JP6670351B2 - Map updating device and map updating method - Google Patents

Description

  The present invention relates to a map updating device and the like for updating a map indicating a movable area of a moving object in a target area.

  2. Description of the Related Art A technique is known in which a moving object such as an unmanned traveling vehicle or an unmanned flying object (for example, a drone) is autonomously moved in a target area and a map indicating a movable area of the moving object in the target area is generated. As described in Patent Document 1 below, this technique is called SLAM (Simultaneous Localization and Mapping).

JP 2018-084699 A

  Here, when the target area is a facility such as a factory or a warehouse, the arrangement of obstacles that hinder movement of the moving object may change. If the arrangement of the obstacles changes, the movable area of the moving body also changes, so that the map needs to be updated.

  However, in SLAM, since maps can be generated and updated only within the range in which the moving object has moved, conventionally, unless the moving object moves each part of the facility every time the position of an obstacle changes, the entire facility is not moved. Map could not be updated.

  An object of one embodiment of the present invention is to provide a map updating device or the like that can update a map so as to match a current state without moving a moving object to all movable areas.

  In order to solve the above-described problem, a map updating device according to one embodiment of the present invention provides a map updating device that moves a moving object in a target area in which a pattern of arrangement of one or more obstacles that hinder movement of the moving object changes. A map updating device for updating a map indicating a movable area, wherein the pattern is limited to a predetermined type, and among the patterns of a predetermined type, the pattern corresponding to the current arrangement of the obstacle is specified. A pattern specifying unit; and a map updating unit that updates the movable area in the map according to the pattern specified by the pattern specifying unit.

  Further, in order to solve the above-described problem, the map updating method according to one embodiment of the present invention provides the map updating method according to one aspect of the present invention, wherein the moving pattern in the target area in which the arrangement pattern of one or more obstacles that hinder the movement of the moving object changes. A map updating method by a map updating device for updating a map indicating a movable area of a body, wherein the pattern is limited to a predetermined type, and corresponds to a current arrangement of the obstacle among the predetermined type of the pattern. And a map updating step of updating the movable area in the map according to the pattern specified in the pattern specifying step.

  According to one embodiment of the present invention, the map can be updated so as to match the current state without moving the moving object to all of the movable areas.

It is a block diagram showing an example of the important section composition of the map updating device concerning Embodiment 1 of the present invention. It is a figure showing the outline of the map update by the above-mentioned map updating device. FIG. 3 is a diagram illustrating an example of patterns corresponding to the arrangement of obstacles and an example of a photographing range for photographing an image in which the features of those patterns appear. It is a flowchart which shows an example of the process which the said map update device performs. It is a figure showing the outline of the map update by the map updating device concerning Embodiment 2 of the present invention. FIG. 9 is a diagram illustrating obstacle information for specifying a pattern of the arrangement of obstacles. It is a block diagram showing an example of the important section composition of the map updating device concerning Embodiment 3 of the present invention.

[Embodiment 1]
(Device overview)
An outline of the map updating device 1 of the present embodiment will be described with reference to FIG. FIG. 2 is a diagram showing an outline of map updating by the map updating device 1. The map updating apparatus 1 is an apparatus that updates a map indicating a movable area of a moving body in a facility (target area) in which a pattern of arrangement of one or more obstacles that hinder movement of the moving body changes. Obstacle arrangement patterns are limited to predetermined types. In the example shown in FIG. 2, the facility is a distribution warehouse, the moving object is an unmanned transport vehicle (unmanned forklift), and the obstacle is luggage loaded on a pallet. However, the moving object is not limited to the unmanned carrier, and may be an unmanned flying object such as a drone. Further, the specific number of the above-described predetermined types is not limited to the present embodiment, but includes, for example, several to several tens of thousands types.

  In the example of FIG. 2, an image obtained by photographing a part of the facility using a fixed point camera or the like is transmitted to the map updating device 1. Fixed point cameras and the like are installed at various places in the facility. The map updating device 1 acquires an image captured by the fixed point camera or the like, and updates a map in the facility based on the image. Then, the map updating device 1 transmits the updated map to the automatic guided vehicle operating in the facility. The automatic guided vehicle moves within the facility with reference to the map received from the map updating device 1.

