dispersity according to the first embodiment of the present invention. [Figure 18] 10 is a flowchart illustrating a procedure for determining an asset group according to the first embodiment of the present invention. [Figure 19] 1 is a flowchart illustrating a main process of determining an asset group according to the first embodiment of the present invention. [Figure 20] 10 is a flowchart illustrating a control allocation process according to the first embodiment of the present invention. [Figure 21] 3 is a flowchart illustrating the operation of the system according to the first embodiment of the present invention. [Figure 22] FIG. 10 is an explanatory diagram illustrating allocation of dispersibility and a table of dispersibility according to the second embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Note that the following description and drawings are merely examples for explaining the present invention, and some omissions and simplifications have been made as appropriate for clarity of explanation. The present invention can also be implemented in various other forms. Furthermore, unless otherwise specified, each component may be singular or plural.

[0014] In the following description, identical or similar components may be assigned the same reference numerals, and redundant description may be omitted. Furthermore, in the following description, various types of information may be described using expressions such as "information" and "table," but the various types of information may also be expressed using other data structures. Furthermore, identification information may be expressed using expressions such as "identification information," "identifier," "name," "ID," and "number," but these are interchangeable. Furthermore, in the following description, "table" will be abbreviated as "TBL."

[0015] First, the system of embodiment 1 will be described with reference to Fig. 1. Fig. 1 is a configuration diagram showing an example of a system configuration including a distributed behavior control device according to embodiment 1. As shown in Fig. 1, for example, the distributed behavior control device 100 wirelessly connects to an asset group 150 including drones 150-1 to 150-N (N is a natural number) which are examples of assets, and issues a command for behavior control such as rescue.

[0016] The distributed behavior control device 100 is configured, for example, as shown internally in FIG. 1, by a processor 301, a data storage device 302, a communication device 303, an input device 304, an output device 305, a program storage device 306, and the like.

[0017] The processor 301 controls the distributed behavior of the first embodiment in accordance with various programs stored in the program storage device 306. The data storage device 302 stores various data when the processor 301 executes the various programs. The program storage device 306 is configured with various programs that control the distributed behavior of the first embodiment.

[0018] The communication device 303 is a device that communicates with each of the drones 150-1 to 150-N of the asset group 150. The input device 304 is a device that includes a keyboard and buttons and is used by an operator to input information such as data, and the output device 305 is a display that displays various types of information on the screen in accordance with the operation of this embodiment.

[0019] Next, the functions of the first embodiment will be explained using Fig. 2. Fig. 2 is a block diagram showing an example of the configuration of the distributed behavior control device according to the first embodiment. The functions of the distributed behavior control device 100 are configured, for example, as shown in FIG. 2, by a data input / output unit 101 that realizes input / output using an input device 304 and an output device 305, a database 102 configured in a data storage device 302, and a distributed behavior control unit 103 that is realized by the processor 301 executing various programs in a program storage device 306.

[0020] The distributed behavior control unit 103 is composed of an environment setting unit 110, an unknown information amount calculation unit 111, a dispersion degree calculation unit 112, an asset group determination unit 113, a control allocation unit 114, a control output unit 115, a visualization unit 116, and a dispersion degree adjustment unit 117. The functions of each of the above units are realized when the processor 301 executes each program in the program storage device 306.

[0021] As will be described in detail later, the environment setting unit 110 has the function of setting the environment, and the unknown information amount calculation unit 111 calculates the amount of unknown information. The dispersion degree calculation unit 112 calculates the dispersion degree for assets to act autonomously and dispersedly, and the asset group determination unit 113 determines the asset group. The control allocation unit 114 allocates control to each asset, i.e., determines behavior information, and the control output unit 115 outputs the behavior information.

[0022] The visualization unit 116 has a display function of the output device 305, and displays and visualizes the behavioral information. The distribution degree adjustment unit 117 has an operation input function of the input device 304, and adjusts the distribution degree by changing the control allocation displayed by the visualization unit 116, and notifies the control allocation unit 1114 of the adjustment result.

[0023] Next, the database will be described in detail. Fig. 3 is an explanatory diagram showing an example of the database according to the first embodiment, Fig. 4 is an explanatory diagram explaining data according to the first embodiment, Fig. 5 is an explanatory diagram explaining parameters according to the first embodiment, and Fig. 6 is an explanatory diagram explaining calculation of the amount of unknown information according to the first embodiment.

[0024] 7 is an explanatory diagram illustrating grid integration and rules according to embodiment 1, FIG. 8 is an explanatory diagram illustrating a rescue method according to embodiment 1, and FIG. 9 is an explanatory diagram illustrating a simulation environment and behavioral goals according to embodiment 1.

[0025] 3, the database 102 is composed of environmental information data 200, asset information data 201, behavioral goal information data 202, unknown information data 203, behavior information data 204, dispersion data 205, cooperation data 206, route information data 207, control method data 208, simulation environment data 209, asset behavior range data 210, grid information data 211, asset group data 212, and rule data 213. Details of each data will be described later.

[0026] As shown in Fig. 4, the explanation range 40 includes a simulation environment 1500 (simulation environment data 209) indicating geospatial information, environment information data 200, and asset information data 201. The simulation environment 1500 will now be explained using Fig. 4. An area 1501 within the simulation environment 1500 indicates an asset movement range 1502 (asset movement range data 210).

