Information processing apparatus, analysis method, and computer-readable recording medium

Abstract

This disclosure includes an information processing apparatus, an analysis method, and a computer-readable recording medium. It efficiently identifies spaces where mobile entities exhibit characteristic behaviors. The information processing apparatus analyzes the world in which the mobile entity moves. Semantic model information represents attribute information for each constituent element constituting the world. Spatial feature quantities representing spaces in the world are defined by item groups that include attribute information of at least multiple constituent elements existing in and around that space. A specific space is the space in the world where the mobile entity exhibits characteristic behaviors. Based on semantic model information related to a first world, the information processing apparatus extracts multiple constituent elements existing in and around a specific space in the first world and defines specific spatial feature quantities representing that specific space. Based on semantic model information related to a second world, the information processing apparatus extracts similar spaces from the second world that have spatial feature quantities similar to those of the specific spatial feature quantities related to the specific space in the first world.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a technology of analyzing a world in which a mobile body moves. BACKGROUND

[0002] Patent Literature 1 discloses a general BIM (Building Information Modeling) model in the field of architectural design. The BIM model is a collection of targets that represent each component of a building. Each target has attribute information. As the attribute information, kinds, positions, sizes, materials, colors, prices, quantities, and the like are exemplified.

[0003] PRIOR ART DOCUMENT

[0004] PATENT LITERATURE

[0005] Patent Literature 1: Japanese Patent Application Publication No. 2020-166327 SUMMARY

[0006] A world (example: building, street) in which a mobile body (example: person, robot) moves is considered. In terms of designing a street or a building that is comfortable for the mobile body, it is useful to know a space in which the mobile body exhibits characteristic behavior. In order to reduce time and cost, a technology capable of efficiently identifying a space in which the mobile body exhibits characteristic behavior is desired.

[0007] One object of the present disclosure is to provide a technology capable of efficiently identifying a space in which a mobile body exhibits characteristic behavior.

[0008] A first viewpoint relates to an information processing apparatus that analyzes a world in which a mobile body moves.

[0009] The information processing apparatus includes one or more processors and one or more storage devices.

[0010] The one or more storage devices store semantic model information that is information of a semantic model of the world, and that indicates attribute information for each constituent element that constitutes the world.

[0011] A space characteristic quantity that characterizes a space in the world is defined by a project group that includes at least attribute information of a plurality of constituent elements that exist in and around the space.

[0012] A specific space is a space in which the mobile body exhibits characteristic behavior in the world.

[0013] A specific space characteristic quantity is a space characteristic quantity that characterizes the specific space.

[0014] The one or more processors extract, from the first world, a plurality of constituent elements existing in and around a specific space in the first world, and define a specific space feature quantity of the specific space by a project group, based on the semantic model information related to the first world.

[0015] Further, the one or more processors extract, from the second world, a similar space having a similar space feature quantity to the specific space feature quantity related to the specific space in the first world, based on the semantic model information related to the second world.

[0016] The second aspect relates to an analysis method of analyzing a world in which a mobile body moves, by a computer.

[0017] The semantic model information is information of a semantic model of a world, and represents attribute information for each constituent element constituting the world.

[0018] A space feature quantity characterizing a space in the world is defined by a project group including at least attribute information of a plurality of constituent elements existing in and around the space.

[0019] The specific space is a space in which a mobile body exhibits a characteristic behavior in the world.

[0020] The specific space feature quantity is a space feature quantity characterizing the specific space.

[0021] The analysis method includes:

[0022] a process of reading out semantic model information related to the first world from one or more storage devices;

[0023] a process of extracting, from the first world, a plurality of constituent elements existing in and around a specific space in the first world, and defining a specific space feature quantity of the specific space by a project group, based on the semantic model information related to the first world;

[0024] a process of reading out semantic model information related to the second world from one or more storage devices; and

[0025] a process of extracting, from the second world, a similar space having a similar space feature quantity to the specific space feature quantity related to the specific space in the first world, based on the semantic model information related to the second world.

[0026] The third aspect relates to an analysis program executed by a computer and causing the computer to execute an analysis process of analyzing a world in which a mobile body moves.

[0027] The semantic model information is information of a semantic model of a world, and represents attribute information for each constituent element constituting the world.

[0028] The spatial feature quantity that characterizes the space in the world is defined by a project group that includes attribute information of a plurality of constituent elements present in and around the space.

