JP6686411B2 - Map making method - Google Patents

Description

  The present invention relates to a map creating method, and more particularly to a map creating method for an autonomous mobile robot.

  An autonomous mobile robot that autonomously moves using a map is known. For example, Patent Document 1 discloses a moving body control device having a route calculating means for calculating a moving route of a moving body on a grid grid map showing a moving environment of the moving body. This moving object control device sets a non-passable potential indicating non-passable on the grid map to avoid passing through a place where the non-passable potential is set. Further, for example, Non-Patent Document 1 discloses observing a movement route of a person in an environment, deriving a route (directed graph) connecting points at which a person stops, and using the directed graph for autonomous movement. There is.

JP, 2014-123200, A

Takeshi Sasaki, Drazen Brscic, and Hideki Hashimoto, "Human Observation Based Extraction of Path Patterns for Mobile Robot Navigation", IEEE Transactions on Industrial Electronics, 2010, Vol.57, No.4, pp.1401-1410

  When it is determined that only the position that is identified as having an obstacle is not passable, the area that is not identified as having an obstacle is determined to be passable regardless of the moving direction. For this reason, even if there is actually a direction in which the movement of the robot should be prohibited, the movement of the robot in that direction is permitted. In addition, in the technique described in Non-Patent Document 1, it is not possible to reflect a fine movement route on a map, and a map that can determine a specific route cannot be obtained. Further, in the technique described in Non-Patent Document 1, the above point is not created in a place where a person continues to move, and thus the movable direction cannot be appropriately set in such a place.

  The present invention has been made in order to solve such a problem, and when a direction in which the movement of the robot should be prohibited actually exists, a map creation method capable of suppressing the movement of the robot in that direction. The purpose is to provide.

A map creating method according to an aspect of the present invention is a map creating method of an autonomous mobile robot that autonomously moves on a map, and includes an unspecified person within a predetermined period within a map creating range. A method of acquiring a moving direction, and setting a movable direction of the autonomous mobile robot for each partial area in which the movement of a person has existed in the map creation range based on the acquired moving direction Is.
According to such a method, the direction in which there is a high probability of being movable is set as the movable direction. Therefore, it is possible to create a map that can prevent the robot from moving in the direction in which the movement should be prohibited.

  According to the present invention, it is possible to provide a map creation method capable of suppressing the movement of a robot in a direction in which the movement of the robot should be prohibited, when the direction actually exists.

It is a schematic diagram which shows an example of a directed grid map. It is a block diagram which shows the structure of the map creation apparatus which concerns on embodiment. It is a flowchart which shows the flow of operation | movement of a map update part. It is a schematic diagram which shows an example of the trajectory group of the person extracted in step 101 of the flowchart shown in FIG. It is a schematic diagram which shows an example of the directed grid map with which the potential value of all directions was set to 100%. It is a schematic diagram explaining the process of steps 105-108 of the flowchart shown in FIG. It is a schematic diagram which shows the example which changes a potential only about the grid to which the point by the side of a starting end belongs. It is a schematic diagram which shows the example which changes the potential of all the grids except the grid to which the point on the terminal side belongs. It is a schematic diagram which shows how to divide | segment the grid in a 4-direction directed grid map. It is a schematic diagram which shows how to divide | segment the grid in a directed grid map of 8 directions. It is a schematic diagram which shows the movement path of the observed actual person. 9 is a schematic diagram showing an example of a map generated using the technique described in Non-Patent Document 1. FIG. It is a schematic diagram which shows an example of the map produced | generated using a grid map.

  Before describing the embodiments of the present invention, first, items to be examined in advance by the inventor will be described.

  In the technique described in Non-Patent Document 1, as described above, the route that connects the points where the person stays is derived after observing the movement route of the person in the environment. For example, when the movement of a person as shown in FIG. 11 is observed, the map as shown in FIG. 12 can be created by using the technique described in Non-Patent Document 1. Note that in FIG. 11, there is an area where intrusion is prohibited due to some circumstances such as rules regarding social movement. Specifically, in FIG. 11, pylons 90 are placed at intervals around the area where entry is prohibited due to some circumstances. More specifically, FIG. 11 shows an environment in which there is an area surrounded by a pylon 90 on the passage between the walls 91. For this reason, the person moves while avoiding the area surrounded by the pylon 90.