[0027] This asset action range 1502 shows an example in which a rescue target 1504 is in an area with a low amount of unknown information, while a rescue target 1505 is in an asset action range 1503 with a high amount of unknown information. Three drones, 150-1, 150-2, and 150-3, are deployed in this asset action range 1502 as an asset group (asset group data 212).

[0028] Next, the simulation environment will be described with reference to Fig. 4. The environment information data 200 is a TBL that stores the behavior area x(m), behavior area y(m), behavior area z(m), obstacle 1 position x(m), obstacle 1 position y(m), obstacle 1 position z(m), grid 1 ID, grid 1 information, etc., linked to an environment ID that identifies the environment.

[0029] The size of the grid is defined as the same size, but in the integration described later, the region indicating the degree of unknowability is variable.

[0030] Simulation environment 1500 includes information about area 1501, and grid 1 includes the amount of unknown information, the degree of cooperation, and the degree of dispersion included in grid 1. When the amount of unknown information is integrated, the grid information is updated. In the example of FIG. 4, the environment ID is associated with "0," and the behavior area x is 1,000 (m), the behavior area y is 1,000 (m), the behavior area z is 30 (m), the obstacle 1 position x is 200 (m), the obstacle 1 position y is 200 (m), the obstacle 1 position z is 10,000 (m), the ID of grid 1 is "0," and the information of grid 1 is (0,0), (1000,500).

[0031] The asset information data 201 will be further described with reference to Fig. 4. The asset information data 201 is a TBL that stores, linked to an asset ID, x (m), y (m), z (m) as the x (axis), y (axis), and z (axis) of the position, and vx (m / s), vy (m / s), and vz (m / s) as the vx (axial direction), vy (axial direction), and vz (axial direction) of the maximum speed.

[0032] In the example of Figure 4, for asset ID "0," the position is x 1 (m), the position y is 2 (m), and the position z is 10 (m), and the maximum speed is vx -0.1 (m / s), vy is 10 (m / s), and vz is 5 (m / s). Similarly, for asset ID "1," the position is x 5 (m), the position y is -0.1 (m), and the position z is -1 (m), and the maximum speed is vx 0.01 (m / s), vy is 0 (m / s), and vz is 5 (m / s). Asset information data 201 includes position information such as the speed of each asset (drone).

[0033] In the case of FIG. 4, two drones 150-1 and 150-2 are assigned to a rescue target 1504 on the upper side (area with low unknown information) based on the degree of coordination, and one drone 150-3 is assigned to a rescue target 1505 on the lower side (asset action range 1503 with high unknown information). The two coordinated drones 150-1 and 150-2 may act together, move to the same area 1501, or act separately. This control is changed according to the degree of coordination. The number of assets assigned is determined according to the degree of unknown information.

[0034] Next, the various data will be described in detail using Fig. 5. As shown in Fig. 5, the explanation area 50 includes behavioral goal information data 202, behavior information data 204, dispersion degree data 205, cooperation degree data 206, asset behavior range data 210, grid information data 211, and asset group data 212.

[0035] The behavioral goal information data 202 is a TBL that stores information such as information range 1, information range 2, and information number of missing persons, linked to a behavioral goal ID that identifies the behavioral goal. As an example, corresponding to a behavioral goal ID of "0," information range 1 is (30, 50, 0), information range 2 is (100, 200, e0), and information number of missing persons is "1." Also, corresponding to a behavioral goal ID of "1," information range 1 is NAN, information range 2 is NAN, and information number of missing persons is NAN.

[0036] The behavior information data 204 is a TBL that stores behaviors, such as rescue or search, indicating the content of the behavior, the degree of cooperation with other assets, and the amount of unknown information, indicating the level of the amount of unknown information, linked to a behavior ID that identifies the behavior. As an example, for behavior ID "0," the behavior is "rescue," the degree of cooperation is "0.5," and the amount of unknown information is "low." Also, for behavior ID "1," the behavior is "search," the degree of cooperation is "0," and the amount of unknown information is "high."

[0037] In the case of action ID "0", the purpose is rescue, and a certain area has been confirmed, so the amount of unknown information is low, and the data structure is for assets to act in coordination with each other. On the other hand, in the case of action ID "1", the purpose is search, and the location is difficult to pinpoint, so the amount of unknown information is high, and the data structure is for assets to act without coordination with each other.

[0038] The dispersion degree data 205 is a TBL that stores grid IDs and dispersion degrees, linked to dispersion degree IDs that identify the dispersion degrees. For example, for a dispersion degree ID of "0", the grid IDs are "0, 1" and the dispersion degree is "1".

[0039] The asset movement range data 210 is a TBL that stores geographical information consisting of movement area x (m), movement area y (m), and movement area z (m) linked to an ID that identifies the asset movement range. As an example, for ID "0", the movement area x is 1,000 (m), y is 1,000 (m), and z is 30 (m). Also, for ID "1", the movement area x is 1,000 (m), y is 1,000 (m), and z is 30 (m).

[0040] The grid information data 211 is a table of contents that defines the x-axis and y-axis of the grid size. For example, the grid size is 1,000 (m) on the x-axis and (m) on the y-axis.