[0029] The specific space is a space in which a moving body exhibits a characteristic behavior in the world.

[0030] The specific space feature quantity is a spatial feature quantity that characterizes the specific space.

[0031] The analysis process includes:

[0032] A process of reading out semantic model information related to the first world from one or a plurality of storage devices;

[0033] A process of extracting, from the first world, a plurality of constituent elements present in and around a specific space in the first world, and defining a specific space feature quantity by a project group, based on the semantic model information related to the first world;

[0034] A process of reading out semantic model information related to the second world from one or a plurality of storage devices; and

[0035] A process of extracting, from the second world, a similar space having a similar spatial feature quantity to the specific space feature quantity related to the specific space in the first world, based on the semantic model information related to the second world.

[0036] According to the present disclosure, by utilizing semantic model information, it is possible to define a specific space feature quantity that characterizes a specific space with high precision while suppressing time and cost. Furthermore, by utilizing a specific space feature quantity that characterizes a specific space, it is possible to efficiently extract a similar space having a similar feature to the specific space. That is, it is possible to efficiently recognize a space in which a moving body exhibits a characteristic behavior (specific space, similar space). BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is a conceptual diagram for explaining a world and a specific space that are objects of analysis in an embodiment of the present disclosure.

[0038] Figure 2 is a conceptual diagram for explaining a definition of a space in an embodiment of the present disclosure.

[0039] Figure 3 is a conceptual diagram showing one example of a semantic model.

[0040] Figure 4 is a conceptual diagram for explaining an example of a project group that defines a spatial feature quantity in an embodiment of the present disclosure.

[0041] Figure 5is a conceptual diagram for explaining a constituent element extraction range in the embodiment of the present disclosure.

[0042] Figure 6 is a block diagram showing a structure example of an information processing apparatus related to the embodiment of the present disclosure.

[0043] Figure 7 is a flowchart showing a specific space definition process related to the embodiment of the present disclosure.

[0044] Figure 8 is a conceptual diagram for explaining a similar space extraction process related to the embodiment of the present disclosure.

[0045] Figure 9 is a flowchart showing a similar space extraction process related to the embodiment of the present disclosure.

[0046] Figure 10 is a flowchart showing a noise removal process related to the embodiment of the present disclosure.

[0047] (Symbol explanation)

[0048] 100: information processing apparatus; 110: user interface; 120: processor; 130: storage; 200: analysis program; 300: semantic model information; 400: specific space feature amount information; C: constituent element; M: mobile body; P: space feature amount; PX: specific space feature amount; RNG: constituent element extraction range; SX: specific space; SY: similar space; W: world; W-1: 1st world; W-2: 2nd world. DETAILED DESCRIPTION

[0049] The embodiment of the present disclosure is explained with reference to the drawings.

[0050] 1. Definition of specific space using semantic model

[0051] Figure 1 is a conceptual diagram for explaining a "world W" and a "specific space SX" as an analysis target in the present embodiment. As the world W, a building, a street (example: smart city), a foundation of a facility, and the like are exemplified. Various objects exist in the world W. In particular, a mobile body M moves in the world W. As the mobile body M, a person, a robot, a vehicle, and the like are exemplified. As the robot, a logistics robot, a work robot, and the like are exemplified. As the vehicle, a human-driven vehicle, an automated driving vehicle, a motorcycle, a bicycle, and the like are exemplified.

[0052] The specific space SX is a space of a part of the world W, particularly a space in which the mobile body M exhibits a "characteristic behavior". For example, the specific space SX is a space of a periphery of a position at which a person is likely to bump into a table. As other examples, the specific space SX is a space of a periphery of a position at which a person is likely to slip his feet and lose his balance. As further other examples, the specific space SX is a space in which a person is likely to feel dazzled and put his hands in front of his eyes. As further other examples, the specific space SX is a space of a periphery of a position at which a person is likely to feel horror or nervousness and stop. As further other examples, the specific space SX is a space of a periphery of a position at which a logistics robot is likely to be stacked. As further other examples, the specific space SX is a space in which a vehicle is likely to make a sudden deceleration.

[0053] It is useful to know the specific space SX in which the mobile body M exhibits a characteristic behavior in designing a street, a building, and the like, which are comfortable for the mobile body M. It is particularly useful to grasp what kind of characteristics the specific space SX has. By grasping the characteristics of the specific space SX, it is possible to find a space having the same characteristics, for example. If the specific space SX, a space similar to the specific space SX, can be found in a design stage before actually constructing a building, a street, and the like, it is also possible to improve the design in advance.