  Here, referring to FIG. 12, in the map in the case of using the technique described in Non-Patent Document 1, only a rough route is set, and a fine movement route cannot be reflected. In particular, since the above points are not created in a place where a person continues to move, it is not possible to reflect the route to such a place on the map. In addition, the map shown in FIG. 12 includes a route passing through the area surrounded by the pylon 90.

  On the other hand, it is also conceivable to set a route as follows by setting a pass-through potential that indicates pass-through on the grid-like grid map. That is, as shown in FIG. 13, it is conceivable to update the grid map so that the position where the person existed can be passed. According to this method, it is possible to set the area where entry is prohibited to be unable to pass, but the moving direction is not set. For this reason, when the robot moves according to the map shown in FIG. 13, there is a possibility that the robot will move against the flow of people and that it will run backward on one way (walking sidewalk, etc.). That is, when there is actually a direction in which the robot movement should be prohibited, it is not possible to suppress the movement of the robot in that direction. Note that, in FIG. 13, the numerical value in each grid is a% value indicating the possibility that the area in the grid is a non-passable area.

EmbodimentsA map creating apparatus according to an embodiment includes a configuration for detecting a position of a person, a configuration for tracking a detected person, a configuration for storing a trajectory of a person, and a configuration for generating a map. There is. Here, in the configuration for detecting the position of the person, the position of the person within the map creation range is continuously detected. At this time, the configuration for detecting the position of the person is not required until the individual is identified, but the position of the unspecified person is continuously detected. It should be noted that the configuration for detecting the position of the person may also identify the individual when the configuration for tracking the detected person requires. The wider the detection range of the configuration for detecting the position of a person, the better. It is preferable that the position of the person can be detected at the same time, but only one person may be detected at the same time.

  The configuration for tracking the detected person uses the position information of the person obtained from the configuration for detecting the position of the person described above to track the person and derive a trajectory of the position of the person. Note that the configuration for tracking the detected person may use only the position information of the person, or may use other information additionally.

  The configuration for storing the trajectory of the person stores the trajectory information of the person obtained by the configuration for tracking the detected person. This configuration stores at least locus information of the position of the person. The trajectory information is composed of, for example, time-sequential position information. The position information included in the trajectory may include detected time information.

  The configuration for generating a map generates a map for an autonomous mobile robot from the stored human trajectory information. Specifically, the configuration for generating a map generates a directed grid map 50 as shown in FIG. The directed grid map 50 is a two-dimensional map as shown in FIG. 1, and is a map obtained by dividing the plane of the map creation range into a grid. Here, the directed grid map 50 is a map that numerically indicates, with respect to each grid, the possibility that the grid is a non-passable area. More specifically, the directed grid map 50 has, for each grid, a numerical value indicating the possibility of being impassable in each direction. Hereinafter, this numerical value may be referred to as a potential. FIG. 1 shows an example in which the potential is set for each grid in eight directions including the direction between them in addition to the top, bottom, left and right, but the number of directions can be set arbitrarily, and for example It may be four directions. In the present embodiment, the potential is indicated by the probability (% value). The number of grids of the directed grid map 50 can be set arbitrarily. Also, the size and shape of the grid can be arbitrarily set. As an example, the grid of the directed grid map 50 is a square whose one side corresponds to several cm of the real environment.

  In the configuration for generating a map, the trajectory of a person is acquired, and the trajectory of the person passing through is extracted for each grid. Then, in the configuration for generating a map, the direction of the trajectory is calculated for each extracted trajectory, and the potential of the calculated direction in the grid is updated each time the direction is calculated. The autonomous mobile robot selects a travel route and autonomously moves in accordance with the potential shown on the map thus generated. The map generated by the configuration for generating a map may be used as the map itself or may be used by being merged with another obstacle map or the like. In addition, the map may be created periodically in the configuration for generating the map, or when a certain instruction is given.