(Main configuration of map updating device)
The main configuration of the map updating device 1 will be described with reference to FIG. FIG. 1 is a block diagram illustrating an example of a main configuration of the map updating device 1. As shown in the figure, the map updating device 1 includes a control unit 10 that controls and controls each unit of the map updating device 1, and a storage unit 20 that stores various data used by the map updating device 1. Further, the map updating apparatus 1 includes a communication unit 30 for communicating with an external device (for example, the automatic guided vehicle in FIG. 2). The map updated by the map updating device 1 is transmitted to the automatic guided vehicle via the communication unit 30. Further, the control unit 10 includes an information acquisition unit 101, a pattern identification unit 102, a map update unit 103, and a map transmission unit 104.

  The information obtaining unit 101 obtains information for specifying an arrangement pattern of obstacles in the facility, and outputs the information to the pattern specifying unit 102. More specifically, the information acquisition unit 101 acquires an image of a part of the facility, for example, from an imaging device as illustrated in FIG. 2 and outputs the acquired image to the pattern identification unit 102.

  The pattern specifying unit 102 specifies a pattern corresponding to the current arrangement of the obstacles among predetermined types of patterns based on the information output from the information obtaining unit 101. Specifically, the pattern identification unit 102 of the present embodiment uses the learned model generated by machine learning to extract a pattern corresponding to the current arrangement of obstacles from the image acquired by the information acquisition unit 101. Identify. Although the details will be described later, the learned model uses a mechanical relationship between an image of a part of a facility in which the feature of the pattern of the arrangement of obstacles is captured and a predetermined type of pattern at the time of the capturing. It is generated by learning.

  According to the above configuration, the pattern of the arrangement of the obstacles in the facility can be specified as long as an image of a part of the current facility is acquired. Therefore, it is possible to easily specify how the obstacle is arranged in the facility without causing the automatic guided vehicle to travel in the entire facility. Then, if it is possible to specify how the obstacle is arranged in the facility, the map updating unit 103 can update the map.

  The map updating unit 103 updates the movable area on the map according to the pattern specified by the pattern specifying unit 102. Specifically, the map update unit 103 sets, as a movable area, an area in the facility where no obstacle is present in the pattern. Further, an area where an obstacle is present in the pattern is defined as an immovable area where a moving object such as the automatic guided vehicle cannot move. Thus, the map is updated so as to match the current obstacle arrangement pattern. Then, the map transmitting unit 104 acquires the map updated by the map updating unit 103, and transmits the map to the automatic guided vehicle via the communication unit 30.

(About trained model and input data)
A learned model used by the pattern identification unit 102 of this embodiment to identify a pattern corresponding to the current arrangement of obstacles will be described. In the following, an example of generating a learned model of a neural network (NN: Neural Network) will be described, but a trained model capable of specifying a pattern corresponding to the current arrangement of obstacles can be constructed. If so, other algorithms can be applied. When the NN trained model is used, it is preferable to use a multilayer NN trained model generated as described below and expected to have high specific accuracy. When an image is used as teacher data as described below, it is preferable to use a CNN (Convolutional Neural Network) suitable for processing the image.

  For generation of the learned model, data in which an image of a part of a facility where a pattern of obstacle placement is known is captured as correct data can be used as teacher data. By preparing a large number of such teacher data for each of predetermined types of patterns that can occur in the facility and performing NN or CNN machine learning, it is possible to generate a learned model used by the pattern identification unit 102. it can. It should be noted that the teacher data only needs to be information that shows the characteristics of the pattern, and is not limited to an image of a part of the facility.

  The correspondence between some of the images in the facility and the above-described patterns will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of patterns corresponding to the arrangement of obstacles and an example of an image capturing range for capturing an image in which the features of those patterns appear. In FIG. 3, the area where the obstacle is arranged is indicated by surrounding it with a broken line. FIG. 3 shows a case where the above-mentioned predetermined types are three types.

