JP5518579B2 - Movable region extraction apparatus and movable region extraction method - Google Patents

Description

  The present invention relates to a movable area extracting apparatus and a movable area extracting method for extracting a movable area where a moving body that moves autonomously can move.

  A mobile body that moves autonomously (hereinafter referred to as an “autonomous mobile robot”) moves according to a movement route set for the autonomous mobile robot. In order for the autonomous mobile robot to move safely, it is important to generate an optimal movement path while avoiding obstacles existing in the movement path.

  A conventional autonomous mobile robot includes a detection device such as a camera, a laser range finder, and an ultrasonic sensor. When the autonomous mobile robot moves, it detects obstacles in the movement path based on the detection information obtained by the detection device, and sequentially updates the movement path so as to avoid the detected obstacle. Move to the desired location. However, the amount of calculation for generating obstacles by extracting obstacles from the detection information obtained by the detection device is very large, and the calculation time required for path generation is long. The large amount of calculation is a restriction on the speed at which the autonomous mobile robot moves. In order to avoid the restriction on the movement speed of the autonomous mobile robot due to the large amount of calculation, for example, in Patent Document 1, in the previous obstacle extraction, the area that was regarded as an obstacle and the detection information outside the previous detection range The time required for the next obstacle extraction and path generation is shortened by using only.

JP 2008-197884 A

  However, in general, it is conceivable that not only fixed obstacles but also moving objects such as humans and other autonomous mobile robots exist on the movement path along which the autonomous mobile robot moves. In Patent Document 1, the area where there is no obstacle in the previous obstacle extraction is not the area where the obstacle is detected next. For this reason, when a moving body enters the movement path of the autonomous mobile robot, the moving body that has entered cannot be detected as an obstacle, and this obstacle cannot be avoided. There is a problem that the robot cannot move safely.

  Conventionally, as a method for detecting an obstacle in the movement path of the autonomous mobile robot, a detection device is installed in the movement area of the autonomous mobile robot, and the autonomous mobile robot is detected based on the detection information obtained by the detection device. There is also a method for extracting obstacles existing in the moving route. For example, when an autonomous mobile robot moves indoors, a distance sensor such as a laser range finder is installed on the ceiling in the moving area, and an obstacle is detected by the laser range finder. And, by setting a movement path in the autonomous mobile robot that avoids the obstacle detected by the laser range finder installed on the ceiling, the autonomous mobile robot can safely be at the target location while avoiding (detouring) the obstacle. Can be moved to.

  As described above, when the laser range finder is installed at a high position in the moving area of the autonomous mobile robot, a wide range of obstacles can be detected by the laser range finder. Then, by performing an operation for extracting an obstacle from the detection information obtained by the laser range finder and generating a movement path using a processing device having a higher processing capacity, the speed at which the autonomous mobile robot moves can be determined. Restrictions can be avoided. It is also possible to detect a moving body that enters the movement path of the autonomous mobile robot at an early stage. However, the obstacle detection range by the laser range finder installed on the ceiling is, for example, conically widened, so the shadow of the obstacle placed on the floor where the autonomous mobile robot moves or the shadow of the moving object For example, an unknown area that cannot be detected by the laser range finder is generated. Normally, this unknown area is also used as an obstacle area, and if it is included in a non-movable area where the autonomous mobile robot cannot move, it may hinder the autonomous mobile robot from moving safely to the target location. Few.

  However, if all unknown areas are included in the non-movable area, the range in which the autonomous mobile robot avoids (bypasses) obstacles becomes large. There is a problem that an efficient movement route cannot be generated. In order to solve this problem and generate an efficient movement path, multiple laser range finders are placed on the ceiling in the movement area of the autonomous mobile robot so that some of the detection ranges of each laser range finder overlap. It is possible to install it. Then, it is conceivable to extract the final obstacle by fusing detection information of a plurality of laser range finders. Thereby, unknown areas, such as the shadow of an obstacle placed on the floor and the shadow of a moving object, can be reduced to some extent. However, since the control for synchronizing the detection timing of the obstacle by each laser range finder requires complicated processing, each laser range finder detects the obstacle asynchronously at present.