  Hereinafter, embodiments of the present invention will be specifically described with reference to the drawings. FIG. 2 is a block diagram showing the configuration of the map creation device 1 according to the embodiment.

  The map creation device 1 includes an information processing device 10 and a camera group 11. The information processing device 10 has a CPU (Central Processing Unit), a memory, and the like (not shown), and has a function as a computer. The information processing device 10 also includes a person detection unit 12, a person tracking unit 13, a trajectory storage unit 14, and a map update unit 15. The person detecting unit 12, the person tracking unit 13, and the map updating unit 15 can be realized by executing a program under the control of the CPU, for example. More specifically, it is realized by executing a program stored in a storage device such as a memory (not shown) under the control of the CPU. Further, each of these constituent elements is not limited to being realized by software by a program, but may be realized by any combination of hardware, firmware, and software. The locus storage unit 14 is realized by a storage device such as a memory.

  In the map creation device 1, the camera group 11 and the person detection unit 12 correspond to the above-described configuration for detecting the position of the person, and the camera group 11 and the person tracking unit 13 track the detected person. This corresponds to the configuration, the trajectory storage unit 14 corresponds to the configuration for storing the above-described person's trajectory, and the map update unit 15 corresponds to the configuration for generating the above-described map.

  The camera group 11 is a plurality of cameras, and is distributed and arranged in the range of the actual environment corresponding to the map creation range. That is, the camera group 11 is arranged so that the movement of a person within the map creation range can be photographed. Preferably, the camera group 11 sets the entire image capturing range as the image capturing range.

  The person detection unit 12 detects the position of a person from the image captured by the camera group 11. Specifically, the person detection unit 12 detects the coordinate position on the two-dimensional plane (map) and the position on the captured image for each person, and outputs the detected position to the person tracking unit 13. The person detection unit 12 detects the coordinate position on the two-dimensional plane from, for example, information about the shooting location of each camera of the camera group 11 and information about the position of the person in the captured image.

  The person tracking section 13 tracks a person using the position of the person detected by the person detecting section 12 and the image captured by the camera group 11. Specifically, for example, the person tracking unit 13 tracks a person by performing image processing using the detected position of the person on the captured image and the color information in the captured image. The person tracking unit 13 outputs a position group on a plane to the trajectory storage unit 14 as a trajectory for each tracked person.

  The locus storage unit 14 stores a locus group including the loci output by the person tracking unit 13. Further, the locus storage unit 14 outputs the stored locus group in response to a request from the map updating unit 15.

  The map updating unit 15 periodically updates the directed grid map 50 based on the information stored in the trajectory storage unit 14. Here, as an example, the directed grid map 50 will be described as a map in which potentials in four directions can be set. The update cycle of the map update unit 15 may be different depending on the environment of the area where the map is to be created. For example, in a place where a lot of people move, such as a passage, the map update unit 15 may update the map at a relatively short cycle such as once every 10 minutes. Further, in a place where movement of people is small, such as an office, the map updating unit 15 may update the map at a relatively long cycle such as twice a day.

  The details of the map updating unit 15 will be described below with reference to the flowchart. FIG. 3 is a flowchart showing the flow of operation of the map updating unit 15.

  In step 100 (S100), the map updating unit 15 acquires the trajectory group of the person stored in the trajectory storage unit 14.

  In step 101 (S101), the map updating unit 15 extracts a trajectory group of a person after the previous update from the trajectory group acquired in step 100. FIG. 4 is a schematic diagram showing an example of the trajectory group of a person extracted in step 101. In addition, in FIG. 4 and FIGS. 6 to 8 to be described later, an arrow indicates a detected locus of a person, and a dot indicates an element of the locus (position of the detected person). As shown in FIG. 4, since a person passes through a passage portion other than the area surrounded by the pylon 90, there is no human trajectory in the area surrounded by the pylon 90. Further, in the example shown in FIG. 4, a person passes by the side of the pylon 90, but at that time, there are two moving directions of the person. One of the moving directions is the moving direction existing on the side closer to the pylon 90, and is the moving direction from right to left in the figure. The other moving direction is the moving direction existing on the side farther from the pylon 90, and is the moving direction from left to right in the figure. That is, the person is driving on the right side. In this way, the map update unit 15 acquires the movement trajectory (movement direction) of an unspecified person within a predetermined period within the map creation range.