[0054] Therefore, in the present embodiment, the definition of "space" is considered. In particular, the definition of "characteristics of a space" is considered. In the following description, the definition of a space and the definition of characteristics of a space are used in the same sense.

[0055] Figure 2 is a conceptual diagram for explaining the definition of a space in the present embodiment. Unlike an object, a space itself does not have a visible characteristic. Therefore, in order to define a space, at least "constituent elements C" present in the space and around the space are used. The constituent elements C refer to elements constituting the world W. Typically, the constituent elements C are stationary objects. For example, the constituent elements C of a building include walls, floors, pillars, ceilings, doors, chairs, tables, windows, and the like. The constituent elements C of the outside of a building include, for example, a terrain, a building, a road, a white line, a sign, a signal, a roadside object, and the like.

[0056] A certain space is characterized by at least the constituent elements C present in and around the space. In other words, information on the constituent elements C present in and around a certain space becomes at least information for characterizing the space. The space to be defined and the constituent elements C for defining the space can be in direct contact or separated. The positional relationship between the constituent elements C, the space, and the positional relationship of the constituent elements C can also become information for characterizing the space. Hereinafter, information for thus characterizing a space is referred to as "space characteristic quantity P". In addition, hereinafter, the space characteristic quantity P for characterizing a specific space SX in which the mobile body M exhibits a characteristic behavior is referred to as "specific space characteristic quantity PX". Defining a certain space means defining a space characteristic quantity P for characterizing the space. For example, defining a certain specific space SX means defining a specific space characteristic quantity PX for characterizing the specific space SX.

[0057] In order to define a space characteristic quantity P related to a certain space with high precision, detailed information on the constituent elements C associated with the space is required. However, it is not necessarily easy to obtain detailed information on the constituent elements C present in the world W. For example, the data amount of point cloud data representing the structure of the entire world W is very large, and in order to correctly obtain information on the kind, size, and the like of each constituent element C, a large amount of time and cost is required. As another example, two-dimensional information of the world W is only graphic information, and it is not possible to automatically determine what each graphic represents, and a large amount of time and cost is spent in determining it.

[0058] It is desirable to define a space characteristic quantity P with high precision while suppressing the time and cost. Therefore, according to the present embodiment, in order to define a space characteristic quantity P related to a space in the world W, a "semantic model" related to the world W is used. The semantic model is a 3D model based on the idea of BIM (Building Information Modeling), CIM (Construction Information Modeling), and the like. However, the semantic model is not only a 3D model, but also includes "attribute information" of each constituent element C constituting the world W. As the attribute information of the constituent elements C, the kind, position, size, material, color, and the like of the constituent elements C are exemplified. Figure 3 An example of BIM data, which is one type of semantic model, is shown.

[0059] By using such a semantic model, it is possible to easily obtain attribute information of a plurality of constituent elements C associated with a space. Furthermore, by using at least the attribute information of the plurality of constituent elements C, it is possible to define a space characteristic quantity P for characterizing the space with high precision. In addition, information other than the attribute information of the constituent elements C can also be additionally used in the definition of the space characteristic quantity P. In general, a "project team" defines a space characteristic quantity P.

[0060] Figure 4 is a conceptual diagram for explaining an example of the item group defining the spatial feature quantity P.

[0061] (1) Attribute information

[0062] The item group defining the spatial feature quantity P includes at least attribute information of each of a plurality of constituent elements C present in and around the space of the definition target. The attribute information of the constituent element C also includes one or more items. As the item constituting the attribute information of the constituent element C, examples are given of a kind, a position, a size, a material, a color, and the like. Examples of definition based on such items are described below.

[0063] <Definition Example 1> The height of a table adjacent to the space is ** m, the color is **, and the material is **.

[0064] <Definition Example 2> The floor under the space is ** color, and the material is **.

[0065] <Definition Example 3> The ceiling height above the space is ** m.

[0066] <Definition Example 4> The window seen from the space is **, and the size of the window frame is ** m x ** m.

[0067] (2) Positional relationship

[0068] The item group defining the spatial feature quantity P can also include a positional relationship of the space of the definition target and at least one constituent element C. The item group defining the spatial feature quantity P can also include a positional relationship between at least two constituent elements C. Examples of definition based on such items are described below.