  In such a state, for example, when a person moves in a region where the detection ranges of the laser range finder overlap, the person is detected by two laser range finders. However, the position of the person detected by the two laser range finders is different for each laser range finder. As described above, in the state where each laser range finder detects an obstacle asynchronously, the detection timing of the two laser range finders is not the same time, but the same person is detected at different timings. It is because it ends. If this state is compared with another structure, it is like taking the same moving person with two cameras at different timings. When the detection information of the same person detected at different positions is merged, the range of the trajectory in which the person has moved, including the shadow of the person, is extracted as an obstacle area. This increases the range in which the autonomous mobile robot avoids (bypasses) obstacles. In addition, even when the person moves and is not at the position previously detected by the laser range finder, the autonomous mobile robot avoids (bypasses) the position of the person previously extracted as an obstacle. Will be set.

  Also, for example, when the autonomous mobile robot moves to the position where the laser range finder detects the autonomous mobile robot itself from the back by moving along the set movement route, the shadow of the autonomous mobile robot itself is also unknown It becomes the area of. In such a state, if the shadow area of the autonomous mobile robot itself is extracted as an obstacle area, the movable area where the autonomous mobile robot can move is lost by the shadow of the autonomous mobile robot itself. It will be.

  In this way, when trying to extract obstacles on the movement path where the autonomous mobile robot moves using the laser range finder installed on the ceiling, fixed obstacles such as tables and chairs that should be extracted originally In addition, there is a problem that a moving object such as a person or another autonomous mobile robot or the autonomous mobile robot itself is extracted as an obstacle. Then, if an obstacle other than the one that should be extracted is extracted as an obstacle that should be avoided (bypassed), there is a problem that it is not possible to generate a movement route that allows the autonomous mobile robot to move efficiently.

  The present invention has been made on the basis of the above-mentioned problem recognition, and a movement that moves from a movement path, such as a person or another autonomous mobile robot, when detecting an obstacle in a movement area where the moving body moves. It is an object of the present invention to provide a movable area extracting apparatus and a movable area extracting method for generating a moving route that allows a moving body to move efficiently to a destination without extracting the body as an obstacle.

In order to solve the above-described problem, the movable region extraction device according to the first aspect of the invention (for example, the movable region extraction unit 30 in the embodiment) is a mobile body that moves autonomously (for example, the embodiment). A first distance sensor (for example, the ceiling LRF 10 in the embodiment) installed so as to observe the lower part of the space from a predetermined first height in the space in which the autonomous mobile robot 100) moves A second distance sensor (for example, a lateral LRF 20 in the embodiment) installed so as to observe the side of the space from a second height lower than the predetermined first height in the space. ) To detect an obstacle existing in the space in which the moving body moves, and information on the area of the detected obstacle (for example, the non-floor grid and the unknown grid in the embodiment). ) Based on the first distance sensor in the movable area extracting device that extracts information on the movable area in which the movable body can move (for example, movable area information in the embodiment). First obstacle detection means (for example, the grid conversion unit 331 in the embodiment) for detecting a first obstacle area representing the obstacle area based on distance information, and the second distance sensor Second obstacle detection means (for example, the mobile object position extraction unit 340 in the embodiment) for detecting a second obstacle region representing the obstacle region based on the information on the distance observed by The first obstacle representing the area of the detected moving object from the first obstacle area by detecting the obstacle continuously moving based on the second obstacle area as a moving object. Exclude object area The area, by judging an immovable area, an obstacle region extracting means for extracting a movable area which can be the mobile moves (e.g., movement in the embodiment body position canceling unit 332), It is characterized by comprising.

  According to the present invention, based on the information on the distance to the obstacle observed by the first distance sensor (for example, the ceiling LRF 10 in the embodiment) installed at the first height, the mobile body (for example, the implementation) The first obstacle region existing in the space in which the autonomous mobile robot 100) moves is detected. Moreover, based on the information on the distance to the obstacle observed by the second distance sensor (for example, the side LRF 20 in the embodiment) installed at the second height lower than the first height, the moving body is A second obstacle area existing in the moving space is detected. In addition, based on the second obstacle area, a moving object continuously moving in the space in which the moving object moves is detected. And the 1st obstacle area | region except the 1st obstacle area | region obtained by observing a moving object is extracted as a movable area | region which a mobile body can move.