[0041] The asset group data 212 is a TBL that stores the number of assets, asset IDs included, degree of collaboration, and amount of unknown information, linked to an asset group ID that identifies the asset group. As an example, for an asset group ID of "0," the number of assets is "2," the asset IDs included are "0.1," the degree of collaboration is "0.5," and the amount of unknown information is "low." For an asset group ID of "1," the number of assets is "1," the asset IDs included are "2," the degree of collaboration is "0," and the amount of unknown information is "high."

[0042] In the above, when there is one degree of dispersion from the dispersion data 205, the grid ID is obtained from the grid ID listed in the simulation environment data 209. Here, the grid is divided into two, "0" and "1", and since only one degree of dispersion is calculated, the two values ​​"0" and "1" are stored in the grid ID column.

[0043] Next, various types of data will be described in detail with reference to Fig. 6. As an example, two examples of explanation areas 1700 and 1710 are shown in the explanation area 60, as shown in Fig. 6. Here, the data in the explanation areas 1700 and 1710 are represented by unknown information data 203.

[0044] The explanation area 1700 shows the amount of unknown information, i.e., the degree of unknownness, in each divided area (grid) in the asset behavior range 1502, and assigns a degree of unknownness to each grid of 4 rows and 4 columns. For example, in the case of grid 1701, the degree of unknownness is "0".

[0045] Here, unknown information is information necessary for an asset to take action and achieve the goal of the action. An unknown degree of "0" indicates a low unknown degree, which is applied when, for example, it is known that a person in need of rescue exists in the area of ​​action.

[0046] On the other hand, an unknown degree of "1" indicates a high unknown degree value, which is applied when, for example, it is not even known that a person in need of rescue exists in the area of ​​operation.

[0047] The explanation region 1710 is a region obtained by reducing the number of divisions compared to the explanation region 1700, and is assigned ranges 1711, 1712, and 1713 that refer to surrounding grids. Only one degree of unknowability is determined for each of the ranges 1711, 1712, and 1713 based on a majority vote. The numbers in the grids may be set when the visualization unit 116 outputs the data.

[0048] Next, grid integration will be explained using Fig. 7. In the explanation area 70, an integration example 1800, control method data 208, and rule data 213 are shown as shown in Fig. 7.

[0049] The control method data 208 is a TBL that stores behavior IDs and rule IDs linked to asset IDs. For example, when asset ID is "0", the behavior ID is "0" and the rule ID is "0". When asset ID is "1", the behavior ID is "0" and the rule ID is "1", and when asset ID is "2", the behavior ID is "1" and the rule ID is "0".

[0050] The rule data 213 is a TBL that stores rule contents linked to rule IDs. For example, a rule that "move to the center of the area as quickly and as short a distance as possible, and move in a circle" is set for the asset in the rule ID "0," and a rule that "move in a serpentine pattern" is set for the asset in the rule ID "1."

[0051] Furthermore, a rule called "random movement" is set for the asset in rule ID "2," and a rule called "asset decision" is set for the asset in rule ID "3." In this case, when the asset makes the decision, the asset may be autonomously distributed and decide to move using a technique such as machine learning.

[0052] As an integrated example 1800, integrated grids 1801 and 1802 are shown within the asset behavior range 1502. The integrated grid 1801 is made up of the top two rows of the asset behavior range 1502 in FIG. 6, and the integrated grid 1802 is made up of the bottom two rows of the asset behavior range 1502 in FIG.

[0053] For integrated grid 1801, a majority vote is taken to determine the degree of unknowability in each of ranges 1711, 1712, and 1713, which are made up of multiple (four or two) adjacent grids. For range 1711, the degree of unknowability is "0," for range 1712, the degree of unknowability is "0," and for range 1713, the degree of unknowability is "0." As a result, for integrated grid 1801, a majority vote is also taken to determine the degree of unknowability in ranges 1711, 1712, and 1713, and the degree of unknowability is determined to be "0." A similar majority vote is taken for integrated grid 1802, and the degree of unknowability is determined to be "1."

[0054] In the control method data 208, an asset ID of "0" is associated with an action ID of "0" and a rule ID of "0." In this case, the rule for the movement of the asset, i.e., the drone, is to "move to the center of the area as quickly and as shortly as possible, and move in a circle." Also, an asset ID of "1" is associated with an action ID of "0" and a rule ID of "1." In this case, the rule for the movement of the asset, i.e., the drone, is to "move in a serpentine pattern."

[0055] Regarding the degree of dispersion, for example, when rescuing a person, it is more effective to have drones, which are assets, work together to some extent rather than dispersing them individually. Also, when searching for a person, for example, it is more effective to have drones work together in a dispersed manner rather than coordinating their actions.

[0056] Here, the degree of cooperation is specified within the range of 0 to 1, and is a numerical value set in advance for the action. In the case of rescue, the degree of cooperation is 0.5, and in the case of search, the degree of cooperation is 0. Furthermore, the degree of unknown is specified within the range of 0 to 1, and the degree of dispersion is calculated by the four basic arithmetic operations of (1 - degree of cooperation) x unknown. In other words, the degree of dispersion is calculated for each area with a different degree of unknown.