[0069] <Definition Example 5> The space is ** m away from the entrance door of the room.

[0070] <Definition Example 6> The distance to the ceiling lighting closest to the space is ** m.

[0071] <Definition Example 7> The “table with a height of ** m” is separated from the “wall with a color of ** and a material of **” by ** m.

[0072] (3) Environmental condition

[0073] The item group defining the spatial feature quantity P can also include, as additional information, an environmental condition in which the space of the definition target is placed. The environmental condition also includes one or more items. As the item constituting the environmental condition, examples are given of a temperature, a humidity, an illuminance, a wind direction, a wind force, a season, a time zone, a state of the ground (dry / wet), a distribution density of people around, a sunlight time, a sunlight change, whether or not it is a moving path of a specific moving body M, and the like. Examples of definition based on such items are described below.

[0074] <Definition Example 8> Sunset is incident to the space from the window (particular season / particular time zone)

[0075] <Definition Example 9> The space is blocked by the shadow of the building (particular season / particular time zone)

[0076] <Definition Example 10> The building wind blows into the space

[0077] <Definition Example 11> There is a floor that can have a wetting effect on the insole when it rains

[0078] <Definition Example 12> The space becomes a movement path of a logistics robot

[0079] <Definition Example 13> The space is used by the visually impaired elderly and the like

[0080] Thus, the space feature quantity P that characterizes the space is defined by the project group. The specific space feature quantity PX that characterizes the specific space SX is also defined by the project group. It can be said that the project group that defines the specific space feature quantity PX represents a factor in which the mobile body M exhibits a characteristic behavior.

[0081] Further, in order to define the space feature quantity P related to the space, it is not necessary to use a constituent element C that is irrelevant to the space and is far away from the space. Hereinafter, the range of the constituent element C used for defining the space feature quantity P is referred to as a "constituent element extraction range RNG".

[0082] Figure 5 is a conceptual diagram for explaining the constituent element extraction range RNG. The constituent element extraction range RNG is a range that includes the space of the definition object (example: specific space SX). For example, the constituent element extraction range RNG is a range within a predetermined distance from a representative point in the space of the definition object. The representative point is, for example, the center position of the space of the definition object. As another example, the representative point can be specified by the user. The predetermined distance is, for example, 10 m. From the constituent element extraction range RNG that includes the space of the definition object, a limited number of constituent elements C used for defining the space feature quantity P that characterizes the space are extracted.

[0083] In addition, when extracting the constituent element C used for defining the space feature quantity P, the field of view (viewing field) of the mobile body M, the dead angle can also be considered. The reason is that the constituent element C that cannot be seen from the mobile body M can not affect the behavior of the mobile body M. For example, the constituent element C that exists in the space of the definition object and cannot be seen from the mobile body M (enters the dead angle) can not be used to define the space feature quantity P of the space. In other words, the constituent element C used to define the space feature quantity P of a certain space can be a constituent element C that enters the field of view of the mobile body M existing in the space. By not increasing the constituent element C used for definition more than necessary, the processing load is reduced.

[0084] As explained above, according to the present embodiment, a space in the world W is defined by utilizing a semantic model. Generally, a semantic model is used in the field of architectural design to manage components of a building, but according to the present embodiment, by utilizing such a semantic model, not "objects" but "spaces" are defined. By utilizing a semantic model, it is possible to define spatial feature quantities P that characterize spaces with high precision while suppressing time and cost. That is, it is possible to define spatial feature quantities P efficiently and with high precision.

[0085] In particular, according to the present embodiment, a specific space SX in which the mobile body M exhibits characteristic behavior is defined. By utilizing a semantic model, it is possible to define specific space feature quantities PX that characterize the specific space SX with high precision while suppressing time and cost. That is, it is possible to define specific space feature quantities PX efficiently and with high precision.

[0086] Definition of spaces based on the semantic model explained above is implemented by a computer. Hereinafter, information processing by a computer according to the present embodiment will be explained in detail.

[0087] 2. Information processing apparatus

[0088] Figure 6 is a block diagram showing a configuration example of an information processing apparatus 100 according to the present embodiment. The information processing apparatus 100 is a computer that performs various information processing.