  According to a second aspect of the present invention, in the second obstacle detection means, the obstacle representing the position of the obstacle based on the information of the distance when the second distance sensor observes the obstacle. A position (for example, a candidate position in the embodiment) is extracted, and the obstacle region extracting unit extracts the obstacle position (for example, a current candidate in the embodiment) extracted by the second obstacle detecting unit. Position) and a past obstacle position (for example, a past person candidate position in the embodiment) previously extracted by the second obstacle detection means, the moving object is detected, and the detected movement is detected. An area excluding the first obstacle area including the obstacle position representing the position of an object is defined as the movable area.

  According to the present invention, based on the information on the distance to the obstacle observed by the second distance sensor, the obstacle position of the observed obstacle (for example, the candidate position in the embodiment) is extracted. The obstacle position obtained by the current observation (for example, the current person candidate position in the embodiment) and the past obstacle position obtained by the previous observation (for example, the past person candidate in the embodiment). And a moving object that is continuously moving in the space in which the moving object moves is detected. The movable body can move in the first obstacle area excluding the first obstacle area including the obstacle position of the moving object (for example, the current candidate position in the embodiment). Extract as a region.

  The second obstacle detection means of the invention described in claim 3 is an obstacle that represents a position of the obstacle based on the information of the distance when the second distance sensor observes the obstacle. A position (for example, a candidate position in the embodiment) is extracted, and the obstacle region extracting unit extracts the obstacle position (for example, a current candidate in the embodiment) extracted by the second obstacle detecting unit. Position) and a past obstacle position (for example, a past person candidate position in the embodiment) previously extracted by the second obstacle detection means, the moving object is detected, and the detected movement is detected. An area excluding the first obstacle area in a predetermined range including the obstacle position that represents the position of an object (for example, the movable object cancellation range in the embodiment) is set as the movable area. It is characterized by.

  According to the present invention, based on the information on the distance to the obstacle observed by the second distance sensor, the obstacle position of the observed obstacle (for example, the candidate position in the embodiment) is extracted. The obstacle position obtained by the current observation (for example, the current person candidate position in the embodiment) and the past obstacle position obtained by the previous observation (for example, the past person candidate in the embodiment). And a moving object that is continuously moving in the space in which the moving object moves is detected. Then, the first obstacle region in a predetermined range (for example, the moving object cancellation range in the embodiment) including the obstacle position of the moving body (for example, the current candidate position in the embodiment) is excluded. The first obstacle area is extracted as a movable area where the moving body can move.

The obstacle area extracting means of the invention described in claim 4 searches the first distance sensor installed at a position closest to the obstacle position representing the position of the moving object, and Information on the distance between the obstacle position and the searched first distance sensor (for example, the distance Lh in the embodiment), and the distance when the searched first distance sensor observes the moving object Based on the height information of the moving object included in the moving object (for example, the height Hs and the height Hh in the embodiment) of the shadow area that cannot be observed by the first distance sensor by the moving object . The length (for example, the length Ls of the shadow in the embodiment) is calculated, and the predetermined range from the obstacle position of the moving object to a position separated by the calculated length of the shadow area. The obstacle position of the moving object And moving at an extension of connecting the installation position of the first distance sensors the search, the region excluding the first obstacle area range in which the predetermined and the movable area, It is characterized by that.

  According to the present invention, the first distance sensor installed at the position closest to the obstacle position of the moving object is searched. Then, the distance between the obstacle position of the moving object and the first distance sensor (for example, the distance Lh in the embodiment) and the height information when the first distance sensor observes the moving object (for example, implementation) Based on the height Hs and the height Hh) in the form of the shaded area, the length of the shaded area in which the first distance sensor cannot observe the distance information (eg, in the embodiment) The shadow length Ls) is calculated. Then, the obstacle position of the moving object between a predetermined range (for example, the moving body canceling range in the embodiment) and a position away from the obstacle position of the current moving object by the length of the shadow area. And the moving body may move in the first obstacle area excluding the first obstacle area included in the shadow area. Extract as possible movable area.

  The obstacle area extracting means according to claim 5 is characterized in that the obstacle position extracted by the second obstacle detecting means is a past obstacle previously extracted by the second obstacle detecting means. Determining that the obstacle corresponding to the obstacle position is the moving object when the position changes by more than a predetermined threshold (for example, a predetermined distance in the embodiment) from the position. Features.