[0057] Next, movement routes will be explained using Figure 8. In the asset movement range 1502, as shown in Figure 8, an example is shown in which movement route 80 is set in a spiral shape in the direction of the arrow in accordance with rule ID "0" in an integrated grid 1801. Here, the movement route of one asset, i.e., a drone, is shown, but if multiple drones are moving, multiple movement routes will be set.

[0058] Next, other simulation environments will be described using a picking robot as an example of an asset with reference to Fig. 9. As shown in Fig. 9, the description range 90 includes a simulation environment 2100 (simulation environment data 209) and behavioral goal information data 202.

[0059] 9, a simulation environment 2100 will be described. An asset movement range 2102 (asset movement range data 210) and picking robots P1 to P5, as an example, are shown in an area 2101 within the simulation environment 2100. Grids 2103 and 2104, to which the picking robots are dispatched, are set in the asset movement range 2102.

[0060] Picking robots P1 and P2 are assigned an asset movement range in grid 2103, and picking robot P3 is assigned an asset movement range in grid 2104. Furthermore, picking robots P4 and P5 are not assigned an asset movement range in grid 2103, and are in a standby state outside asset movement range 2102.

[0061] In the example of Figure 9, the behavioral goal information data 202 is a TBL that stores information: position x, information: position y, information: asset speed (m / s), information: amount of luggage that has not been picked, initial luggage amount (pieces), current luggage amount (pieces), applied rule, and dispatch time (minutes), linked to the behavioral goal ID.

[0062] In the example of Figure 9, when the behavioral goal ID is "0", the information: position x is [0, 100], information: position y is [0, 100], information: asset speed is 1 (m / s), information: amount of luggage that has not been picked is NAN, initial luggage amount is 100 (pieces), current luggage amount is NAN (pieces), applied rule is "0", and dispatch time is 30 (minutes).

[0063] Also, when the action goal ID is "1" (top row), the information: position x is [0, 100], information: position y is [0, 100], information: asset speed is 1 (m / s), information: amount of luggage that has not been picked is NAN, initial luggage amount is 100 (pieces), current luggage amount is NAN (pieces), applied rule is "0", and dispatch time is 30 (minutes).

[0064] Furthermore, when the action goal ID is "1" (bottom row), the information: position x is [0, 100], the information: position y is [100, 200], the information: asset speed is 1 (m / s), the information: amount of luggage that has not been picked is NAN, the initial luggage amount is 50 (pieces), the current luggage amount is NAN (pieces), the applied rule is "3", and the dispatch time is 15 (minutes).

[0065] Based on the approximate position, speed, and initial baggage volume of the picking robot, the workload within the dispatch destination grid can be predicted and the dispatch of the next picking robot can be controlled. Note that the rule ID is handled in the same way as described above.

[0066] Here, the control rules are summarized. Control allocation is performed in the following way. <When calculating only one dispersity> 1. Multiply the number of assets in the asset group you are taking action on by the diversification factor. Round up the figure. 2. It is necessary to distinguish whether the commanded action is for an area with high unknownness or an area with low unknownness. In this case, in the case of rescue, since it is known that the person to be rescued is in the area to be acted upon, the asset moves to the person to be rescued in an area with low unknownness and sends a rescue signal. In the case of search, since it is not known that the person to be rescued is in the area to be acted upon, the asset conducts a search as if the area were high in unknownness. 3. The group of assets containing the total number of assets minus (the number obtained in 1 above) is designated as one of the asset subgroups and assigned to the region of unknownness identified in 2 above. 4. In areas other than those listed above, create subgroups of assets based on the unknown level of each area, using the number of assets in the asset group minus the number of assets included in the areas listed above. 5. For the subgroups of assets assigned in 4 above, assign actions in order of areas of greatest uncertainty. <Calculating the degree of dispersion in each region> 1. A group containing a number of assets equal to the total number of assets minus the number multiplied by the unknownness of the area with the highest unknownness is defined as one of the asset subgroups. 2. Multiply the number of assets excluding the number in 1 above by the diversification of each area. Round up the number. 3. Create sub-asset groups for each area with different degrees of ambiguity, and allocate the number of assets indicated in 2 above to each asset group. If the number of assets exceeds the total number of assets, reduce the number of assets in sub-asset groups with low diversification. 4. Assign actions to sub-asset groups.

[0067] It is also possible to have drones disperse within the same unknown area depending on the degree of dispersion. For example, when searching, drones in a certain unknown area can be created separately.

[0068] Next, the operation of the first embodiment will be described with reference to Fig. 10 to Fig. 20. Fig. 10 is a flowchart explaining the decentralized behavior control procedure according to the first embodiment, Fig. 11 is a flowchart explaining the environment setting procedure according to the first embodiment, Fig. 12 is a flowchart explaining the procedure for calculating the amount of unknown information according to the first embodiment, Fig. 13 is a flowchart explaining the main processing for calculating the amount of unknown information according to the first embodiment, and Fig. 14 is a flowchart explaining a modified example of the main processing for calculating the amount of unknown information according to the first embodiment.

[0069] 15 is a flowchart explaining the procedure for calculating the degree of dispersion according to embodiment 1, Fig. 16 is a flowchart explaining the main processing for calculating the degree of dispersion according to embodiment 1, Fig. 17 is a flowchart explaining a modified example of the main processing for calculating the degree of dispersion according to embodiment 1, Fig. 18 is a flowchart explaining the procedure for determining an asset group according to embodiment 1, Fig. 19 is a flowchart explaining the main processing for determining an asset group according to embodiment 1, and Fig. 20 is a flowchart explaining the control allocation processing according to embodiment 1. The processing of each flowchart is realized when the processor 301 executes a program in the program storage device 306.

