Map generation device, map generation method, and map generation program

Abstract

To provide a technology that does not depend on the specifications of a moving type machine such as a moving robot which travels using a two-dimensional map, and can generate a two-dimensional map on which obstacle information on a range unmeasurable by a two-dimensional measurement sensor is reflected.SOLUTION: A map generation system 10 includes a space division unit 104 that divides a space in a three-dimensional map into plural divisional spaces in a height direction, an obstacle determination unit 105 that determines an obstacle in each of the divisional spaces, a divisional space compression unit 106 that compresses two-dimensionally in the height direction the obstacle in each of the divisional spaces, and generates plural compressed divisional spaces each including the compressed two-dimensional obstacle, and a two-dimensional map generation unit 108 that combines one or more compressed divisional spaces, which coincide with a width of an arbitrary height, out of the plural compressed divisional spaces, and generates the combined compressed divisional spaces as a two-dimensional map.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a map generating device, a map generating method, and a map generating program, and in particular to a technique for generating a two-dimensional map from a three-dimensional map. [Background technology]

[0002] Patent document 1 discloses a technology for acquiring geometric characteristics of a landmark from a mobile robot, projecting the three-dimensional coordinates of the acquired geometric characteristics onto an image, and calculating two-dimensional coordinates relative to the three-dimensional coordinates (see paragraph

[0020] ).

[0003] Patent Document 2 discloses a technology that generates a two-dimensional map with a large area and low resolution when a mobile robot moves at high speed, based on external environmental information, and generates a two-dimensional map with a small area and high resolution when the mobile robot moves at low speed (see paragraphs

[0018] to

[0023] ). [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-57312 A [Patent Document 2] JP 2020-87248 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, Patent Document 1 is merely a technology for converting coordinate systems, and the generated 2D map depends on the specifications of the mobile robot being used, such as the position and attitude of the mobile robot and its built-in camera, so the 2D map cannot be applied to other mobile robots with different specifications, and even if the 2D map is used, other mobile robots cannot flexibly determine obstacles.

[0006] In addition, cited document 2 is merely a technology for generating two-dimensional maps having different areas and resolutions according to differences in the moving speed of a mobile robot, and the generated two-dimensional map is affected by the moving speed of the mobile robot used, so there are problems similar to those for patent document 1.

[0007] In particular, many of the sensors installed on mobile robots in the city are two-dimensional measurement sensors, which cannot detect obstacles at a different height than the scanning surface of the mobile robot (for example, obstacles above or below the sensor mounting position), so there are cases where the mobile robot becomes stranded due to obstacles that are difficult to measure, making it difficult to move.

[0008] In this regard, Cited Document 2 observes the outside world in three dimensions using a LiDAR sensor and generates an outside world map based on the observation results (see paragraphs

[0032] to

[0036] ). However, there is a problem in that the observation results of the LiDAR sensor often contain noise, making it impossible to generate a two-dimensional map that accurately reflects obstacles.

[0009] The present invention has been made in consideration of the above, and aims to provide a technology that can generate a 2D map for a mobile machine, such as a mobile robot that travels using a 2D map, that is independent of the specifications of the mobile machine and that reflects obstacle information in areas that are difficult to measure with a 2D measurement sensor. [Means for solving the problem]

[0010] A map generation device of one embodiment of the present invention includes a space division unit that divides a space in a three-dimensional map into a plurality of divided spaces in the height direction; an obstacle determination unit that determines obstacles in each divided space; a divided space compression unit that compresses the obstacles in each divided space two-dimensionally in the height direction and generates a plurality of compressed divided spaces including the compressed two-dimensional obstacles; and a two-dimensional map generation unit that combines one or more compressed divided spaces corresponding to a width of any height among the plurality of compressed divided spaces in the height direction and generates the combined compressed divided spaces as a two-dimensional map.

[0011] A map generation method according to one embodiment of the present invention is a map generation method performed by a map generation device, which divides a space in a 3D map into a plurality of divided spaces in the height direction, determines obstacles in each divided space, compresses the obstacles in each divided space into two dimensions in the height direction, generates a plurality of compressed divided spaces containing the compressed two-dimensional obstacles, combines one or more compressed divided spaces corresponding to a width at any height among the plurality of compressed divided spaces in the height direction, and generates the combined compressed divided spaces as a 2D map.

[0012] A map generating program according to one aspect of the present invention causes a computer to function as the map generating device. Effect of the Invention

[0013] According to the present invention, it is possible to provide a technology that can generate a 2D map for a mobile machine such as a mobile robot that travels using a 2D map, which is independent of the specifications of the mobile machine and reflects obstacle information in areas that are difficult to measure with a 2D measurement sensor. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a map generating system. [Diagram 2] FIG. 2 is a sequence diagram showing a processing flow of the map generating system. [Diagram 3] FIG. 3 is a diagram showing an example of the generation of a three-dimensional map. [Figure 4] FIG. 4 is a diagram showing an example of division of a three-dimensional map. [Diagram 5] FIG. 5 is a diagram illustrating an example of compression of a divided space. [Figure 6] FIG. 6 is a diagram showing an example of generating a two-dimensional map. [Figure 7] FIG. 7 is a flowchart showing the process flow of step S109. [Figure 8] FIG. 8 is a diagram showing the positional relationship of processing target points with respect to the xy plane of the compressed divided space. [Figure 9] FIG. 9 is a diagram showing an example of the formation of the determination region. [Figure 10] FIG. 10 is a diagram showing an example of determining an obstacle and noise based on spatial connections. [Figure 11] FIG. 11 is a diagram showing an example of determining an obstacle and noise based on a temporal connection. [Figure 12] FIG. 12 is a diagram illustrating an example of the configuration of a map generating system. [Figure 13] FIG. 13 is a sequence diagram showing the flow of processing in the map generating system. [Figure 14] FIG. 14 is a diagram showing an example of generating a common map. [Figure 15] FIG. 15 is a diagram showing an example of a slice of a common map. [Figure 16] FIG. 16 is a diagram showing an example of compression of the common map. [Figure 17] FIG. 17 is a diagram showing an example of expanding an obstacle area. [Figure 18] FIG. 18 is a diagram showing an example of expanding an obstacle area. [Figure 19] FIG. 19 is a diagram showing an example of compositing a two-dimensional map. [Figure 20] FIG. 20 is a diagram showing an example of a distributed two-dimensional map. [Figure 21] FIG. 21 is a diagram illustrating an example of a hardware configuration of a map generating device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0016] [Summary of the Invention] The present invention is a technology for generating a 2D map from a 3D map and a technology for converting a 3D map into a 2D map. The object of the present invention is to provide a technology for a mobile machine such as a mobile robot that travels using a 2D map, which is independent of the specifications of the mobile machine and can generate a 2D map that accurately reflects obstacle information in a range that is difficult to measure with a 2D measurement sensor.