  Through steps 102 to 110, the map update section 15 creates an updated directed grid map. Specifically, in step 102 (S102), the map update unit 15 creates a map with 100% potential. That is, as shown in FIG. 5, the map updating unit 15 creates a directed grid map in which the potential values in all directions are set to 100% in all grids.

  In step 103 (S103), the map updating unit 15 extracts a trajectory from the trajectory group extracted in step 101.

  In step 104 (S104), the map updating unit 15 calculates the grid to which each of the points on the trajectory extracted in step 103 belongs. That is, the map updating unit 15 associates each grid forming the locus with which grid on the map. For example, the map updating unit 15 performs this association by calculating using the coordinates of each point and the position information regarding each grid.

  By the following steps 105 to 108, processing for correcting the potential in the direction corresponding to the actual moving direction of the person is performed. FIG. 6 is a schematic diagram for explaining the processing of steps 105 to 108 of the flowchart shown in FIG.

  In step 105 (S105), the map updating unit 15 extracts the coordinates of two adjacent points forming the trajectory extracted in step 103.

  In step 106 (S106), the map updating unit 15 determines whether or not the coordinates of two adjacent points extracted in step 105 belong to the same grid. If the coordinates of the two points do not belong to the same grid, the process proceeds to step 107. On the other hand, when the coordinates of the two points belong to the same grid, the process proceeds to step 109.

  In step 107 (S107), the map updating unit 15 calculates the intersection of the boundary line of the grid and the line segment connecting the two points.

  In step 108 (S108), the map updating unit 15 changes the potential of the grid to which the point on the starting end side of the two points belongs and in the direction in which the intersection exists to 0%.

  In step 109 (S109), the map updating unit 15 determines whether or not the processes of steps 106 to 108 have been performed for all of the two adjacent points forming the trajectory extracted in step 103. When the processes of steps 106 to 108 have not been completed for all the two adjacent points, the process returns to step 105. On the other hand, when the processes of steps 106 to 108 have been completed for all the two adjacent points, the process proceeds to step 110.

  In step 110 (S110), the map updating unit 15 determines whether or not the processes of steps 103 to 109 have been executed for all the trajectories forming the trajectory group extracted in step 101. When the processes of steps 103 to 109 have not been executed for all the trajectories, the process returns to step 103. On the other hand, when the processes of steps 103 to 100 have been completed for all trajectories, the process proceeds to step 111.

  In this way, the directed grid map 50 is generated based on the movement trajectory of an unspecified person within a predetermined period within the map creation range acquired in step 100. Here, in the area (grid) in which the movement of a person exists in the map creation range, the potential is changed from 100% of the initial value. That is, such an area is set as a passable area. However, in this passable region, the potential is changed from the initial value only in the direction corresponding to the actual moving direction of the person. Therefore, in the passable area, a direction corresponding to the actual moving direction of the person is set as the movable direction. Further, the area (grid) in which the movement of the person does not exist in the map creation range has the potential of 100%, and such an area is set as the impassable area.

  That is, the flowchart shown in FIG. 3 shows the following map creation method of the autonomous mobile robot that autonomously moves on the map. That is, the movement direction of an unspecified person within a predetermined period within the map creation range is acquired (see steps 100 to 107), and the person within the map creation range is acquired based on the acquired movement direction. The map creating method is shown in which the movable direction of the autonomous mobile robot is set for each partial area (grid) in which the movement of (1) has existed (see step 108).

  In step 111 (S111), the map update unit 15 updates the map by merging the past map and the map obtained by the above process. That is, the map update unit 15 merges the directed grid map obtained at the time of the previous update with the map obtained by the processing of steps 100 to 110 in the current update, and updates the map by the update processing of this time. A customized map. Any method can be adopted as the method of merging maps, and for example, merging by Bayesian estimation or weighted sum may be performed. Further, here, a new map is generated by merging the past map and the map obtained by the above processing, but the map obtained by the processing of step 100 to step 110 without merging. May be a map used by the autonomous mobile robot.