[0070] First, the operation of the distributed behavior control unit 103 will be described with reference to Fig. 10. The environment setting unit 110 acquires environment information, asset information, and behavior goal information, and sets the environment information data 200 in the DB 102 (step S401).

[0071] The unknown information amount calculation unit 111 calculates the amount of unknown information that has not been acquired as information among the information necessary for the action goal within the action range of the asset defined by the environmental information (step S402). Note that various data and information being processed are temporarily stored in the data storage device 402.

[0072] The dispersion degree calculation unit 112 calculates the dispersion degree of the asset's behavior range using the unknown information amount calculated in step S402 and the degree of cooperation recorded in the behavior information of the behavior information data 204 (step S403).

[0073] The asset group determination unit 113 acquires the unknown information amount calculated in step S402, the dispersion degree calculated in step S403, and the asset information acquired in step S401. The asset group determination unit 113 further determines the number of actions to be assigned to different areas (grids) within the asset's behavioral range, and determines sub-asset groups of the asset corresponding to that number (step S404).

[0074] The control allocation unit 114 allocates a control for carrying out an action to the sub-asset group determined in step S404 (step S405). This allocation of control means the behavioral information of the asset. The control output unit 115 outputs the behavioral information allocated to each asset in step S405 (step S406).

[0075] The visualization unit 116 outputs the environment and asset's range of movement, the amount of unknown information in the environment, the degree of dispersion, the degree of coordination, the control assigned to the sub-asset group, and the expected movement route of the asset to the data input / output unit 101 (step S407).

[0076] Next, the processing of each unit shown in Fig. 10 will be specifically described with reference to Fig. 11 to Fig. 20. In the environment setting unit 110, as shown in Fig. 11, environmental information, asset information, and behavioral goal information are acquired (step S501), and the asset behavior range data 210 is set using the acquired various information, and the simulation environment data 209 is set (step S502).

[0077] 12, the unknown information amount calculation unit 111 acquires action goal information, asset action range, simulation environment, and grid information (step S601). Furthermore, it is determined whether information on each object (such as a rescue target) in the action goal information is unknown information in the simulation environment, and the amount of unknown information is calculated according to the number of pieces of unknown information (step S602). Then, the degree of unknownness is calculated according to the amount of unknown information (step S603).

[0078] Step S602 will now be described in detail with reference to Fig. 13. In step S602, first, a process is executed to divide the asset movement range in the simulation environment into fixed time intervals based on the grid size of the grid information (step S6021).

[0079] Next, for each divided area, it is determined whether the information of each object in the behavioral goal information is unknown information in the simulation environment (step S6022), and a process is executed to calculate the amount of unknown information according to the number of unknown information (step S6023).

[0080] Here, the calculation of the degree of unknownness by the unknown information amount calculation unit 111 will be described with reference to Fig. 14. After the processes of steps S6021 to S6023 are executed, the unknown information amount of the grids surrounding each divided area is referenced, and the unknown information amount of each divided area is calculated (step S6024).

[0081] Furthermore, for each divided region, if the unknown information amounts of adjacent regions are equal, the corresponding regions are merged into one and the unknown information amount is assigned (step S6025). Note that various methods are available for assigning the unknown information amount of the merged region, such as majority vote, average, maximum, or minimum value of the unknown information amounts of the surrounding regions. Then, the degree of unknownness is calculated based on the unknown information amount of each region (step S6026).

[0082] 15, the dispersion degree calculation unit 112 first acquires behavioral information (step S901), and then acquires the amount of unknown information or the degree of unknownness set in the asset behavior range of the simulation environment (step S902). Then, the dispersion degree is calculated based on the degree of cooperation and the amount of unknown information or the degree of unknownness described in the behavioral information (step S903).

[0083] Here, the dispersion degree calculation method in step S903 will be described in detail with reference to Fig. 16. First, the amount of unknown information or the degree of unknownness set in the asset behavior range of the simulation environment is acquired (step S9031), and the largest of the acquired amounts of unknown information or the degree of unknownness is acquired (step S9032). Then, the dispersion degree is calculated based on the largest amount of unknown information or the degree of unknownness and the degree of cooperation described in the behavior information (step S9033).

[0084] The dispersion degree may be calculated by the method shown in Fig. 17. After the processes of steps S901 and S902 are executed, the dispersion degree is calculated for each divided area based on the cooperation degree and the unknown information amount or unknown degree described in the behavior information (step S1101).

[0085] 18, the asset group determination unit 113 acquires asset information and calculates the total number of assets (step S1201). The asset group determination unit 113 further acquires the degree of dispersion, behavioral information, and simulation environment, and determines sub-asset groups from the total number of assets (step S1202).

[0086] The asset group determination method in step S1202 will now be described in detail with reference to Fig. 19. First, the degree of diversification is acquired (step S1301), and it is determined whether there are two or more degrees of diversification (step S1302).