  Specifically, in the example shown in FIG. 3, there are three types of patterns A, B, and C. In pattern A, the area where the obstacle is arranged has the shape of the letter “E”. In the pattern B, the area where the obstacle is arranged has a shape obtained by removing the central horizontal bar from the letter “E”. In the pattern C, the area where the obstacle is arranged has a shape obtained by removing the vertical bar from the alphabet “E”.

  In the example shown in FIG. 3, the features of each pattern are shown in the imaging range indicated by a circle. That is, the range of the pattern A includes a part of the region T where the obstacles are arranged vertically and a part of the region Y where the obstacles are arranged horizontally, while the range of the pattern B includes Includes only a part of the area T, and the range of the pattern C includes only a part of the area Y.

  In such a facility, a learned model generated by machine learning using the following three types of teacher data may be used. Note that the following images (1) to (3) may be images captured at an angle overlooking the imaging range by, for example, a fixed-point camera fixed to a ceiling, a wall, or the like.

  (1) Teacher data that associates pattern A as correct data with an image obtained by shooting the shooting range when the obstacle is arranged in pattern A.

(2) Teacher data in which pattern B is associated as correct data with respect to an image obtained by photographing the photographing range when the arrangement of obstacles is pattern B. (3) When the arrangement of obstacles is pattern C Teacher data in which a pattern C is associated as the correct answer data with respect to the image of the shooting range.

  If an image photographed in the photographing range is input to the learned model when the arrangement of the obstacle is unknown, the probability corresponding to each of the patterns A to C is output from the learned model. Then, the pattern specifying unit 102 specifies a pattern corresponding to the current obstacle arrangement based on the output. For example, the pattern specifying unit 102 may specify the pattern having the highest probability as the pattern corresponding to the current obstacle arrangement.

  In the following, for simplicity of explanation, an example will be described in which the arrangement of the current obstacle corresponds to one of a plurality of types of patterns from an image captured in one imaging range as described above. The shooting range may be different for each pattern. In this case, a learned model is generated for each shooting range. Then, the information acquisition unit 101 acquires images captured in each imaging range, and the pattern identification unit 102 identifies a pattern using a learned model corresponding to the imaging range.

  For example, it is assumed that the photographing ranges Aa to Ac for the patterns A to C in FIG. 3 are set, and the learned models Ma to Mc corresponding to the patterns A to C have been generated in advance. In this case, the pattern specifying unit 102 inputs an image captured in the capturing range Aa to the learned model Ma, and if the output probability value is equal to or greater than the threshold, the current obstacle arrangement is the pattern A. And specify. On the other hand, if the value of the probability is less than the threshold value, the image captured in the capturing range Ab is input to the learned model Mb, and the same processing as described above is performed, and the current obstacle arrangement is the pattern B. It is determined whether or not. If the current arrangement of the obstacle is not the pattern B, the pattern specifying unit 102 inputs the image photographed in the photographing range Ac to the learned model Mc, performs the same processing as described above, and performs the same processing as described above. It is determined whether or not the arrangement of the objects is the pattern C.

  In the following, an example will be described in which the image acquired by the information acquiring unit 101 is used as input data of a learned model as it is, but an image obtained by performing a predetermined process on the image may be used as input data. The predetermined process may be any process that can reduce information not related to a feature point without losing the feature point in the image. For example, if the image is a color image, a grayscale image may be used as the input data.

  Regardless of whether the acquired image is used as it is or subjected to processing, the image used for the input data was taken under the same shooting conditions as the image used for the teacher data for machine learning of the trained model. Preferably, there is. The shooting conditions include, for example, a positional relationship between the shooting device and the shooting range, ambient brightness, and the like. In order to make the photographing conditions uniform, it is preferable to generate teacher data from an image photographed by a fixed-point camera fixed at a predetermined position with respect to the photographing range, and use the image photographed by the same fixed-point camera as input data.