[0017] In order to achieve the above-mentioned object of the present invention, the present invention divides a three-dimensional space in a three-dimensional map into a plurality of spaces, determines obstacles for each divided space, compresses the determined obstacles into two dimensions, and combines one or more compressed divided spaces containing the compressed two-dimensional obstacles to generate a two-dimensional map.

[0018] According to the present invention, a 2D map is generated using processing that can be applied to any 3D map regardless of the position or attitude of the mobile machine, such as dividing the 3D space, determining obstacles for each divided space, and compressing the obstacles. Therefore, a 2D map can be generated that is not dependent on the specifications of the mobile machine and reflects obstacle information in areas that are difficult to measure with a 2D measurement sensor.

[0019] Furthermore, when determining whether an obstacle exists, the present invention takes into consideration the spatial connection between the divided spaces and the temporal connection between the divided spaces to determine whether an obstacle is noise.

[0020] According to the present invention, whether an obstacle is noise or not is determined by taking into consideration the spatial connection between divided spaces and the temporal connection between divided spaces, so that a two-dimensional map that accurately reflects obstacle information can be generated.

[0021] [First embodiment] [Map generation system configuration] 1 is a diagram showing an example of the configuration of a map generation system according to this embodiment. The map generation system 1 includes a map generation device 10 that generates a 2D map from a 3D map, a sensor device 20 that measures a 3D space, an input terminal 30 for inputting various information, and a mobile robot 40 that travels using the 2D map. These components of the map generation system 1 can communicate with each other via a communication network NT.

[0022] [Map Generation Device] The map generating device 10 includes a sensor information acquisition unit 101, an input information receiving unit 102, a 3D map generating unit 103, a space division unit 104, an obstacle determination unit 105, a divided space compression unit 106, an obstacle determination unit 107, a 2D map generating unit 108, a format conversion unit 109, a map information distribution unit 110, and a data storage unit 111.

[0023] The sensor information acquisition unit 101 has a function of acquiring sensor information required for generating a 3D map of the outside world from the sensor device 20. The sensor information is, for example, point cloud information that captures the external shape of an object existing in the outside world as a point cloud, and image information or video information that captures the outside world.

[0024] The input information receiving unit 102 has a function of receiving input information required to convert a 3D map into a 2D map from the input terminal 30. The input information is, for example, the height of the divided space, the number of space divisions, and the height width of the 3D space suitable for the mobile robot 40.

[0025] The 3D map generating unit 103 has a function of generating a 3D map relating to the external world, using the sensor information acquired by the sensor information acquiring unit 101.

[0026] The space dividing unit 104 has a function of generating a three-dimensional space of the above-mentioned outside world using the three-dimensional map generated by the three-dimensional map generating unit 103, and dividing the generated three-dimensional space into a plurality of divided spaces in the height direction.

[0027] The obstacle determination unit 105 has a function of determining an obstacle in each of the multiple divided spaces divided by the space dividing unit 104.

[0028] The divided space compression unit 106 has a function of compressing the obstacles in each divided space determined by the obstacle determination unit 105 two-dimensionally in the height direction, and generating a plurality of compressed divided spaces including the compressed two-dimensional obstacles.

[0029] The obstacle determination unit 107 has a function of using the multiple compressed divided spaces generated by the divided space compression unit 106 to determine whether or not a two-dimensional obstacle in each compressed divided space is an obstacle, based on whether or not there is a spatial connection in the height direction of the two-dimensional obstacle in each compressed divided space, and whether or not there is a temporal connection of the two-dimensional obstacle in each compressed divided space.

[0030] For example, the obstacle determination unit 107 determines a two-dimensional obstacle that is included only in a compressed divided space in a middle layer among the multiple compressed divided spaces as a false obstacle (e.g., noise). For example, the obstacle determination unit 107 determines a two-dimensional obstacle that is included only in one of the compressed divided spaces at the same position among the multiple compressed divided spaces at different times as a false obstacle.

[0031] The two-dimensional map generation unit 108 has a function of combining, in the height direction, one or more compressed divided spaces that reflect the determination result of the authenticity of the obstacle based on the determination result of the authenticity of the obstacle determined by the obstacle determination unit 107, and generating the combined compressed divided spaces as a two-dimensional map.

[0032] For example, the two-dimensional map generating unit 108 generates a two-dimensional map by combining one or more compressed divided spaces corresponding to a height width suitable for the mobile robot 40. The two-dimensional map generating unit 108 generates a two-dimensional map by combining one or more compressed divided spaces corresponding to an arbitrary height width among the multiple compressed divided spaces.

[0033] The format conversion unit 109 has a function of converting the format of the two-dimensional map generated by the two-dimensional map generation unit 108 into the format of the mobile robot 40 .

[0034] The map information distribution unit 110 has a function of distributing the two-dimensional map after the format conversion by the format conversion unit 109 to the mobile robot 40.

[0035] The data storage unit 111 has a function of storing various pieces of information, various pieces of data, various intermediate products, various processing results, etc. acquired or generated by the above-mentioned functional units of the map generating device 10 in a readable manner.

[0036] [Sensor device] The sensor device 20 is a device that measures and captures images of the outside world or objects present in the outside world. For example, the sensor device 20 is a LiDAR (Light Detection and Ranging) sensor, a RADAR (Radio Detecting and Ranging) sensor, an RGB camera, an infrared camera, or a depth camera.

[0037] [Input device] The input terminal 30 is a terminal for inputting input information required to convert a 3D map into a 2D map. For example, the input terminal 30 is a tablet terminal, a smartphone terminal, or a personal computer. It includes a web browser function, a communication function, an input function, etc.

[0038] [Mobile robot] The mobile robot 40 is an example of a mobile machine that travels using a two-dimensional map. Examples of mobile machines include a transport robot that transports products from shelves in a warehouse to a specified position, and a delivery robot that delivers ordered menu items from the kitchen to seats in a restaurant. Other examples include a cleaning robot that cleans an office and a security robot that guards a building. The mobile machine may be any mobile machine that travels using a two-dimensional map, but may also include drones, automobiles, etc.

[0039] [Positions and positional relationships of the map generating device, the sensor device, and the mobile robot] The map generating device 10 may be disposed in the mobile robot 40 or in the cloud. The sensor device 20 may be built into the mobile robot 40, may be attached externally to the mobile robot 40, may be disposed in the input terminal 30, or may be fixedly installed inside or outside the target environment. The mobile robot 40 may be configured to include the map generating device 10, the sensor device 20, and the input terminal 30.

[0040] [System Operation] Next, an example of the operation of the map generating system 1 will be described with reference to Fig. 2. Fig. 2 is a sequence diagram showing the flow of processing of the map generating system 1. It is assumed that the sensor device 20 is mounted on a mobile robot 40 and measures the target external environment that is the measurement target.