[0089] The information processing apparatus 100 includes a user interface 110. The user interface 110 accepts input of information from a user and provides various information to the user. The user interface 110 includes an input device and an output device. As the input device, a keyboard, a mouse, a touch panel, and the like are exemplified. As the output device, a display device, a touch panel, a speaker, and the like are exemplified. The user interface 110 can also be a GUI (Graphical User Interface).

[0090] The information processing apparatus 100 further includes one or more processors 120 (hereinafter simply referred to as processors 120) and one or more storage devices 130 (hereinafter simply referred to as storage devices 130). The processors 120 perform various processing. For example, the processors 120 include a CPU (Central Processing Unit). The storage devices 130 store various information required in processing performed by the processors 120. As the storage devices 130, a volatile memory, a non-volatile memory, an HDD (Hard Disk Drive), an SSD (Solid State Drive), and the like are exemplified.

[0091] The analysis program 200 is a computer program executed by the processor 120. By executing the analysis program 200 by the processor 120, the function of the information processing apparatus 100 is realized. The analysis program 200 is stored to the storage apparatus 130. Alternatively, the analysis program 200 can also be recorded to a computer-readable recording medium. The analysis program 200 can also be provided via a network.

[0092] The semantic model information 300 is information of a semantic model related to the world W of the analysis target. As described above, the semantic model includes attribute information of each constituent element C constituting the world W. An environmental condition (incidental information) of each location can also be associated with the semantic model. In this case, the semantic model information 300 includes the semantic model and the environmental condition (incidental information) associated therewith. The semantic model information 300 is generated in advance and stored to the storage apparatus 130.

[0093] The specific space feature quantity information 400 is information defining a specific space feature quantity PX characterizing the specific space SX. This specific space feature quantity information 400 is generated by the "specific space definition processing" described below and stored to the storage apparatus 130.

[0094] 3. Specific space definition processing

[0095] Figure 7 is a flowchart showing the specific space definition processing according to the present embodiment.

[0096] In step S100, the processor 120 reads out the semantic model information 300 related to the world W of the analysis target from the storage apparatus 130. Then, the processor 120 displays the semantic model of the world W on the user interface 110 (display apparatus).

[0097] In step S101, the processor 120 acquires the position information of the specific space SX in the world W. For example, the user grasps the position where the mobile body M exhibits a characteristic behavior through a questionnaire survey or the like. Then, the user specifies the specific space SX in the world W using the user interface 110 (input apparatus). Alternatively, the processor 120 can also automatically specify the specific space SX. The processor 120 acquires the position information of the specified specific space SX from the semantic model information 300.

[0098] In step S102, the processor 120 extracts a plurality of constituent elements C present in and around the specific space SX from the semantic model information 300. At this time, the processor 120 extracts the constituent elements C from the extraction range RNG including the constituent elements of the specific space SX (refer to FIG. 2B). Alternatively, the processor 120 can also extract only the constituent elements C entering the field of view of the mobile body M present in the specific space SX. Figure 5 ). Alternatively, the processor 120 can also extract only the constituent elements C entering the field of view of the mobile body M present in the specific space SX.

[0099] In step S103, the processor 120 acquires attribute information of each of the extracted components C based on the semantic model information 300. Then, the processor 120 defines a specific space feature quantity PX that characterizes the specific space SX by using the above-mentioned item group (refer to FIG. 4). The item group includes at least the attribute information of the extracted components C. The item group can also include the positional relationship of the specific space SX and the components C. The item group can also include the positional relationship among the components C. The item group can also include environmental conditions in which the specific space SX is located as incidental information. The item group that defines the specific space feature quantity PX corresponds to a factor that exhibits a characteristic behavior of the mobile body M. Figure 4 ). The item group includes at least the attribute information of the extracted components C. The item group can also include the positional relationship of the specific space SX and the components C. The item group can also include the positional relationship among the components C. The item group can also include environmental conditions in which the specific space SX is located as incidental information. The item group that defines the specific space feature quantity PX corresponds to a factor that exhibits a characteristic behavior of the mobile body M.

[0100] Thus, the processor 120 defines the specific space feature quantity PX that characterizes the specific space SX based on the semantic model information 300. The specific space feature quantity information 400 indicates the defined specific space feature quantity PX. It can also be said that the specific space feature quantity information 400 indicates the characteristics of the specific space SX. The specific space feature quantity information 400 is stored in the storage device 130.