  According to this invention, the difference between the current obstacle position obtained by observation by the second distance sensor and the past obstacle position obtained by previous observation by the second distance sensor is determined in advance. The obstacle located at the current obstacle position is continuously moving in the space in which the moving object moves when the threshold is equal to or greater than the threshold (for example, a predetermined distance in the embodiment). It is determined that the object is moving.

  When the space in which the moving body of the invention described in claim 6 moves is in a building, a plurality of the first distance sensors are installed at a predetermined interval on the ceiling in the building, and the first The obstacle detection means detects the first obstacle region based on distance information in a space obtained by merging the observation ranges of the plurality of first distance sensors installed.

  According to this invention, when the space in which the moving body moves is in the building, the first height at which the first distance sensor is installed is set as the height of the ceiling in the building. A plurality of first distance sensors are installed at predetermined intervals on the ceiling in the building. And the space which united the observation range of each 1st distance sensor installed in multiple becomes space where a mobile body moves. Then, the first obstacle detection means detects the first obstacle region based on distance information in a space in which the observation ranges of the respective first distance sensors are merged.

  The second distance sensor of the invention described in claim 7 is configured to observe an observation range parallel to the floor surface in the building at a predetermined height between the ceiling and the floor surface. A plurality of second obstacle detection means are installed at a predetermined interval, and the second obstacle detection means is configured to select the second obstacle based on distance information in a plane obtained by merging observation ranges of the plurality of second distance sensors installed. The obstacle area is detected.

  According to this invention, when the space in which the moving body moves is in the building, the second height at which the second distance sensor is installed is set to the height of the ceiling in the building and the height of the floor in the building. It is set to a predetermined height. In addition, a plurality of second distance sensors are installed at predetermined intervals so that the observation range of the second distance sensor is parallel to the floor surface in the building. And the plane which united the observation range of each 2nd distance sensor installed in plurality becomes a plane in the space where a mobile object moves. Then, the second obstacle detection means detects the second obstacle region based on distance information on a plane obtained by merging the observation ranges of the respective second distance sensors.

  The obstacle area extracting means of the invention described in claim 8 is a path generating means for generating a path along which the moving body moves based on the extracted movable area (for example, the path generating unit 40 in the embodiment). , Further comprising.

  According to the present invention, the obstacle region extraction unit further includes a route generation unit (for example, the route generation unit 40 in the embodiment) that generates a route along which the moving body moves based on the extracted movable region.

The movable region extraction apparatus according to the invention described in claim 9 (for example, the movable region extraction unit 31 in the embodiment) is such that the mobile body that moves autonomously (for example, the autonomous mobile robot 100 in the embodiment) moves. And a distance sensor (for example, the ceiling LRF 10 in the embodiment) installed so as to observe the lower part of the space from a predetermined height in the space to be present in the space in which the moving body moves. An obstacle is detected, and based on information on the area of the detected obstacle (for example, non-floor grid and unknown grid in the embodiment), information on a movable area in which the moving body can move (for example, In the movable area extracting device for extracting the movable area information in the embodiment, the obstacle is based on the distance information observed by the distance sensor. Based on the information of the amount of reflected wave when the distance sensor observes the obstacle, the obstacle detection means (for example, the grid conversion unit 331 in the embodiment) for detecting the obstacle area representing the area of the object, The shape detection means for detecting the shape of the obstacle (for example, the moving body position extraction unit 380 in the embodiment) and the shape of the obstacle detected by the shape detection means are compared with a predetermined object shape. , When the shape of the obstacle is equivalent to the shape of the predetermined object, the obstacle sensor continuously moves based on the distance information when the distance sensor observes the obstacle. The obstacle area representing the area of the obstacle that is continuously moving from the obstacle area , the obstacle area set according to the speed of the obstacle Area excluding , By determining the immovable area, comprises, an obstacle region extracting means for extracting a movable area which can be the mobile moves (e.g., mobile location cancellation unit 332 in the embodiment), It is characterized by that.