[0041] First, the sensor information acquisition unit 101 of the map generating device 10 acquires sensor information required to generate a 3D map of the target external world from the sensor device 20 (step S101).

[0042] For example, the sensor information acquisition unit 101 acquires point cloud information indicating the external shape of an object in a target space measured by a 3D LiDAR, and image information and video information of the target space captured by an RGB camera.

[0043] Next, the 3D map generating unit 103 of the map generating device 10 generates a 3D map of the target external world using the sensor information (step S102).

[0044] For example, the 3D map generating unit 103 generates a 3D map including object information in the target external world based on point cloud information of objects in the target external world using image information of the target external world as a base. At this time, the 3D map generating unit 103 may utilize a technique such as SLAM (Simultaneous Localization And Mapping) for the measurement results by the 3D LiDAR.

[0045] An example of a generated 3D map of the target external world is shown in Fig. 3. For example, a 3D map MAP(1) is generated that includes a four-legged desk 51, a desk mat 52 placed under the desk 51, a trash can 53 placed next to the desk 51, a cable 54 hanging down from the top of the desk 51 to the desk mat 52, and an unknown object 55.

[0046] The processes after step S102 may be executed within the mobile robot 40 equipped with the map generating device 10, or may be executed within the cloud after uploading the sensor information to the cloud.

[0047] Next, the user inputs the height H of the divided spaces for dividing the three-dimensional space in the three-dimensional map to the input terminal 30 (step S103). At this time, the user may input the number N of space divisions instead of the height H of the divided spaces.

[0048] Next, the input terminal 30 transmits the height H of the divided space or the number N of space divisions input by the user to the map generating device 10 (step S104).

[0049] Next, the input information receiving unit 102 of the map generating device 10 receives the height H of the divided space or the number of space divisions N transmitted from the input terminal 30, and the space dividing unit 104 of the map generating device 10 divides the three-dimensional space in the 3D map by the height H of the divided space or the number of space divisions N received (step S105).

[0050] Here, an example of division of the 3D map MAP(1) shown in Fig. 3 is shown in Fig. 4. For example, the space dividing unit 104 slices the 3D map MAP(1) from the bottom to the top in the height direction of the z axis every H cm, and forms divided spaces R(0) to R(3) each having a height of H cm.

[0051] At this time, the space dividing unit 104 may or may not allow two adjacent divided spaces to overlap each other at a height Mcm. Note that the heights H and M can be changed for each target external environment and each divided space.

[0052] On the other hand, if the number of space divisions N has been transmitted from the input terminal 30, the space dividing unit 104 slices the 3D map MAP(1) from the bottom to the top by the number of space divisions N.

[0053] Next, the obstacle determination unit 105 of the map generating device 10 assigns an initial value of '0' to i of the divided space R(i), and determines whether i of the divided space R(i) matches the number of space divisions N (step S106).

[0054] If the i of the divided space R(i) does not match the number of space divisions N, the obstacle determination unit 105 of the map generating device 10 determines the obstacles in the divided space R(i) (step S107), and the divided space compression unit 106 compresses the determined obstacles and the divided space R(i) into two-dimensional information (step S108).

[0055] Any method can be used to determine whether an obstacle exists. For example, if an object occupies a certain volume or more in a specific space of 5 cm length x 5 cm width x H cm height, the specific space is determined to be an obstacle.

[0056] An example of compression of a divided space having obstacle information compressed into two-dimensional information is shown in Fig. 5. The compressed divided space R'(0) obtained by compressing the divided space R(0) includes two-dimensional obstacles 51' corresponding to the four legs of a desk 51, a two-dimensional obstacle 52' corresponding to a desk mat 52, a two-dimensional obstacle 53' corresponding to a trash can 53, and a two-dimensional obstacle 54' corresponding to a cable 54.

[0057] Thereafter, the obstacle determination unit 105 of the map generating device 10 increments i, and repeatedly executes steps S107 to S until i of the divided space R(i) matches the number N of space divisions.

[0058] As a result, as shown in FIG. 5, compressed divided spaces R'(1) to R'(3) are obtained for the divided spaces R(1) to R(3).

[0059] The compressed divided space R'(1) obtained by compressing the divided space R(1) includes a two-dimensional obstacle 51' corresponding to the four legs of a desk 51, a two-dimensional obstacle 53' corresponding to a trash can 53, a two-dimensional obstacle 54' corresponding to a cable 54, and a two-dimensional obstacle 55' corresponding to an unknown object 55.

[0060] The compressed divided space R'(2) obtained by compressing the divided space R(2) includes a two-dimensional obstacle 51' corresponding to the four legs of the desk 51, a two-dimensional obstacle 54' corresponding to the cable 54, and a two-dimensional obstacle 55' corresponding to the unknown object 55.

[0061] The compressed divided space R'(3) obtained by compressing the divided space R(3) includes a two-dimensional obstacle 51' corresponding to the top of the desk 51 and a two-dimensional obstacle 54' corresponding to the cable 54.

[0062] Next, when i of the divided space R(i) coincides with the number of space divisions N, the obstacle determination unit 107 of the map generating device 10 determines the two-dimensional obstacles in each compressed divided space R'(i) in consideration of the presence or absence of spatial connection in the height direction of the two-dimensional obstacles in each compressed divided space R'(i) and the presence or absence of temporal connection of the two-dimensional obstacles in each compressed divided space R'(i) (step S109). This process will be described later.

[0063] Next, the user inputs, into the input terminal 30, the height A of the bottom surface and the height B of the top surface of the height space that the user wishes to distribute to the mobile robot 40 (step S110).

[0064] The heights A and B correspond to a suitable range of heights for the mobile robot 40. For example, if the height of the mobile robot 40 is 20 cm, the user inputs A=0 cm and B=20 cm. Generally, when the mobile robot 40 runs on the ground, A=0 cm, but when the mobile robot 40 runs in the air, A=0 cm is not necessarily the case.

[0065] Next, the input terminal 30 transmits the height A of the bottom surface and the height B of the top surface input by the user to the map generating device 10 (step S111).

[0066] Next, the two-dimensional map generation unit 108 of the map generation device 10 generates a two-dimensional map corresponding to the mobile robot 40 by aggregating two-dimensional obstacles in the compressed divided space R'(i) included between the height A of the bottom surface and the height B of the top surface (step S112).

[0067] For example, if the height H of the divided space is 3 cm, adjacent divided spaces do not overlap each other vertically (M = 0 cm), and A = 0 cm and B = 3 cm, a two-dimensional map MAP(2) consisting of only the compressed divided space R'(0) is generated as shown in Figure 6. If A = 3 cm and B = 9 cm, a two-dimensional map MAP(2) is generated in which the compressed divided space R'(1) and the compressed divided space R'(2) are superimposed in the height direction.