(Processing flow)
The flow of processing executed by the map updating device 1 will be described with reference to FIG. FIG. 4 is a flowchart illustrating an example of a process (a map updating method) executed by the map updating device 1. Strictly speaking, the process from S2 onward in FIG. 4 is a process of the map updating device 1, and the process S1 is a process executed by a photographing device such as a fixed point camera installed at a specific place of the facility.

  In S1, the photographing device photographs a part of the facility. Then, the photographing device transmits the photographed image to the map updating device 1. In S2, the information acquisition unit 101 acquires the image transmitted in S1, and outputs the acquired image to the pattern identification unit 102. The pattern identification unit 102 identifies the current obstacle arrangement pattern using the learned model described above. (Pattern identification step).

  In S3, the pattern specifying unit 102 compares the previously specified pattern with the currently specified pattern to determine whether the pattern has been changed. If it is determined that the pattern has been changed (YES in S2), the process proceeds to S4. On the other hand, when it is determined that the pattern has not been changed (NO in S2), the process returns to S1.

  In S4, the map updating unit 103 updates the map according to the pattern specified in S2 (map updating step). Then, in S5, the map transmitting unit 104 transmits the map updated in S4 to the automatic guided vehicle. Thereafter, the process returns to S1. Thus, the automatic guided vehicle can always travel in the facility using a map that matches the latest obstacle arrangement pattern.

(Summary of Embodiment 1)
As described above, the map updating apparatus 1 updates the map indicating the movable area of the automatic guided vehicle in a facility where the arrangement pattern of one or more obstacles that hinder the movement of the automatic guided vehicle changes. The pattern is limited to a predetermined type. The pattern specifying unit 102 specifies a pattern corresponding to the current arrangement of obstacles from among the predetermined types of patterns. The map updating unit 103 updates the movable area on the map according to the pattern specified by the pattern specifying unit 102. As described above, according to the map updating apparatus 1, it is specified which of the patterns the current obstacle arrangement corresponds to, and the map is updated according to the result. The map can be updated to match the current situation without moving the carrier.

  Note that, instead of using the learned model, the pattern identifying unit 102 uses an image analysis method such as pattern matching to determine the current obstacle arrangement in the facility from an image of a part of the facility. It may be specified.

[Embodiment 2]
Embodiment 2 of the present invention will be described below. For convenience of description, members having the same functions as those described in the above embodiment are denoted by the same reference numerals, and description thereof will not be repeated. This is the same in the third and subsequent embodiments.

  An outline of the map updating device 1 of the present embodiment will be described with reference to FIG. FIG. 5 is a diagram showing an outline of map updating by the map updating device 1 of the present embodiment. In the example of FIG. 5, obstacle information indicating the arrangement of obstacles in the facility is transmitted to the map updating device 1 from a plurality of automatic guided vehicles (moving objects). The map updating device 1 acquires the transmitted obstacle information, and updates the map in the facility based on the obstacle information. The processing after updating the map is the same as in the first embodiment.

  The main configuration of the map updating device 1 is the same as that of FIG. However, the information acquiring unit 101 of the present embodiment acquires, from each of a plurality of automatic guided vehicles moving in the facility, obstacle information indicating the location of the obstacle in the facility. Further, the pattern specifying unit 102 of the present embodiment specifies a pattern corresponding to the current arrangement of obstacles, using the obstacle information acquired from each of the plurality of automatic guided vehicles.

(Pattern identification method)
A method of specifying a pattern from obstacle information will be described with reference to FIG. FIG. 6 is a diagram illustrating obstacle information for specifying a pattern of the arrangement of obstacles. In FIG. 6, a region where an obstacle is arranged is surrounded by a broken line as in FIG. In FIG. 6, the three automatic guided vehicles are indicated by symbols a, b, and c. Further, each automatic guided vehicle is provided with a photographing device. In FIG. 6, the photographing directions of these photographing devices are indicated by white arrows. As described above, the images captured at various places in the facility, the positions where the images are captured, and the capturing direction can be used as obstacle information.