  According to this invention, based on the information on the distance to the obstacle observed by the distance sensor (for example, the ceiling LRF 10 in the embodiment) installed at a predetermined height, the mobile body (for example, the embodiment) The obstacle region existing in the space in which the autonomous mobile robot 100) moves is detected. Further, the shape of the obstacle is detected based on the information on the amount of the reflected wave when the distance sensor observes the obstacle. In addition, when the shape of the detected obstacle is equal to the predetermined shape of the object, the obstacle is determined based on the distance information when the distance sensor observes the obstacle. It is determined whether the moving object is continuously moving in the moving space. And the obstacle area | region except the obstacle area | region obtained by observing a moving object is extracted as a movable area | region which a mobile body can move.

The movable region extraction method according to the invention described in claim 10 is installed so as to observe the lower part of the space from a first predetermined height in the space in which the moving body that moves autonomously moves. A first distance sensor and a second distance sensor installed to observe a side of the space from a second height lower than the predetermined first height in the space; A movable area that detects an obstacle existing in the space in which the moving body moves and extracts information on a movable area in which the moving body can move based on information on the area of the detected obstacle. In the extraction method, a first obstacle detection procedure for detecting a first obstacle region representing the obstacle region based on information on a distance observed by the first distance sensor, and the second distance Based on distance information observed by sensors A second obstacle detection procedure for detecting a second obstacle region representing the obstacle region, and the obstacle continuously moving based on the second obstacle region as a moving object The moving object moves by detecting and determining an area excluding the first obstacle area representing the detected area of the moving object from the first obstacle area as a non-movable area. And an obstacle region extraction procedure for extracting a movable region that can be performed.

  According to this invention, based on the information on the distance to the obstacle observed by the first distance sensor installed at the first height, the first obstacle area existing in the space in which the moving body moves. Based on the information on the distance to the obstacle observed by the second distance sensor installed at the second height lower than the first height, the second existing in the space in which the moving body moves Detect obstacle areas. Then, based on the second obstacle region, the moving object continuously moving in the space in which the moving object moves is detected, and the first obtained from the information of the distance at which the detected moving object is observed. The first obstacle area excluding the obstacle area is extracted as a movable area where the moving body can move.

A movable region extraction method according to an eleventh aspect of the present invention is a distance sensor installed so as to observe a lower portion of the space from a predetermined height in the space in which the autonomously moving mobile body moves. , Detecting an obstacle existing in the space in which the moving body moves, and extracting information on a movable area in which the moving body can move based on information on the area of the detected obstacle. In the movable region extraction method, an obstacle detection procedure for detecting an obstacle region representing the obstacle region based on information on a distance observed by the distance sensor, and when the distance sensor observes the obstacle Based on the information on the amount of reflected wave, the shape detection procedure for detecting the shape of the obstacle is compared with the shape of the obstacle detected by the shape detection procedure and the shape of the predetermined object. When the shape is equivalent to the shape of the predetermined object, the distance sensor detects continuous movement of the obstacle based on the information of the distance when the obstacle is observed, The obstacle area representing the area of the detected obstacle that is continuously moving from the obstacle area , excluding the obstacle area set in accordance with the speed of the obstacle. And an obstacle region extraction procedure for extracting a movable region in which the moving body can move by determining the region as a non-movable region .

  According to this invention, based on the information on the distance to the obstacle observed by the distance sensor installed at a predetermined height, the obstacle area existing in the space in which the moving body moves is detected, and the distance The shape of the obstacle is detected based on information on the amount of reflected wave light when the sensor observes the obstacle. Then, when the detected obstacle shape is equivalent to the predetermined object shape, the distance sensor continues in the space where the moving body moves based on the distance information when the obstacle is observed. It is determined whether the moving object is moving. And the obstacle area | region except the obstacle area | region obtained from the information of the distance which observed the moving object is extracted as a movable area | region which a mobile body can move.