[0101] By using the semantic model information 300, it is possible to define the specific space feature quantity PX that characterizes the specific space SX with high precision while suppressing the time and cost. That is, it is possible to define the specific space feature quantity PX efficiently and with high precision.

[0102] 4. Similar space extraction processing

[0103] The specific space feature quantity information 400 that indicates the characteristics of the specific space SX is useful. For example, by referring to the specific space feature quantity information 400, it is possible to efficiently extract a space that has similar characteristics to the specific space SX. That is, if the characteristics of a certain specific space SX are defined once, it is possible to efficiently extract a space similar to the specific space SX.

[0104] Hereinafter, a space similar to the specific space SX will be referred to as a "similar space SY". The similar space SY is a space that has a space feature quantity P similar to the specific space feature quantity PX that characterizes the specific space SX. Hereinafter, the processing of extracting the similar space SY will be referred to as "similar space extraction processing".

[0105] Figure 8is a conceptual diagram for explaining the similar space extraction processing involved in the present embodiment. There is a specific space SX in a 1st world W-1. The specific space SX in the 1st world W-1 is defined by the specific space definition processing described above. A 2nd world W-2 is a world that is the target of extraction of a similar space SY. The 2nd world W-2 can be the same as the 1st world W-1 or different from the 1st world W-1. From the 2nd world W-2, a similar space SY that is similar to the specific space SX is extracted. In the case where the 1st world W-1 and the 2nd world W-2 are the same, from the same world W (example: a building), a similar space SY that is similar to the specific space SX is extracted.

[0106] Figure 9 is a flowchart showing the similar space extraction processing involved in the present embodiment.

[0107] In step S110, the processor 120 reads out the semantic model information 300-1 related to the 1st world W-1 from the storage device 130.

[0108] In step S111, the processor 120 performs the specific space definition processing that defines the specific space SX in the 1st world W-1, based on the semantic model information 300-1 related to the 1st world W-1. The specific space definition processing is as explained in the section 3 described above. By the specific space definition processing, the specific space feature quantity PX that characterizes the specific space SX in the 1st world W-1 is defined. The specific space feature quantity information 400 that represents the defined specific space feature quantity PX is stored to the storage device 130.

[0109] In step S120, the processor 120 reads out the semantic model information 300-2 related to the 2nd world W-2 from the storage device 130.

[0110] In step S121, the processor 120 extracts (searches) a similar space SY that is similar to the specific space SX from the 2nd world W-2, based on the semantic model information 300-2 related to the 2nd world W-2 and the specific space feature quantity information 400. That is, the processor 120 extracts (searches) a similar space SY that has a space feature quantity P similar to the specific space feature quantity PX related to the specific space SX in the 1st world W-1, from the 2nd world W-2.

[0111] The similarity of the specific space feature quantity PX related to the specific space SX in the 1st world W-1 and the space feature quantity P related to the comparison target space in the 2nd world W-2 means that the degree of coincidence of the item group in which the specific space feature quantity PX is defined and the item group in which the space feature quantity P is defined is above a coincidence threshold. The degree of coincidence is, for example, the number of items that coincide in the item groups.

[0112] For example, assume that a specific space feature amount PX related to a specific space SX in the 1st world W-1 is defined by 7 kinds of items D1, D2, D3, D4, D5, D6, and D7. On the other hand, assume that a space feature amount P related to a comparative object space in the 2nd world W-2 is defined by 5 kinds of items D1, D2, D3, D4, and D5. Further, the number of items that are identical between the specific space feature amount PX related to the specific space SX and the space feature amount P related to the comparative object space is calculated. In the case where the items include numerical values, it can also be determined that two items are identical if the difference between the numerical values included in the two items is less than a predetermined threshold value. In the above example, assume that 5 kinds of items D1, D2, D3, D4, and D5 are identical between the specific space feature amount PX related to the specific space SX and the space feature amount P related to the comparative object space. In this case, the degree of identity of the item group is calculated to be 5 / 7. In the case where the degree of identity threshold value is 70%, it is determined that the degree of identity is higher than the degree of identity threshold value, and the comparative object space is a similar space SY that is similar to the specific space SX. In the case where it is assumed that only 3 kinds of items D1, D2, and D3 are identical, the degree of identity is 3 / 7, and it is determined that the comparative object space is not similar to the specific space SX.