[0087] If the determination result indicates that there are two or more degrees of dispersion (YES route in step S1302), a number is obtained by subtracting the value obtained by performing arithmetic operations on the degree of uncertainty of the area with the highest degree of uncertainty from the total number of assets, and the group containing the number of main assets with this obtained value is determined to be one of the sub-asset groups (step S1303).

[0088] Next, the dispersion degree of each divided area is applied to the total number of assets excluding the number of main assets by arithmetic operations to calculate the number of dispersed assets (step S1304). Furthermore, sub-asset groups are created for each area with a different degree of unknownness, and the number of assets is calculated by allocating the dispersed asset number to each asset group (step S1305).

[0089] Next, if the sum of the calculated total number of assets and the asset main number exceeds the total number of assets, the number of assets in the sub-asset group with low diversification is reduced (step S1306). Then, the asset sub-groups and the number of assets to be allocated to each group are determined (step S1311).

[0090] Also, in step S1302, if the determination result is that the dispersion degree is 1 (NO route in step S1302), arithmetic operations are performed using the dispersion degree on the number of assets in the asset group to be subjected to the action, and the asset dispersion number is calculated (step S1307).

[0091] Then, it is identified whether the behavior commanded from the behavior information is a behavior in an area with a high degree of unknownness or a behavior in an area with a low degree of unknownness (step S1308). Next, a group of the number of assets including the number of main assets obtained by subtracting (the number 1) from (the total number of assets) is assigned as one of the asset subgroups to the unknown area of ​​the identified behavior (step S1309).

[0092] Then, asset subgroups are created based on the number of assets obtained by subtracting the main number of assets from the total number of assets, according to the degree of unknownness of each area (step S1310). After this, the process proceeds to step S1311, where the asset subgroups and the number of assets to be assigned to each group are determined.

[0093] As shown in FIG. 20, the control allocation unit 114 first acquires asset groups and behavior information (step S1401), and allocates behavior commands to each divided area based on information on behavioral goals for each divided area in the simulation environment (step S1402).

[0094] As described above, according to embodiment 1, it is possible to allocate control to assets efficiently and in a balanced manner by estimating variance from the amount of unknown information and allocating control rules based on the variance.

[0095] Specifically, in an environment in which assets are made to act, it is important to grasp the amount of unknown information relative to the information necessary for the asset's behavioral goal, calculate the degree of dispersion for an asset group consisting of multiple assets from the linkages linked to the behavior and the amount of unknown information, and divide the assets within the asset group into certain groups based on this calculated degree of dispersion, and disperse each group to select rule control for behavior.

[0096] Furthermore, it is possible to determine to what extent each asset should be distributed to act, or whether the assets should be linked together and act in coordination, based on whether each asset should be linked to an action and how much unknown information is contained in the area in which the action will be carried out. In other words, when each asset is automatically controlled based on rules, it leads to efficient and balanced movement of assets, rather than moving them haphazardly.

[0097] In this way, the control action of each asset can be determined based on the amount of information of the destination where the asset will be executed, so that the allocation of assets can be efficiently and balancedly allocated. Also, since the degree of asset distribution can be confirmed, the degree of distribution can be adjusted in accordance with the user's desire to move assets in a moderately distributed manner, or conversely, to move assets in a widely distributed manner. In other words, if the degree of distribution is not what the user desires, it can be adjusted. Furthermore, the user can confirm the future information amount of the destination where each asset will execute its action.

[0098] Furthermore, the user can be assisted in adjusting the asset distribution degree by a distribution degree adjustment unit 117 such as a GUI (Graphical User Interface) that can change the amount of unknown information of the environment, distribution degree information, and distribution degree.

[0099] Furthermore, if the existence of a person to be rescued is unknown in the target area, the environment is unknown, i.e., the area has a large amount of unknown information, and information is collected by dispersing assets (drones) as much as possible. In this case, the degree of dispersion is set to large.

[0100] Furthermore, if it is known that rescue targets exist in some areas, but their existence is unknown in other areas, the environment is unknown, but the amount of unknown information is medium, and information is collected in these areas with different distributions. In this case, the degree of distribution is set to medium, and assets in these areas are not distributed, but are moved in a coordinated manner.

[0101] Also, if it is known that there are multiple people to be rescued in a certain area of ​​the target area, the environment is unknown, but the amount of unknown information is small, so the degree of dispersion is set to small. In this case, the assets are not dispersed, but are moved in a coordinated manner.

[0102] Here, the operation of the entire system will be described with reference to Fig. 21. Fig. 21 is a flowchart explaining the operation of the system according to the present embodiment 1. Fig. 21 shows, as a premise, the processing from the stage when the behavioral information of each asset has been determined, and each asset, i.e., each drone, is in a standby state.

[0103] When the distributed behavior control device 100 transmits behavior information to each drone (step S2201), each drone receives the behavior information (step S2101) and starts an action (for example, rescue) based on the behavior information (step S2102).

[0104] Each drone continues the instructed operation until the end of the action (NO route in step S2103). Then, when it is determined that the action has ended (YES route in step S2103), an end notification is sent to the distributed behavior control device 100 (step S2104). In this way, processing on the asset side is executed.

[0105] When the distributed behavior control device 100 receives an end notification from each drone, it executes an end process, and stores and manages the completion of the process based on the current behavior information in association with each drone (step S2202).