  For example, in FIG. 6, when an obstacle is arranged in the pattern A, an image captured by the automatic guided vehicle a from the position where the area T falls within the imaging range toward the area T has a state in which the obstacle exists. The region T of FIG. The information acquisition unit 101 receives the image captured by the automatic guided vehicle a, specifies the position and the shooting direction of the automatic guided vehicle, and outputs the information to the pattern specifying unit 102 as obstacle information. Then, the pattern specifying unit 102 specifies that there is an obstacle in the region T by analyzing the image. It should be noted that the automatic guided vehicle has a function of traveling by SLAM, and since the position and the photographing direction (vehicle direction) of the automatic guided vehicle can be specified by this function, the information acquisition unit 101 communicates with the automatic guided vehicle. Thus, the position and the photographing direction of the automatic guided vehicle can be specified.

  Similarly, the information acquisition unit 101 receives the image captured by the automatic guided vehicle b, identifies the position and the imaging direction of the automated guided vehicle, and outputs the information as obstacle information to the pattern identification unit 102. I do. Then, the pattern specifying unit 102 specifies that there is an obstacle in the region Y by analyzing the image. Note that whether or not an obstacle is present in the captured image can also be specified using a learned model generated by machine learning using the image in which the obstacle is present as teacher data.

  In the example of FIG. 6, only the pattern A has an obstacle in both the regions T and Y. Therefore, the pattern specifying unit 102 determines that the current arrangement of the obstacle is the pattern A based on the above specification result. Identify that. As described above, the information acquiring unit 101 acquires the images photographed at various places in the facility, and the photographing position and the photographing direction as the obstacle information, so that the pattern identifying unit 102 can determine the current arrangement of the obstacles. Which pattern is applicable can be specified. Needless to say, in specifying the pattern A, an image taken from the automatic guided vehicle c may be used.

  The same applies to the pattern B. The pattern identification unit 102 determines that an obstacle exists in the region T from an image or the like captured from the automatic guided vehicle a and an obstacle exists in the region Y from an image or the like captured from the automatic guided vehicle c. By specifying that there is no obstacle, the current obstacle arrangement is specified to be pattern B. The same applies to the pattern C. The pattern identification unit 102 determines that there is an obstacle in the area Y from an image or the like taken from the automatic guided vehicle b, and that the area T is in the area T from an image or the like taken from the automatic guided vehicle c. By specifying that there is no obstacle, it is specified that the current obstacle arrangement is pattern C.

(Summary of Embodiment 2)
As described above, the map updating apparatus 1 includes the information acquisition unit 101 that acquires, from each of a plurality of automatic guided vehicles moving within a facility, obstacle information indicating an arrangement of obstacles within the facility. The pattern specifying unit 102 specifies a pattern corresponding to the current arrangement of the obstacles using the obstacle information acquired from each of the plurality of automatic guided vehicles. According to this configuration, the pattern corresponding to the current arrangement of the obstacles is specified using the obstacle information acquired from each of the plurality of automatic guided vehicles, so that the automatic guided vehicle is moved to the entire movable area. It is possible to update the map accurately in a short time without any problem.

  In addition, the obstacle information acquired from each automatic guided vehicle only needs to indicate the arrangement of obstacles in a part of the facility, and is not limited to the above example. For example, the information obtaining unit 101 obtains, as obstacle information, a map updated by the automatic guided vehicle by the SLAM function (a map of a portion updated by the most recent traveling of the map used by the automatic guided vehicle). May be. In this case, the pattern specifying unit 102 specifies the presence or absence of an obstacle peculiar to each pattern (in the example of FIG. 6, the presence or absence of an obstacle in the regions T and Y) from the acquired map in the same manner as described above. Based on the identification result, a pattern corresponding to the current location of the obstacle can be identified.

[Embodiment 3]
Embodiment 3 of the present invention will be described below. A configuration of a main part of the map updating device 1 according to the present embodiment will be described with reference to FIG. FIG. 7 is a block diagram illustrating an example of a main configuration of the map updating device 1. As shown in FIG. 7, the map updating apparatus 1 differs from the map updating apparatus 1 of FIG. 1 in that an event detecting unit 105 is provided instead of the information acquiring unit 101.