[0068] Next, the user inputs the specifications and address of the mobile robot 40 into the input terminal 30 (step S113).

[0069] Next, the input terminal 30 transmits the specifications and address of the mobile robot 40 input by the user to the map generating device 10 (step S114).

[0070] Next, the format conversion unit 109 of the map generating device 10 converts the format of the two-dimensional map so that it matches the specifications (communication protocol, data format, etc.) of the mobile robot 40 (step S115).

[0071] Finally, the map information distribution unit 110 of the map generating device 10 distributes the two-dimensional map after the format conversion to the mobile robot 40 (step S116).

[0072] (Modification) The above-described operations and processing order of the map generating system 1 are merely an example. For example, steps S103, S104, S110, S111, S113, and S114 related to the user's operations may be executed before step S101. This can eliminate the waiting time for receiving data from the input terminal 30 in the map generating device 10, and can efficiently and quickly generate a two-dimensional map.

[0073] (Modification) Various information and data input by the user may be stored in advance in the data storage unit 11 of the map generating device 10, and the map generating device 10 may sequentially read and process necessary information from the data storage unit 11 each time the map generating device 10 proceeds with processing. This allows the map generating device 10 to eliminate waiting time for receiving data from the input terminal 30, and enables efficient and rapid generation of a two-dimensional map.

[0074] [Operation of step S109] Next, the operation of step S109 shown in Fig. 2 will be described in detail with reference to Fig. 7. Fig. 7 is a flow chart showing the process flow of step S109.

[0075] Here, the x+1th smallest point on the x axis and the y+1st smallest point on the y axis of the compressed divided space R'(i) consisting of the xy plane are defined as the processing point P(x,y) (where x and y are non-negative integers). In this case, the x coordinate of the processing point P(x,y) is x P(x,y) and the y coordinate is expressed as y P(x,y) For example, as shown in FIG. 8, the x coordinate of P(0,0) is x P(0,0) =0.32, and the y coordinate is y P(0,0) =-0.12.

[0076] In addition, a judgment area is used to determine whether or not there is a spatial or temporal connection between two-dimensional obstacles, and variables a and b (where a and b are positive real numbers) for forming the size of the judgment area are set in advance.

[0077] First, the obstacle determination unit 107 assigns an initial value of '0' to i in the compressed divided space R'(i), and assigns initial values ​​of '0' to x and y of the processing target point P(x,y) in the compressed divided space R'(i) (step S109-1). At this time, the obstacle determination unit 107 may start processing from the minimum value of the processing target point P(x,y) that includes two-dimensional obstacles.

[0078] Next, the obstacle determination unit 107 determines whether or not there is a two-dimensional object at the processing target point P(x,y) (step S109-2). If there is a two-dimensional object at the processing target point P(x,y), the process proceeds to step S109-3, and if there is no two-dimensional object at the processing target point P(x,y), the process skips steps S109-3 and S109-4 and proceeds to step S109-5.

[0079] Next, if there is a two-dimensional object at the processing point P(x, y), the obstacle determination unit 107 determines whether the two-dimensional object exists at the processing point P(x, y) in the two compressed divided spaces R'(i-1) and R'(i+1) adjacent to each other above and below. P(x,y) -a,y P(x,y) -b) and (x P(x,y) +a,y P(x,y) It is then determined whether or not a two-dimensional obstacle is present within a rectangular determination area having a diagonal line at (a+b) (step S109-3).

[0080] An example where an obstacle is present within the judgment area is shown in Figure 9. The judgment area is defined as a rectangular judgment area AR that is slightly larger than the processing target point P(x,y) selected on the compressed divided space R'(i) in the middle part of the figure by the vertical and horizontal dimensions of the variables a and b. The processing target point P(x,y) in the compressed divided space R'(i) includes a two-dimensional obstacle 54' that corresponds to a cable 54, and each of the compressed divided spaces R'(i-1) and R'(i+1) on the top and bottom of the figure also includes a two-dimensional obstacle 54' within the judgment area AR.

[0081] Next, if there is a two-dimensional obstacle within each judgment area AR in two adjacent divided spaces R'(i-1) and R'(i+1) vertically, the obstacle determination unit 107 determines that the two-dimensional obstacle at the processing target point P(x, y) selected in the compressed divided space R'(i) is a connected object with the two-dimensional obstacles determined in the two divided spaces R'(i-1) and R'(i+1) (step S109-4).

[0082] When the compressed divided space R'(i) is the bottom surface, the obstacle determination unit 107 determines whether there is a two-dimensional obstacle or a connected object between the compressed divided space R'(i+1) adjacent to the upper side only. When the compressed divided space R'(i) is the top surface, the obstacle determination unit 107 determines whether there is a two-dimensional obstacle or a connected object between the compressed divided space R'(i-1) adjacent to the lower side only.

[0083] On the other hand, if there is no two-dimensional obstacle within the respective determination regions AR of the two vertically adjacent divided spaces R'(i-1) and R'(i+1), the obstacle determination unit 107 does not execute step S109-4.

[0084] Next, the obstacle determination unit 107 determines the current x P(x,y) It is determined whether the current x value is greater than the maximum value X in the x-axis direction in the compressed divided space R'(i) (step S109-5). P(x,y) If the value is not greater than the maximum value X, the obstacle determination unit 107 increments the x value and returns to step S109-2. By repeating steps S109-2 to S109-5, all processing target points P in the 0th row in the y-axis direction are scanned.

[0085] On the other hand, the current x P(x,y) If the value is greater than the maximum value X, the obstacle determination unit 107 determines the current y P(x,y) It is determined whether the current y value is greater than the maximum value Y in the y-axis direction in the compressed divided space R'(i) (step S109-6). P(x,y) If the value is not greater than the maximum value Y, obstacle determination unit 107 increments the y value and returns to step S109-2. By repeating steps S109-2 to S109-6, all processing points P in all rows are scanned in the y-axis direction.

[0086] On the other hand, the current y P(x,y)If the value is greater than the maximum value Y, the obstacle determination unit 107 determines whether the current i matches the space division number N minus 1 (step S109-7). If i is smaller than the space division number N minus 1, the obstacle determination unit 107 increments i and returns to step S109-2. By repeating steps S109-2 to S109-7, all processing target points P in the compressed divided space R'(i) are scanned.

[0087] Next, the obstacle determination unit 107 determines the two-dimensional obstacles in each compressed divided space R'(i) in consideration of the presence or absence of spatial connection in the height direction of the two-dimensional obstacles in each compressed divided space R'(i) (step S109-8). This process will be described in detail below.