[0106] Regarding this storage management, the progress of the action may be managed by linking the action goal information data 202 with an action goal ID and providing an end flag, for example. This storage information is a matter of design and can be modified in various ways.

[0107] Next, a second embodiment of the present invention will be described with reference to the drawings. Fig. 22 is an explanatory diagram illustrating the allocation of dispersion degrees and a dispersion degree table according to the second embodiment. Fig. 22 uses a picking robot as an example of an asset, and in an explanation area 2200, an unknown degree is assigned to each divided area (grid), and a dispersion degree is assigned to each area.

[0108] Two integrated grid areas 2202 and 2203 are set within the asset movement range 2201. In Fig. 22, two picking robots are dispatched to the integrated grid area 2202 on the upper side.

[0109] According to rule ID "0", the robot first moves to the center and then moves in a circle from there to work, so there will be few packages near the center, and there will likely still be more packages on the edges of the area, but it is unknown how many there are.

[0110] In Figure 22, in the integrated grid area 2203 on the lower side, the picking robots are performing picking work using machine learning, so it is unclear how much of the work has been completed overall, and the degree of uncertainty is high. Because the picking robot on the lower side was dispatched from the right side, the degree of uncertainty on the right side is estimated to be slightly smaller. This is estimated based on machine learning and mathematical methods.

[0111] In the second embodiment, the degree of cooperation is set to "0", and the degree of dispersion is determined only by the degree of unknownness. However, since the method for calculating the degree of dispersion is merely an example, the degree of dispersion may take a different value.

[0112] Grid ID numbers are assigned from the top left to the right, three columns in, then the next row, three columns in, and so on. Grid ID "0" has an unknown degree of 0.6, and grid ID "5" has an unknown degree of 0.3. Finally, the two additional picking robots will be dispatched to grid IDs "6" and "9."

[0113] In this second embodiment, as in the first embodiment described above, even if information regarding behavioral goals is unknown in the area to which the assets are dispatched, it is possible to autonomously distribute and control the behavior of each asset within the asset group.

[0114] In the above-described first embodiment, in the example of FIG. 6, when the rescue target 1504 is found in an area of ​​the asset action range 1502 with an unknownness level of "0" and the rescue is completed, the action of the asset in charge of that area, i.e., the drone, is completed in accordance with the rules. The present invention is not limited to this. In the example of FIG. 6, if the rescue target 1504 is found but the rescue target is not yet found in the area of ​​the asset action range 1503 with an unknownness level of "1," the rule may be set so that the drone assigned to the rescue of the rescue target 1504 is not returned, but the rescue action is controlled in cooperation with the asset assigned to that area. Of course, if the battery or the like needs to be charged, the drone may be returned once and then used for the rescue.

[0115] In the above-described embodiment, drones and picking robots are used as examples of assets, but the present invention is not limited to this and can be applied to a dispatch system, such as dispatching an ambulance in the event of a disaster. In this way, the behavioral control of asset groups in highly unknown areas can be fully applied to the field of mobile objects, including taxis.

[0116] Furthermore, the above-described configurations, functional units, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a memory, a hard disk, a recording device such as an SSD (Solid State Drive), an IC card, an SD card, a DVD, or other recording media.

[0117] Furthermore, the above-described layout of the various functional units, processing units, and databases of the information collection system 10 is merely an example. The layout of the various functional units, processing units, and databases can be changed to an optimal layout in terms of the performance, processing efficiency, communication efficiency, etc. of the hardware and software provided in these devices.

[0118] Furthermore, the configuration (schema, etc.) of the database that stores the various types of data described above can be flexibly changed from the perspective of efficient use of resources, improved processing efficiency, improved access efficiency, improved search efficiency, and the like. [Explanation of symbols]

[0119] 100 Distributed behavior control device 101 Data input / output unit 102 DB 103 Distributed Behavior Control Unit 110 Environment Settings 111 Unknown information calculation unit 112 Dispersion degree calculation section 113 Asset Group Determination Department 114 Control Allocation Unit 115 Control output section 116 Visualization section 117 Dispersity adjustment section 150 Asset Groups 150-1 Drone 150-N Drone 200 Environmental Information Data 201 Asset Information Data 202 Action Target Information Data 203 Unknown Information Data 204 Behavioral Information Data 205 Dispersion Data 206 Collaboration Data 207 Route Information Data 208 Control Method Data 209 Simulation Environment Data 210 Asset Activity Range Data 211 Grid Information Data 212 Asset Group Data 213 Rule Data 301 processor 302 Data storage device 303 Communication Equipment 304 Input Device 305 Output Device 306 Program storage device

Claims

1. A distributed behavior control device that controls behavior of one or more assets, an environment setting unit that sets environment information including a goal and an area regarding the behavior of the one or more assets; an unknown information amount calculation unit that calculates an unknown information amount required to achieve a behavioral goal of the one or more assets based on the set environmental information; a dispersion degree calculation unit that calculates a dispersion degree for the one or more assets to act autonomously and in a dispersed manner based on the calculated amount of unknown information; a control allocation unit that determines an allocation of control for autonomously and decentralized behavior with respect to the one or more assets based on the calculated degree of decentralization; a control output unit that outputs the determined control allocation as behavior information of the one or more assets; A distributed behavior control device comprising:

2. 2. The distributed behavior control device according to claim 1, wherein the one or more assets are asset groups, and further comprising an asset group determination unit that performs arithmetic operations on the degree of dispersion using information on the number of assets in the asset group to determine the number of sub-asset groups, and determines sub-asset groups in areas with a large amount of unknown information, and wherein the control allocation unit allocates control to the determined sub-asset groups.