According to the first aspect of the present invention, based on the information on the distance to the obstacle observed by the first distance sensor installed at the first height, 1 obstacle area is detected. Further, the second obstacle existing in the space in which the moving body moves based on the information on the distance to the obstacle observed by the second distance sensor installed at the second height lower than the first height. Detect object areas. For this reason, the obstacle in the space in which the moving body moves can be observed over a wide range from the first height, and the obstacle having the second height lower than the first height can be extracted.
Further, in order to detect a moving object continuously moving in the space in which the moving body moves based on the second obstacle region, the first moving in the space in which the moving body moves is detected. A moving object having a height of 2 or more can be detected.
In addition, the first distance sensor is used to extract the first obstacle area excluding the first obstacle area obtained by observing the moving object as a movable area where the moving body can move. Thus, information obtained by removing a moving object having a height greater than or equal to the second height from information on the area of the obstacle observed over a wide range can be output as information on the movable area.

According to the second aspect of the invention, the obstacle position of the observed obstacle is extracted based on the information on the distance to the obstacle observed by the second distance sensor. The center position can be obtained.
In addition, based on the obstacle position obtained by the current observation and the past obstacle position obtained by the previous observation, the moving object moving continuously in the space where the moving object moves is detected. Therefore, it is possible to obtain information on the amount of movement (for example, distance) when the position representing the moving object (for example, the center position) has moved from the past position to the current position.
Further, in order to extract the first obstacle area excluding the first obstacle area including the obstacle position of the moving body as a movable area where the moving body can move, a predetermined amount of movement is determined. By removing only the moving object from the information on the obstacle area, the range of the movable area in which the moving body can move can be expanded without removing the fixed obstacle area.

According to the invention described in claim 3, in order to extract the obstacle position of the observed obstacle based on the information of the distance to the obstacle observed by the second distance sensor, the obstacle is represented, for example, The center position can be obtained.
In addition, based on the obstacle position obtained by the current observation and the past obstacle position obtained by the previous observation, the moving object moving continuously in the space where the moving object moves is detected. Therefore, it is possible to obtain information on the amount of movement (for example, distance) when the position representing the moving object (for example, the center position) has moved from the past position to the current position.
In addition, in order to extract the first obstacle area excluding the first obstacle area in a predetermined range including the obstacle position of the moving body as a movable area where the moving body can move, Movement that allows the moving object to move without removing the fixed obstacle area by removing only the predetermined range including the moving object of the predetermined moving amount from the obstacle area information. The range of possible areas can be expanded.

  According to the invention described in claim 4, the first distance sensor installed at the position closest to the obstacle position of the moving object is searched. Then, based on the distance between the obstacle position of the moving object and the first distance sensor, and the height information when the first distance sensor observes the moving object, the first object becomes the shade of the moving object. The length of the shadow area where the distance sensor cannot observe the distance information is calculated. Then, the obstacle position of the moving object and the installation position of the first distance sensor are connected within a predetermined range from the obstacle position of the current moving object to a position separated by the length of the shadow area. The first obstacle area excluding the first obstacle area included in the shadow area by moving on the extension line is extracted as a movable area where the moving body can move. For this reason, in the area of the obstacle remaining outside the predetermined range, the area that cannot be detected by the first distance sensor due to the shadow of the moving object can be excluded from the information on the area of the obstacle. It is no longer necessary to narrow the movable area where the moving body can move.

  According to the invention described in claim 5, the difference between the current obstacle position obtained by the observation by the second distance sensor and the past obstacle position obtained by the previous observation by the second distance sensor. Is equal to or greater than a predetermined threshold value, it is determined that the obstacle located at the current obstacle position is a moving object that is continuously moving in the space in which the moving object moves. The moving object that needs to be handled in the same manner as the fixed obstacle due to the slow movement speed is not excluded from the information of the obstacle area. For this reason, the movable area where the moving body can move is not expanded too much carelessly.

According to the invention described in claim 6, when the space in which the moving body moves is in the building, the first height for installing the first distance sensor is set in advance as the ceiling height in the building. A plurality of first distance sensors are installed at regular intervals. And since the space which united the observation range of each 1st distance sensor installed in multiple numbers becomes the space where a mobile body moves, the area | region which cannot detect the obstacle in the space where a mobile body moves ( Occlusion area) can be reduced, and obstacles in the space where the moving body moves can be observed over a wide range.
The first obstacle detection means detects the first obstacle region based on the distance information in the space obtained by merging the observation ranges of the respective first distance sensors. Obstacles in the entire range can be observed, and moving objects entering the moving path of the moving object can be observed at an early stage.