[0113] As explained above, by using the specific space feature amount PX that characterizes the specific space SX, a similar space SY having a similar feature to the specific space SX can be efficiently extracted (searched). That is, if the feature of a certain specific space SX is defined once, a similar space SY similar to the specific space SX can be efficiently extracted (searched). The 2nd world W-2 that is the object of searching for a similar space SY is arbitrary. In various worlds, a space in which a mobile body M exhibits a characteristic behavior (a specific space SX, a similar space SY) can be efficiently recognized. If such a space (a specific space SX, a similar space SY) can be found before a building or a street is actually constructed, it is also possible to improve the design in advance. This is useful in designing a street or a building that is comfortable for a mobile body M.

[0114] 5. Noise removal processing

[0115] There is a possibility that items (noise) having low relevance to the characteristic behavior of the moving body M are included in the item group defining the specific space feature quantity PX. The accuracy of the specific space feature quantity PX defined by the item group including the items having low relevance is low. When the specific space feature quantity PX having low accuracy is used, there is a possibility that a space not similar to the specific space SX is erroneously extracted as the similar space SY. In addition, there is also a possibility that a space that should be extracted as the similar space SY is not correctly extracted. In order to suppress such erroneous extraction and omission, it is preferable to remove the items (noise) having low relevance to the characteristic behavior of the moving body M from the item group. Such processing is referred to as "noise removal processing".

[0116] Figure 10 is a flowchart illustrating the noise removal processing according to the present embodiment.

[0117] In step S130, the processor 120 performs the specific space definition processing described in Section 3 above with respect to a plurality of specific spaces in which the moving body M exhibits the same characteristic behavior (example: a person collides with a table). Thus, the processor 120 acquires a plurality of specific space feature quantities PX related to each of the plurality of specific spaces.

[0118] In step S131, the processor 120 statistically calculates the relevance between the "characteristic behavior of the moving body M" and "each item included in the item group defining each of the plurality of specific space feature quantities PX" on the basis of the plurality of specific space feature quantities PX. When a certain item is commonly included in the plurality of specific space feature quantities PX, it can be said that the relevance between the item and the characteristic behavior of the moving body M is very high. That is, the more the number of specific space feature quantities PX including a certain item commonly, the higher the relevance between the item and the characteristic behavior of the moving body M. Conversely, the less the number of specific space feature quantities PX including a certain item commonly, the lower the relevance between the item and the characteristic behavior of the moving body M.

[0119] In step S132, the processor 120 removes items having relevance less than a certain level from the item group defining the specific space feature quantity PX related to the specific space SX. In other words, the processor 120 selectively uses only items having relevance of a certain level or more to define the specific space feature quantity PX related to the specific space SX again.

[0120] Through the noise removal processing described above, items having low relevance to the characteristic behavior of the moving body M are removed from the specific space feature quantity PX. Thus, the accuracy of the specific space feature quantity PX is improved. As a result, the accuracy of the similar space extraction processing (see Section 4) using the specific space feature quantity PX is also improved. That is, erroneous extraction and omission of the similar space SY are suppressed.

[0121] If the number of items (factors) defining the specific spatial feature amount PX can be reduced, each item can also be analyzed in more detail. For example, the degree of influence of each item can be calculated. There is also a possibility that a "necessary factor" that definitely causes a characteristic behavior of the mobile body M can be determined. If the necessary factor is determined, means for canceling out the necessary factor from each other can also be studied. For example, if insufficient illumination is the necessary factor, countermeasures such as additional setting of lighting can be studied.

Claims

1. An information processing apparatus for analyzing a world in which a mobile body moves, wherein The information processing apparatus includes: one or more processors; and one or more storage devices that store semantic model information that is information of a semantic model of the world, and that indicates attribute information for each constituent element that constitutes the world, a spatial feature quantity that characterizes a space in the world is defined by a project group that includes at least the attribute information of a plurality of constituent elements that exist in and around the space, a specific space is a space in which the mobile body exhibits a characteristic behavior in the world, a specific spatial feature quantity is the spatial feature quantity that characterizes the specific space, the one or more processors are configured to extract, from the first world, a plurality of constituent elements that exist in and around the specific space in the first world, based on the semantic model information related to the first world, define the specific spatial feature quantity by the project group, extract, from the second world, a similar space that has a similar spatial feature quantity to the specific spatial feature quantity related to the specific space in the first world, based on the semantic model information related to the second world, the one or more processors are further configured to acquire a plurality of specific spatial feature quantities related to each of a plurality of specific spaces in which the mobile body exhibits the characteristic behavior, statistically calculate a degree of correlation between the characteristic behavior of the mobile body and each item included in the project group, based on the plurality of specific spatial feature quantities, remove, from the project group that defines the specific spatial feature quantity, an item whose degree of correlation is less than a certain level.