[0088] Through the processing up to step S109-7, it has been determined whether the two-dimensional obstacle in the compressed divided space R'(i) is a connected object that is connected to a two-dimensional obstacle in the upper compressed divided space R'(i+1), a two-dimensional obstacle in the lower compressed divided space R'(i-1), or both of these two-dimensional obstacles, or whether it is an unconnected object that is not connected to them.

[0089] However, even if they are connected objects, a connected object floating in the air in a three-dimensional space is physically impossible on the ground surface where gravity exists. Therefore, the obstacle determination unit 107 deletes, as noise, connected objects that are not connected to the bottom or top surface of the multiple connected objects, that is, connected objects that are included only in the middle compressed divided spaces among the multiple compressed divided spaces that are continuous in the height direction.

[0090] For example, as shown in Fig. 10, the obstacle determination unit 107 removes, as noise, a continuous object (two-dimensional obstacle 55' equivalent to unknown object 55) contained only in two compressed divided spaces R'(1) and R'(2) located in the middle layer. On the other hand, the obstacle determination unit 107 determines, as an obstacle, a continuous object (two-dimensional obstacle 54' equivalent to cable 54) contained in all compressed divided spaces R'(0) to R'(3).

[0091] In addition, even if a continuous object is floating in the air, the obstacle determining unit 107 may determine that the continuous object is an obstacle if it is connected from the bottom or top surface.

[0092] The same applies to an unconnected object that is included in only one compressed divided space. The obstacle determination unit 107 removes, as noise, an unconnected object that is included in only one compressed divided space in the middle layer among multiple compressed divided spaces that are continuous in the height direction.

[0093] Finally, the obstacle determination unit 107 determines the two-dimensional obstacles in each compressed divided space R'(i) in consideration of the presence or absence of a temporal connection between the two-dimensional obstacles in the compressed divided space R'(i) (step S109-9). This process will be described in detail below.

[0094] Even if a connected object is determined to be an obstacle, it is impossible for the connected object to disappear in a short time in the present day when teleportation is generally impossible. Therefore, the obstacle determination unit 107 deletes, as noise, a connected object that does not exist at the same position across multiple consecutive different times, that is, a connected object that is included in only one of multiple compressed divided spaces at the same position among multiple compressed divided spaces at different times.

[0095] For example, as shown in FIG. 11, in two compressed divided spaces R'(1) and R'(2), if the compressed divided space at time t contains a continuous object of a two-dimensional obstacle 55' equivalent to an unknown object 55, but the compressed divided space at time t+1 does not contain the continuous object at the same position, the obstacle determination unit 107 deletes the continuous object as noise.

[0096] In addition, the obstacle determination unit 107 may determine as an obstacle a continuous object that moves at a specific period, for example, a continuous object that is included in the compressed divided space at time t, is not included in the same position in the compressed divided space at time t+1, but is again included in the same position in the compressed divided space at time t+2.

[0097] The same applies to unconnected objects that are included in only one compressed divided space. The obstacle determination unit 107 also removes unconnected objects that are included in only one compressed divided space in the middle layer among the multiple compressed divided spaces, as noise, if the unconnected objects are non-contiguous objects that do not exist across time.

[0098] (Modification) The method of aggregating two-dimensional obstacles described in step S112 above is an example. For example, two or more compressed divided spaces R'(i) among the multiple compressed divided spaces R'(i) may be arbitrarily overlapped, such as overlapping the compressed divided space R'(1) and the compressed divided space R'(3).

[0099] (Modification) The rectangular judgment area AR described in step S109-3 above is just an example. For example, instead of a rectangle, a circle with a radius a may be used as the judgment area AR, or a complex area formed by combining multiple figures of the same or different sizes and shapes may be used as the judgment area AR.

[0100] (Modification) The operation and processing order of step S109 described above are merely an example. For example, step S109-9 may be executed before step S109-8, or step S109-8 and step S109-9 may be executed simultaneously.

[0101] (Modification) Only one of step S109-8 and step S109-9 may be executed, or both may be executed in parallel taking into account the processing results of each other, or both may be executed in parallel without taking into account the processing results of each other.

[0102] [Effects of the embodiment] According to this embodiment, the map generating device 10 divides the space in the three-dimensional map into a plurality of divided spaces in the height direction, determines the obstacles in each divided space, compresses the obstacles in each divided space into two dimensions in the height direction, generates a compressed divided space containing the compressed two-dimensional obstacles, determines the authenticity of the two-dimensional obstacles in each compressed divided space as obstacles, combines one or more compressed divided spaces in the height direction that reflect the determined authenticity of the obstacles, and generates the combined compressed divided spaces as a two-dimensional map.

[0103] As described above, in this embodiment, the map generating device 10 generates a 2D map using processes that can be applied to any 3D map regardless of the position or posture of the mobile robot, such as dividing the 3D space, determining obstacles for each divided space, and compressing the obstacles. Therefore, it is possible to generate a 2D map that is not dependent on the specifications of the mobile robot and reflects obstacle information in areas that are difficult to measure with a 2D measurement sensor.

[0104] Furthermore, according to this embodiment, the map generating device 10 determines whether or not a two-dimensional obstacle in each compressed divided space is an obstacle, based on whether or not there is a spatial connection in the height direction of the two-dimensional obstacle in each compressed divided space, and whether or not there is a temporal connection of the two-dimensional obstacle in each compressed divided space.

[0105] As described above, in this embodiment, the map generating device 10 determines whether or not a two-dimensional obstacle in each compressed divided space is an obstacle, based on whether or not there is a spatial connection in the height direction of the two-dimensional obstacle in each compressed divided space, and whether or not there is a temporal connection of the two-dimensional obstacle in each compressed divided space, so that a two-dimensional map that accurately reflects obstacle information can be generated.

[0106] As described above, in this embodiment, it is possible to generate a 2D map that can be used by various mobile machines. It is possible to reflect obstacles that cannot be detected by sensors mounted on machines, such as mobile robots in cities that are equipped with 2D measurement sensors, on the map. It is possible to avoid events that have not been determined as obstacles and have caused problems such as running aground. The present invention according to this embodiment is particularly effective in cases where there is something, such as a cable laid on the floor, that is difficult to determine as an obstacle but is an obstacle to a mobile robot and can cause it to run aground.

[0107] [Second embodiment] In the second embodiment, a 3D map of the entire movement environment of a mobile machine is generated as a common map using 3D measurement sensors such as LiDAR or a camera, and from the common map, a machine map that holds information on obstacles that can be detected within the measurement range of the sensor mounted on the mobile machine, and an additional obstacle map that holds information on obstacles that are in a range that cannot be detected by the sensor mounted on the mobile machine are generated. Both of these maps are then distributed according to the specifications of the mobile machine, enabling the mobile machine to recognize all obstacle information within its movement space and move smoothly.