3. In the distributed behavior control device described in claim 2, the amount of unknown information is expressed as an unknown degree, the unknown information amount calculation unit calculates the unknown degree at predetermined intervals in the area where the behavior of the one or more assets is performed and assigns an unknown degree to each of the areas, the dispersion degree calculation unit calculates the dispersion degree in each area based on the unknown degree determined for each of the areas, the asset group determination unit determines the number of sub-asset groups using the largest dispersion degree among the dispersion degrees in each of the areas and determines sub-asset groups for the areas with the high unknown degree, and the control allocation unit allocates control to the sub-asset groups.

4. In the distributed behavior control device described in claim 2, the unknown information amount calculation unit calculates the degree of unknowability at predetermined intervals in the area where the asset behavior is performed, takes a majority vote on the unknowability of each adjacent area to update it to the largest unknowability, and connects areas with the same updated unknowability to make them larger, the dispersion calculation unit calculates the dispersion for the unknowability of each area, the asset group determination unit determines the number of sub-asset groups using the largest dispersion among the dispersions in each area, determines sub-asset groups for areas with high unknowability, and the control allocation unit allocates control to the sub-asset groups.

5. The distributed behavior control device of claim 2 further comprises a visualization unit that displays the degree of dispersion and a dispersion adjustment unit that adjusts the degree of dispersion in accordance with user operation, wherein the asset group determination unit determines the number of sub-asset groups based on the degree of dispersion adjusted by the dispersion adjustment unit, determines sub-asset groups in areas with a large amount of unknown information, the control allocation unit allocates control to the sub-asset groups, the control output unit outputs the control allocation to the visualization unit, and the dispersion adjustment unit adjusts the degree of dispersion by changing the control allocation displayed on the visualization unit.

6. In the distributed behavior control device described in claim 5, the control allocation unit performs arithmetic operations on the degree of dispersion using the number of assets in the asset group to determine the number of assets that will act in a dispersed manner, calculates a weight for each area from the area with the largest amount of unknown information, calculates the number of assets included in a sub-asset group that will act in a dispersed manner according to the calculated weight, and allocates control to the sub-asset group.

7. 3. The distributed behavior control device according to claim 2, wherein the unknown information amount calculation unit, when calculating the unknown information amount in the behavioral area of ​​the one or more assets, calculates the unknown information amount based on movement information such as the speed of targets related to the behavioral target of each area, the dispersion degree calculation unit calculates a dispersion degree that determines the degree to which assets in an asset group including multiple assets will autonomously distribute their behavior based on the unknown information amount, and the control allocation unit allocates control to cause the assets in the asset group to autonomously distribute their behavior based on the dispersion degree.

8. 2. The distributed behavior control device according to claim 1, further comprising an unknown information amount calculation unit that calculates an unknownness degree corresponding to the proportion of unknown information based on the amount of information regarding the target of the behavior to which the asset is commanded and the amount of unknown information therein, and a dispersion degree calculation unit that calculates a dispersion degree based on the unknownness degree.

9. In the distributed behavior control device described in claim 1, the control allocation unit allocates control to autonomously distribute and act on the one or more assets based on the degree of coordination linked to the behavior of the one or more assets and the degree of distribution calculated by the degree of distribution calculation unit.

10. 2. The distributed behavior control device according to claim 1, wherein when there are a plurality of assets to behave in a distributed manner in the area with a high degree of unknownness, the assets are autonomously distributed and act without being coordinated at the destination of the action.

11. A distributed behavior control system, comprising: one or more assets; a device that communicates with the one or more assets and controls the behavior of each of the assets; Equipped with The device sets an environment based on the environmental information, which is information specifying information on behavioral goals, asset information, and asset behavior areas, calculates an amount of unknown information from information regarding the behavioral goals to which assets are commanded in the area in which the asset's behavior is to be performed, calculates a degree of dispersion that determines the degree to which assets in an asset group containing multiple assets will act autonomously and in a distributed manner based on the calculated degree of dispersion, determines a control allocation for autonomously distributing the actions of the assets in the asset group based on the calculated degree of dispersion, and outputs the control allocation for each asset.

12. A distributed behavior control method for a device that controls behavior of one or more assets, comprising: an environment setting step in which the environment information is information on behavioral goals, asset information, or information specifying a behavioral area of ​​an asset, and the environment is set based on the environment information; an unknown information amount calculation step of calculating an unknown information amount from information regarding a goal of an action commanded to the asset in an area where the asset's action set in the environment setting step is performed; a dispersion degree calculation step of calculating a dispersion degree that determines the degree of autonomous dispersion of assets within an asset group including a plurality of assets based on the amount of unknown information calculated in the unknown information amount calculation step; a control allocation step of determining control allocation for autonomously distributing and acting on the assets within the asset group based on the degree of diversification calculated in the degree of diversification calculation step; a control output step of outputting the allocation of control for each asset determined by the control allocation step; A decentralized behavior control method comprising:

Images