2. The information processing apparatus according to claim 1, wherein the project group that defines the spatial feature quantity that characterizes the space further includes a positional relationship of the space and at least one of the plurality of constituent elements.

3. The information processing apparatus according to claim 1 or 2, wherein the project group that defines the spatial feature quantity that characterizes the space further includes a positional relationship between at least two of the plurality of constituent elements.

4. The information processing apparatus according to claim 1 or 2, wherein the project group that defines the spatial feature quantity that characterizes the space further includes an environmental condition in which the space is located.

5. The information processing apparatus according to claim 1 or 2, wherein the plurality of constituent elements used to define the spatial feature quantity that characterizes the space are constituent elements that exist within a predetermined distance from a representative point in the space.

6. The information processing apparatus according to claim 1 or 2, wherein the plurality of constituent elements used to define the spatial feature quantity that characterizes the space are constituent elements that enter a field of view of the mobile body that exists in the space.

7. The information processing apparatus according to claim 1 or 2, wherein The similarity of the specific space feature quantity related to the specific space in the first world and the space feature quantity related to the space in the second world means that the degree of coincidence of the item group defining the specific space feature quantity and the item group defining the space feature quantity is above a threshold value.

8. An analysis method of analyzing a world in which a mobile body moves, by a computer, wherein semantic model information is information of a semantic model of the world, and attribute information is indicated for each constituent element constituting the world, a space feature quantity characterizing a space in the world is defined by an item group including at least the attribute information of a plurality of constituent elements existing in and around the space, a specific space is a space in which the mobile body exhibits a characteristic behavior in the world, a specific space feature quantity is the space feature quantity characterizing the specific space, the analysis method includes: a process of reading out the semantic model information related to a first world from one or a plurality of storage devices; a process of extracting, from the first world, a plurality of constituent elements existing in and around a specific space in the first world, based on the semantic model information related to the first world, and defining the specific space feature quantity by the item group; a process of reading out the semantic model information related to a second world from the one or the plurality of storage devices; and a process of extracting, from the second world, a similar space having a space feature quantity similar to the specific space feature quantity related to the specific space in the first world, based on the semantic model information related to the second world, the analysis method further includes: a process of acquiring a plurality of specific space feature quantities related to each of a plurality of specific spaces in which the mobile body exhibits the characteristic behavior, a process of statistically calculating a degree of correlation between the characteristic behavior of the mobile body and each item included in the item group, based on the plurality of specific space feature quantities, a process of removing, from the item group defining the specific space feature quantity, an item having a degree of correlation less than a certain level.

9. A computer-readable recording medium recording an analysis program to be executed by a computer, the analysis program causes the computer to execute an analysis process of analyzing a world in which a mobile body moves, wherein semantic model information is information of a semantic model of the world, and attribute information is indicated for each constituent element constituting the world, a space feature quantity characterizing a space in the world is defined by an item group including at least the attribute information of a plurality of constituent elements existing in and around the space, a specific space is a space in which the mobile body exhibits a characteristic behavior in the world, a specific space feature quantity is the space feature quantity characterizing the specific space, the analysis process includes: a process of reading out the semantic model information related to a first world from one or a plurality of storage devices; extracting a plurality of elements existing in and around the specific space in the first world, based on the semantic model information related to the first world, and defining the specific space characteristic quantity by the item group; reading out the semantic model information related to the second world from the one or more storage devices; and extracting a similar space having a similar space characteristic quantity to the specific space characteristic quantity related to the specific space in the first world, from the second world, based on the semantic model information related to the second world, the analysis process further includes: acquiring a plurality of specific space characteristic quantities related to each of a plurality of specific spaces in which the mobile body exhibits the characteristic behavior, statistically calculating a degree of correlation between the characteristic behavior of the mobile body and each item included in the item group, based on the plurality of specific space characteristic quantities, removing an item having a degree of correlation less than a certain level from the item group defining the specific space characteristic quantity.

Images