[0108] For example, a 3D map of the mobile robot's movement space is generated in advance as a common map. Then, a 2D map that can be measured by a sensor mounted on the mobile robot to be used in the service is generated from the common map. At the same time, a 2D map is generated from the common map by extracting obstacle information within the movement space, taking into account the size of the mobile robot, etc. Then, the 2D map with the obstacle information added as metadata of no-travel areas is converted to match the specifications of the mobile robot and distributed.

[0109] [Map generation system configuration] The second embodiment uses a three-dimensional measuring sensor capable of sensing a wide area, and transmits to the mobile machine not only map information that can be recognized by the sensor mounted on the mobile machine, but also information about obstacles that cannot be detected by the sensor, so that the mobile machine can recognize them, thereby enabling the mobile machine to avoid collisions with obstacles that it cannot recognize.

[0110] 12 is a diagram showing an example of the configuration of a map generation system according to this embodiment. The map generation system 1 includes a map generation device 10 that generates a 2D map from a 3D map, a sensor device 20 that measures a 3D space, an input terminal 30 for inputting various information, and a mobile robot 40 that travels using the 2D map. These components of the map generation system can communicate with each other via a communication network NT.

[0111] In the second embodiment, the functions of the map generating device 10 are different from those in the first embodiment. The functions of the map generating device 10 will be described below.

[0112] [Map Generation Device] The map generating device 10 includes a sensor information acquisition unit 101, an input information receiving unit 102, a 3D map generating unit 103, a first 2D map generating unit 201, a second 2D map generating unit 202, a 2D map synthesis unit 203, a format conversion unit 109, a map information distribution unit 110, and a data storage unit 111.

[0113] The sensor information acquisition unit 101 has a function of acquiring, from the sensor device 20, sensor information required for generating a three-dimensional map of the outside world.

[0114] The input information receiving unit 102 has a function of receiving, from the input terminal 30, input information required for converting a 3D map into a 2D map.

[0115] The 3D map generating unit 103 has a function of generating a 3D map related to the above-mentioned outside world by using the sensor information acquired by the sensor information acquiring unit 101. Specifically, the 3D map generating unit 103 generates a 3D map of a 3D space including obstacle information within the movement space of the mobile robot 40.

[0116] In this embodiment, the 3D map generated by the 3D map generation unit 103 is used in common by the first 2D map generation unit 201 and the second 2D map generation unit 202, and therefore the 3D map is called a common map.

[0117] The first 2D map generating unit 201 has a function of generating a 2D map (a map for a mobile robot) that is aligned with the measurement range of the sensor of the mobile robot 40, using a common map that is a 3D map.

[0118] For example, the first 2D map generator 201 generates a 2D map by slicing the common map at the height at which the 2D measurement sensor of the mobile robot 40 is mounted. This 2D map holds information about obstacles included in the measurement range of the sensor of the mobile robot 40.

[0119] The second 2D map generator 202 has a function of generating a 2D map (additional obstacle map) by compressing the 3D space using the common map, which is a 3D map. This 2D map holds information about obstacles that cannot be detected by the mobile robot 40.

[0120] Since the second 2D map generation unit 202 is required to generate a 2D map that contains information about obstacles that cannot be detected by the mobile robot 40, it may generate a 2D map from a 3D space that includes both inside and outside the measurement range of the sensor of the mobile robot 40, or it may generate a 2D map from a 3D space only outside the measurement range of the sensor of the mobile robot 40.

[0121] When generating a two-dimensional map from a three-dimensional space including both inside and outside the measurement range, the second two-dimensional map generation unit 202 compresses, for example, obstacle information within the height space from the feet to the top of the head of the mobile robot 40 onto the two-dimensional map.

[0122] When generating a two-dimensional map from a three-dimensional space only outside the measurement range, the second two-dimensional map generation unit 202 compresses onto the two-dimensional map obstacle information within the height space, for example, from the feet of the mobile robot 40 to the mounting position of the two-dimensional measurement sensor and from the top of the mounting position of the two-dimensional measurement sensor to the top of the mobile robot 40 (i.e., within the movement space of the mobile robot 40 and outside the measurement range of the two-dimensional measurement sensor).

[0123] The second 2D map generating unit 202 also has a function of expanding the obstacle area included in the compressed 2D map. For example, the second 2D map generating unit 202 adds an area based on the size and movement amount of the mobile robot 40 to the periphery of the obstacle area.

[0124] The 2D map synthesis unit 203 has a function of comparing the 2D map (map for a mobile robot) generated by the first 2D map generation unit 201 with the 2D map (additional obstacle map) generated by the second 2D map generation unit 202, and synthesizing the difference with the additional obstacle map onto the map for a mobile robot.

[0125] The format conversion unit 109 has a function of converting the format of the one 2D map synthesized by the 2D map synthesis unit 203 and the format of the two 2D maps (the map for the mobile robot and the additional obstacle map) before synthesis into the format of the mobile robot 40.

[0126] The map information distribution unit 110 has a function of distributing the 2D map after the format conversion by the format conversion unit 109 to the mobile robot 40. The map information distribution unit 110 may distribute only one of the combined 2D maps, only two of the 2D maps before the combination, or all of the 2D maps.

[0127] The data storage unit 111 has a function of storing various pieces of information, various pieces of data, various intermediate products, various processing results, etc. acquired or generated by the above-mentioned functional units of the map generating device 10 in a readable manner.

[0128] [Functional classification of functional parts] A supplement to the functional units. Functional units having similar functions or processing contents may be realized by one functional unit, or may be provided by other functional units. For example, the first 2D map generating unit 201, the second 2D map generating unit 202, and the 2D map synthesis unit 203 may be executed by one processing unit. The divided space compression unit 106 and the 2D map generating unit 108 shown in FIG. 1 may also be executed by one processing unit. The divided space compression unit 106 may be provided with the second 2D map generating unit 202. The user can appropriately set the classification method of the functions and the names of the functional units.

[0129] [System Operation] Next, an example of the operation of the map generating system will be described with reference to Fig. 13. Fig. 13 is a sequence diagram showing the flow of processing of the map generating system 1. Note that the mobile robot 40 is a machine that travels on a two-dimensional map, and is equipped with a function that can recognize no-travel areas as metadata.

[0130] First, the sensor information acquisition unit 101 of the map generating device 10 acquires sensor information required to generate a 3D map of the target external world from the sensor device 20 (step S201).

[0131] Next, the 3D map generating unit 103 of the map generating device 10 generates a 3D map of the target external world as a common map MAP(1) using the sensor information (step S202).

[0132] For example, the 3D map generating unit 103 generates a common map MAP(1) of a 3D space by using a sensor or camera capable of 3D measurement and utilizing technology such as SLAM. The common map MAP(1) may be generated using a 3D measurement sensor mounted on the mobile robot 40, a camera installed in the environment, or a combination of these. Alternatively, the map may be generated from BIM / CIM (Building / Construction Information Modeling) or 3D-CAD information.