  By the way, when a transportation vehicle such as a truck for carrying a load arrives at a facility, it is expected that the luggage will be placed in a specific area in the facility. On the other hand, when the transport vehicle that unloads cargo arrives at the facility, it is expected that the load will be unloaded from a specific area in the facility. As described above, a pattern corresponding to the arrangement of obstacles can be estimated based on the arrival of the transport vehicle.

  The event detection unit 105 detects a predetermined event, such as the above event, which triggers a change in the pattern of the arrangement of obstacles. Then, when the event detecting unit 105 detects a predetermined event, the pattern specifying unit 102 specifies that the changed pattern is the pattern of the obstacle arrangement (the pattern after the completion of the trigger). I do.

(Specific examples of patterns)
For example, three types of patterns as shown in FIG. 3 are defined, and when the vehicle carrying the luggage arrives at the facility in a state where the obstacle (for example, luggage) is arranged in the pattern B, the pattern is carried in. It is assumed that it has been decided that the load is to be carried into the area Y. Further, it is assumed that the area T is an area in which luggage carried out of the facility is accumulated.

  In this case, the event detection unit 105 detects that the carry-in vehicle that carries the load arrives at the facility, as a predetermined event that triggers a change in the pattern of the obstacle arrangement from the pattern B to the pattern A. In addition, the event detection unit 105 detects that the unloading vehicle that unloads the cargo arrives at the facility as a predetermined event that triggers the change in the pattern of the obstacle arrangement from the pattern A to the pattern C.

  Note that a method of detecting that a carry-in vehicle or a carry-out vehicle has arrived is not particularly limited. For example, the event detection unit 105 can perform the above-described detection by analyzing an image of a parking space in a facility, or can perform detection using various sensors. In addition, for example, when it is determined that the vehicle is to be carried in or out at a predetermined time, the event detection unit 105 may detect that the incoming vehicle or the outgoing vehicle has arrived at that time. Good. Further, for example, the event detection unit 105 may monitor the state of the region T or Y by, for example, analyzing an image of the region. Then, the event detection unit 105 determines that the operation of moving the obstacle (for example, luggage) in the region T or Y (the operation of carrying the obstacle to the region or the operation of carrying the obstacle out of the region) has been started. Event may be detected.

(Summary of Embodiment 3)
As described above, the map updating apparatus 1 includes the event detection unit 105 that detects a predetermined event that triggers a change in a pattern. When the event detection unit 105 detects the event, the pattern identification unit 102 identifies the pattern after the change as a pattern of the obstacle arrangement. According to this configuration, an event that triggers a change in the pattern of obstacle placement is detected, and a pattern corresponding to the location of the obstacle is identified based on the detected event, so that only a predetermined trigger is detected. The map can be updated quickly.

(Modification)
The configuration for pattern identification described in each of the above embodiments can be used in combination. That is, the map updating apparatus 1 includes an information acquisition unit 101 according to the first embodiment that acquires an image of a part of a facility, an information acquisition unit 101 according to the second embodiment that acquires obstacle information, A plurality of the event detection units 105 may be provided. In this case, the pattern specifying unit 102 may specify a pattern by using the method described in each embodiment, and may output a final pattern specifying result by integrating the specifying results. For example, when the pattern specified based on the image obtained by the information obtaining unit 101 of the first exemplary embodiment matches the pattern specified based on the event specified by the event detecting unit 105 of the third exemplary embodiment, Alternatively, the arrangement of obstacles may be specified as the pattern. With such a configuration, the pattern identification accuracy can be improved.

[About distributed processing]
A part of the processing executed by the map updating apparatus 1 described in each of the above embodiments may be executed by one or a plurality of apparatuses communicatively connected to the map updating apparatus 1. For example, when performing an operation using the learned model, the operation process may be executed by an AI server connected to the map updating apparatus 1 in communication. In this case, the map updating apparatus 1 generates input data to be input to the learned model, transmits the generated data to the AI server, receives output data from the AI server, and specifies a pattern.

[Example of software implementation]
A control block (particularly, each unit included in the control unit 10) of the map updating apparatus 1 may be realized by a logic circuit (hardware) formed on an integrated circuit (IC chip) or the like, or may be realized by software. Good.