  According to the map creation device 1 according to the above-described embodiment, the direction in which there is a high probability of being movable is set as the movable direction. Therefore, it is possible to create a map that can prevent the robot from moving in the direction in which the movement should be prohibited. That is, it is possible to create a map that can reduce the possibility that the robot will move against the flow of people and the possibility that it will reverse one-way. In addition, since the directed grid map 50 as described above is created, the fine movement rule can be reflected in the map. Further, unlike the technique described in Non-Patent Document 1, it can be applied to an area where a person is constantly moving. In addition, since a map is created based on the actual movement of people, locations where entry is prohibited by some rule, or locations where entry is difficult to detect with sensors such as glass and dents are detected by the autonomous mobile robot. It can also be reflected in the map.

  In the above description, in step 108, the potential is changed only for the grid to which the point on the starting end side of the two points belongs (see FIG. 7), but the end of the grid existing on the line segment connecting the two points is changed. The potentials of all grids except the grid to which the side point belongs may be changed (see FIG. 8). Note that FIG. 7 is a schematic diagram showing an example in which the potential is changed only for the grid to which the point on the start end side among the two points belongs, and FIG. 8 shows the potential on all grids except the grid to which the point on the end side belongs. It is a schematic diagram which shows the example changed. In FIGS. 7 and 8, the grid surrounded by the bold frame is the grid whose potential is to be changed.

  Further, in the above description, the map updating unit 15 sets the potential in the direction corresponding to the moving direction of the person to 0%, but may decrease the potential from the initial value (100%) by a predetermined value. (See the potential on the lower left of Figure 1). The lower limit value of the potential in this case is 0%.

  Further, in the above description, the map updating unit 15 specifically shows an example in which the four-direction directed grid map 50 is generated, but as described above, the number of directions is not limited to four, and may be eight, for example. It may be a direction (see FIG. 1). However, in the case of four directions, the direction of the potential update target was determined depending on where each of the upper, lower, left, and right sides of the grid intersected (see FIG. 9). It is necessary to divide it (Fig. 10). FIG. 9 is a schematic diagram showing how to divide the grid in the directed grid map 50 in four directions, and FIG. 10 is a schematic diagram showing how to divide the grid in the directed grid map 50 in eight directions. Note that in FIGS. 9 and 10, the division of the sides of the grid is shown by the difference in the line type (solid line or broken line).

  The present invention is not limited to the above-mentioned embodiments, but can be modified as appropriate without departing from the spirit of the present invention. For example, although the camera group 11 is used in the above-described embodiment, a laser range finder group may be used that is distributed and arranged in the range of the actual environment corresponding to the map creation range. In this case, set the height of each laser range finder to about the height of a person's foot, and detect the person's position information by detecting the position of the person's foot from the output of the laser range finder to track the person. May be. Further, instead of disposing the camera group 11 or the laser range finder group, an autonomous mobile robot equipped with the camera or laser range finder may be configured to detect the position of a person while moving.

1 map creation device 10 information processing device 11 camera group 12 human detection unit 13 human tracking unit 14 trajectory storage unit 15 map update unit 50 directed grid map 90 pylon 91 wall

Claims (1)

A method for creating a map of an autonomous mobile robot that autonomously moves on a map by a map creating device , comprising:
The map creating device acquires a trajectory of movement of an unspecified person within a predetermined period within a map creating range ,
The map creation device has two adjacent points of the plurality of coordinate points of the person forming one trajectory, and the coordinates of the two points belong to different partial areas within the map creation range. Find a point,
The map creation device identifies an intersection of a boundary line of the partial area and a line segment connecting the two points,
The map creation device sets a direction in which the intersection exists to a direction in which the autonomous mobile robot can move, with respect to a partial region that is partitioned by the boundary line where the intersection exists and is a partial region on the starting end side of the trajectory. How to create a map.

Images