[0133] An example of the creation of a common map MAP(1) is shown in Fig. 14. For example, a common map MAP(1) is created that includes a four-legged desk 51, a desk mat 52 placed under the desk 51, and a trash can 53 placed next to the desk 51.

[0134] The processes after step S202 may be executed within the mobile robot 40 equipped with the map generating device 10, or may be executed within the cloud after uploading the sensor information to the cloud.

[0135] Next, the user inputs the height H at which the sensor is mounted on the mobile robot 40 to the input terminal 30 (step S203).

[0136] Next, the input terminal 30 transmits the sensor mounting height H of the mobile robot 40 input by the user to the map generating device 10 (step S204).

[0137] Next, the first two-dimensional map generation unit 201 of the map generation device 10 generates a two-dimensional map MAP(2) by slicing the common map MAP(1) at the sensor mounting height H (step S205).

[0138] An example of slicing the common map MAP(1) is shown in Fig. 15. The common map MAP(1) is sliced ​​on a two-dimensional plane V at the sensor mounting height H of the mobile robot 40 to generate a two-dimensional map MAP(2) that holds obstacle information included in the measurement range of the sensor. The two-dimensional map MAP(2) includes, for example, a two-dimensional obstacle 51' corresponding to the four legs of a desk 51 and a two-dimensional obstacle 53' corresponding to a trash can 53.

[0139] Next, the user inputs the height A of the bottom surface and the height B of the top surface of the mobile robot 40 into the input terminal 30 (step S206). The two heights A and B input here are values ​​that take into account the range that the mobile robot 40 cannot detect with its own sensor.

[0140] Next, the input terminal 30 transmits the height A of the bottom surface and the height B of the top surface input by the user to the map generating device 10 (step S207).

[0141] Next, the second 2D map generation unit 202 of the map generation device 10 generates a 2D map MAP(3) including obstacle information included within the height range of the mobile robot 40 by compressing into two dimensions the obstacles in the three-dimensional space included between the height A of the bottom surface and the height B of the top surface in the common map MAP(1) (step S208).

[0142] An example of the compression of the common map MAP(1) is shown in Fig. 16. The height space from the feet to the top of the mobile robot 40 (for example, when the mobile robot 40 runs on a floor, the space from 0 cm to the height Bcm of the top) is specified, and all information on obstacles etc. within that height space h is compressed onto the two-dimensional map. The compressed two-dimensional map MAP(3) holds information on obstacles that the mobile robot 40 can and cannot detect with its own sensors. For example, it includes a two-dimensional obstacle 51' corresponding to the four legs of a desk 51, a two-dimensional obstacle 52' corresponding to a desk mat, and a two-dimensional obstacle 53' corresponding to a trash can 53.

[0143] Here, depending on the mobile robot, the position of the mobile robot may be determined by the position of a mounted sensor, etc. In this case, the physical body (casing) of the mobile robot may make it difficult to move. Therefore, the area of ​​obstacles included in the two-dimensional map MAP(3) is expanded taking into account the size of the mobile robot 40, etc.

[0144] Then, the user inputs the length and width R of the mobile robot 40 (step S209).

[0145] Next, the input terminal 30 transmits the size R of the moving robot 40 input by the user to the map generating device 10 (step S210).

[0146] Next, the second two-dimensional map generation unit 202 of the map generation device 10 expands the obstacle area included in the two-dimensional map MAP(3) by the size R of the mobile robot 40 (step S211).

[0147] Specifically, as shown in FIG. 17, when the mobile robot 40 is projected onto the XY plane, if the distance from the center position of the mobile robot 40 to the farthest end point of the mobile robot 40 is R, the area is expanded by R from the boundary line of the obstacle included in the two-dimensional map MAP(3).

[0148] In addition, there are mobile robots whose running patterns differ according to the type of machine, such as those that run in two dimensions by moving forward and backward and tilting their wheels like a car, and those that can run in three dimensions by turning on the spot and moving sideways. In that case, as shown in Figure 18, the obstacle area may be expanded by the maximum value of the movement amount R(x, y) at one sampling rate of the mobile robot, that is, the maximum value Max(Vxmax Δt, Vymax Δt) when the movement speed in the x direction and y direction is integrated over a certain time.

[0149] Next, the two-dimensional map synthesis unit 203 of the map generating device 10 compares the two-dimensional map MAP(2) with the two-dimensional map MAP(3), and synthesizes the obstacle information contained only in the two-dimensional map MAP(3) into the two-dimensional map MAP(2) as a new no-travel area (step S212).

[0150] For example, as shown in FIG. 19(a), a two-dimensional map MAP(2)' to which a new no-driving area has been added is generated by combining a two-dimensional obstacle 52' that is included only in the two-dimensional map MAP(3) generated in step S208 with the two-dimensional map MAP(2) generated in step S205.

[0151] For example, as shown in FIG. 19(b), a two-dimensional map MAP(2)' is generated to which a new no-driving area has been added by combining the expanded obstacle area contained only in the two-dimensional map MAP(3) generated in step S211 with the two-dimensional map MAP(2) generated in step S205.

[0152] Next, the user inputs the specifications and address of the mobile robot 40 into the input terminal 30 (step S213).

[0153] Next, the input terminal 30 transmits the specifications and address of the mobile robot 40 input by the user to the map generating device 10 (step S214).

[0154] Next, the format conversion unit 109 of the map generating device 10 converts the format of the single 2D map MAP(2)' after the synthesis and the formats of the two 2D maps MAP(2) and MAP(3) before the synthesis to conform to the specifications (communication protocol, data format, etc.) of the mobile robot 40 (step S215).

[0155] Finally, the map information distribution unit 110 of the map generating device 10 distributes the single synthesized 2D map MAP(2)' after format conversion and / or the two unsynthesized 2D maps MAP(2), MAP(3) after format conversion to the mobile robot 40 (step S216).

[0156] The two-dimensional map MAP(3) may be used, for example, in such a way that the mobile robot 40 recognizes obstacle information included in the two-dimensional map MAP(3) as a no-travel area and moves around it. As shown in Fig. 20, in the case where there is a map distribution destination management device 60, the distribution destination management device 50 may use the two-dimensional map MAP(2) as a basic map for the mobile robot, and distribute the obstacle areas included in the two-dimensional map MAP(3) as no-travel areas.

[0157] Thereafter, the mobile robot 40 uses existing technology such as SLAM to move through space based on the two-dimensional map MAP(2)' or the two-dimensional maps MAP(2) and MAP(3). The two-dimensional map MAP(2)' and the two-dimensional map MAP(3) contain no-travel areas with added obstacle information that cannot be detected by the sensors of the mobile robot 40, so the mobile robot 40 can move through the target space while reliably avoiding obstacles. In addition, the no-travel areas contained in the two-dimensional map MAP(2)' and the two-dimensional map MAP(3) are expanded taking into account the size and movement amount of the mobile robot 40, so the mobile robot 40 can move through the target space while more reliably avoiding obstacles.