  In the latter case, the map updating apparatus 1 includes a computer that executes instructions of a program that is software for realizing each function. This computer includes, for example, one or more processors and a computer-readable recording medium storing the program. Then, in the computer, the object of the present invention is achieved by the processor reading the program from the recording medium and executing the program. As the processor, for example, a CPU (Central Processing Unit) can be used. Examples of the recording medium include “temporary tangible medium” such as a ROM (Read Only Memory), a tape, a disk, a card, a semiconductor memory, and a programmable logic circuit. Further, a RAM (Random Access Memory) for expanding the program may be further provided. Further, the program may be supplied to the computer via an arbitrary transmission medium (a communication network, a broadcast wave, or the like) capable of transmitting the program. Note that one embodiment of the present invention can also be realized in the form of a data signal embedded in a carrier wave, in which the above-described program is embodied by electronic transmission.

  The present invention is not limited to the embodiments described above, and various modifications are possible within the scope shown in the claims, and embodiments obtained by appropriately combining technical means disclosed in different embodiments. Is also included in the technical scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Map update apparatus 101 Information acquisition part 102 Pattern identification part 103 Map update part 105 Event detection part

Claims (4)

A map updating apparatus that updates a map indicating a movable area of the moving body in a target area in which a pattern of arrangement of one or more obstacles that hinder movement of the moving body changes,
The above patterns are limited to a predetermined type,
A pattern specifying unit that specifies the pattern corresponding to the current arrangement of the obstacle, among the predetermined types of patterns,
A map updating unit that updates the movable area in the map according to the pattern specified by the pattern specifying unit .
The pattern identification unit uses a learned model in which the correspondence between the image of a part of the target area in which the feature of the pattern appears and the pattern at the time of the imaging is machine-learned. A pattern corresponding to the current arrangement of the obstacle from an image of the map. A map updating apparatus that updates a map indicating a movable area of the moving body in a target area in which a pattern of arrangement of one or more obstacles that hinder movement of the moving body changes,
The above patterns are limited to a predetermined type,
A pattern specifying unit that specifies the pattern corresponding to the current arrangement of the obstacle, among the predetermined types of patterns,
A map updating unit that updates the movable area in the map according to the pattern specified by the pattern specifying unit;
From each of the plurality of the movable body that moves the target area, and an information acquisition unit for acquiring obstacle information showing the arrangement of the obstacles in the target area,
The pattern specifying unit, a plurality of by using the obstacle information obtained from each of the mobile, maps update device you and identifies a pattern corresponding to the current configuration of the obstacle. A map updating method by a map updating device that updates a map indicating a movable area of the moving body in a target area in which a pattern of arrangement of one or more obstacles that hinder movement of the moving body changes,
The above patterns are limited to a predetermined type,
A pattern specifying step of specifying the pattern corresponding to the current arrangement of the obstacle among the predetermined types of the patterns,
See contains and a map update step of updating the movable area in the map, depending on the particular pattern in the pattern specifying step,
In the pattern identification step, the image of a part of the target area in which the feature of the pattern appears is captured, and the correspondence between the image and the pattern at the time of the capture is machine-learned using a learned model. A map corresponding to a current arrangement of the obstacle from an image obtained by capturing a map. A map updating method by a map updating device that updates a map indicating a movable area of the moving body in a target area in which a pattern of arrangement of one or more obstacles that hinder movement of the moving body changes,
The above patterns are limited to a predetermined type,
A pattern specifying step of specifying the pattern corresponding to the current arrangement of the obstacle among the predetermined types of the patterns,
A map updating step of updating the movable area in the map according to the pattern specified in the pattern specifying step,
From each of the plurality of moving objects moving in the target area, further including a step of obtaining obstacle information indicating the arrangement of the obstacle in the target area,
In the above-mentioned pattern specification step, a map updating method is characterized in that a pattern corresponding to a current arrangement of the obstacle is specified using the obstacle information obtained from each of the plurality of moving objects.

Images