[0158] (Modification) The above operations and processing sequence of the map generation system 1 are just an example.

[0159] For example, steps S203, S206, S209, and S213 related to the user's actions may be executed before step S201. This allows the map generating device 10 to eliminate the waiting time for receiving data from the input terminal 30, and allows the two-dimensional map to be generated efficiently and quickly.

[0160] For example, the processing order of steps S203 to S205 and steps S206 to S211 may be reversed, or may be executed simultaneously.

[0161] For example, if there is no need to expand the obstacle area, steps S209 to S211 may be omitted.

[0162] [Effects of this embodiment] According to this embodiment, the map generating device 10 generates a 3D map of the moving space of the mobile machine as a common map using a sensor capable of 3D measurement, and then uses the common map to convert a map including information on obstacles in the space that can be recognized by the mobile machine to match the map specifications of the mobile machine and distributes it to the mobile machine. This eliminates the need for advance map generation, which is required when the mobile machine is actually introduced, and significantly reduces map generation costs.

[0163] Furthermore, according to this embodiment, the map generating device 10 converts information about all obstacles, etc. in the moving space that cannot be recognized by the sensors of the mobile machine, taking into account the size, etc., of the mobile machine, to conform to the map specifications of the mobile machine and distributes the converted information to the mobile machine, so that the mobile machine can recognize information about obstacles at the feet, above, etc. that cannot be recognized by the mobile machine, and can avoid collisions with obstacles that it cannot recognize.

[0164] [Other embodiments] It is also possible to combine the first and second embodiments.

[0165] [others] The present invention is not limited to the above-described embodiment, and various modifications are possible within the scope of the present invention.

[0166] The map generating device 10 of the present embodiment described above can be realized, for example, as shown in Fig. 20, by using a general-purpose computer system including a CPU 901, a memory 902, a storage 903, a communication device 904, an input device 905, and an output device 906. The memory 902 and the storage 903 are storage devices. In the computer system, the CPU 901 executes a predetermined program loaded on the memory 902, thereby realizing each function of the map generating device 10.

[0167] The map generating device 10 may be implemented by one computer. The map generating device 10 may be implemented by multiple computers. The map generating device 10 may be a virtual machine implemented in a computer. The program for the map generating device 10 may be stored in a computer-readable recording medium such as an HDD, an SSD, a USB memory, a CD, or a DVD. The computer-readable recording medium is, for example, a non-transitory recording medium. The program for the map generating device 10 may also be distributed via a communication network. [Explanation of symbols]

[0168] 1: Map generation system 10: Map generation device 20: Sensor device 30: Input terminal 40:Mobile robot 51: Desk 52: Desk mat 53: Trash can 54: Cable 55:Unknown object 51': 2D obstacle equivalent to a desk 52': 2D obstacle equivalent to a desk mat 53': 2D obstacle equivalent to a trash can 54': 2D obstacle equivalent to a cable 55': 2D obstacle corresponding to unknown object 60: Delivery destination management device 101: Sensor information acquisition unit 102: Input information reception unit 103: 3D map generation unit 104: Space division part 105: Obstacle detection unit 106: Partitioned space compression unit 107: Obstacle determination section 108: 2D map generation unit 109: Format conversion section 110: Map Information Distribution Department 111: Data storage unit 201: First 2D map generating unit 202: Second 2D map generator 203: 2D map synthesis unit 901:CPU 902: Memory 903: Storage 904: Communication equipment 905: Input device 906: Output device

Claims

1. a space dividing unit that divides a space in the three-dimensional map into a plurality of divided spaces in a height direction; an obstacle determination unit that determines an obstacle in each of the divided spaces; a divided space compression unit that compresses obstacles in each divided space two-dimensionally in a height direction and generates a plurality of compressed divided spaces including the compressed two-dimensional obstacles; an obstacle determination unit that determines whether a two-dimensional obstacle in each compressed divided space is an obstacle based on the presence or absence of a temporal connection between the two-dimensional obstacles in the compressed divided spaces; a two-dimensional map generating unit that combines one or more compressed divided spaces corresponding to a width of an arbitrary height among the plurality of compressed divided spaces in a height direction and generates the combined compressed divided spaces as a two-dimensional map, The obstacle determination unit is determining, as a false obstacle, a two-dimensional obstacle that is included in only one of a plurality of compressed divided spaces at the same position among the plurality of compressed divided spaces at the same height at different times; The two-dimensional map generating unit includes: Generate the two-dimensional map that reflects the result of determining whether the obstacle is genuine or not. Map generator.

2. The obstacle determination unit is The map generating device according to claim 1 , wherein the authenticity of a two-dimensional obstacle in each compressed divided space is determined based on the presence or absence of spatial connection in the height direction of the two-dimensional obstacle in each of the plurality of compressed divided spaces.

3. The obstacle determination unit is 2. The map generating device according to claim 1, wherein a two-dimensional obstacle included only in a middle compressed divided space among a plurality of compressed divided spaces positioned in the height direction is determined as a false obstacle.

4. The two-dimensional map generating unit includes:

2. The map generating device of claim 1, wherein the two-dimensional map is generated by combining one or more compressed divided spaces corresponding to a width of a height suitable for a given mobile machine.

5. The two-dimensional map generating unit includes:

2. The map generating device according to claim 1, wherein the area of ​​two-dimensional obstacles contained in the one or more compressed divided spaces is expanded based on the size or movement amount of a predetermined mobile machine.

6. A map generating method performed by a map generating device, A first step of dividing a space in a three-dimensional map into a plurality of divided spaces in a height direction; a second step of determining an obstacle in each of the divided spaces; a third step of compressing the obstacles in each divided space into two dimensions in a height direction to generate a plurality of compressed divided spaces including the compressed two-dimensional obstacles; A fourth step of determining whether or not the two-dimensional obstacles in each compressed divided space are obstacles based on the presence or absence of a temporal connection between the two-dimensional obstacles in the compressed divided spaces; a fifth step of combining, in a height direction, one or more compressed divided spaces corresponding to a width of an arbitrary height among the plurality of compressed divided spaces, and generating the combined compressed divided spaces as a two-dimensional map; In the fourth step, determining, as a false obstacle, a two-dimensional obstacle that is included in only one of a plurality of compressed divided spaces at the same position among the plurality of compressed divided spaces at the same height at different times; In the fifth step, Generate the two-dimensional map that reflects the result of determining whether the obstacle is genuine or not. Map generation method.

7. A map generating program that causes a computer to function as the map generating device according to claim 1.

Images