  According to the seventh aspect of the present invention, when the space in which the moving body moves is in the building, the second height at which the second distance sensor is installed is set to the ceiling height in the building and the floor in the building. A plurality of second distance sensors are arranged at predetermined intervals so that the observation range of the second distance sensor is parallel to the floor surface in the building as a predetermined height between the surface heights. Install. And the plane which united the observation range of each 2nd distance sensor installed in plurality becomes a plane in the space where a mobile object moves. Then, the second obstacle detection means detects the second obstacle region based on distance information on a plane obtained by merging the observation ranges of the respective second distance sensors. For this reason, it is possible to detect a moving object (for example, a person or another autonomous mobile robot) having a height higher than the height assumed to exist in the space in which the moving body moves.

  According to the eighth aspect of the present invention, the obstacle region extracting means further includes route generating means for generating a route along which the moving body moves based on the extracted movable region, so that information on the extracted movable region is provided. Based on the above, it is possible to efficiently generate a moving path along which the moving body moves.

According to the ninth aspect of the present invention, based on the information on the distance to the obstacle observed by the distance sensor installed at a predetermined height, the obstacle area existing in the space in which the moving body moves Is detected. Further, the shape of the obstacle is detected based on the information on the amount of the reflected wave when the distance sensor observes the obstacle. In addition, when the shape of the detected obstacle is equal to the predetermined shape of the object, the obstacle is determined based on the distance information when the distance sensor observes the obstacle. It is determined whether the moving object is continuously moving in the moving space. For this reason, it is possible to detect a predetermined specific moving object that is continuously moving in the space in which the moving body moves.
In addition, in order to extract the obstacle area excluding the obstacle area obtained by observing the moving object as a movable area where the moving body can move, information on the area of the obstacle observed by the distance sensor Therefore, information excluding a specific moving object can be output as information on the movable area.

  According to the invention described in claim 10, based on the information on the distance to the obstacle observed by the first distance sensor installed at the first height, In the space where the moving body moves based on the information on the distance to the obstacle detected by the second distance sensor installed at the second height lower than the first height. The second obstacle area existing in the area is detected. Then, based on the second obstacle region, the moving object continuously moving in the space in which the moving object moves is detected, and the first obtained from the information of the distance at which the detected moving object is observed. The first obstacle area excluding the obstacle area is extracted as a movable area where the moving body can move. For this reason, information obtained by removing a moving object having a height equal to or higher than the second height from the information on the area of the obstacle observed over a wide range by the first distance sensor may be output as information on the movable area. it can.

  According to the eleventh aspect of the present invention, the obstacle region exists in the space in which the moving body moves based on the information on the distance to the obstacle observed by the distance sensor installed at a predetermined height. And the shape of the obstacle is detected based on the information on the amount of reflected wave light when the distance sensor observes the obstacle. Then, when the detected obstacle shape is equivalent to the predetermined object shape, the distance sensor continues in the space where the moving body moves based on the distance information when the obstacle is observed. It is determined whether the moving object is moving. And the obstacle area | region except the obstacle area | region obtained from the information of the distance which observed the moving object is extracted as a movable area | region which a mobile body can move. For this reason, information obtained by removing a specific moving object from information on the area of the obstacle observed by the distance sensor can be output as information on the movable area.

It is the figure which showed typically the movement area | region where the route generation system by the 1st Embodiment of this invention was used. 1 is a block diagram showing a schematic configuration of a route generation system according to a first embodiment of the present invention. It is a figure explaining the classification of the grid registered in the movable area extraction part of the 1st embodiment. It is a figure explaining the example of the determination method of the immovable area | region in the movable area | region extraction part of the 1st embodiment. It is a figure explaining the example of the calculation method of the mobile body position in the movable area | region extraction part of the 1st embodiment. It is a figure explaining the 1st example of the cancellation method of the moving body position in the movable area extraction part of the 1st embodiment. It is a figure explaining the 2nd example of the cancellation method of the moving body position in the movable area extraction part of the 1st embodiment. It is a figure explaining the example of the mobile body detected in the route generation system of the 1st embodiment. It is a figure explaining the example of the cancellation method of the area | region of the shadow of a moving body in the movable area | region extraction part of the 1st embodiment. It is the flowchart which showed the process sequence of the movement route generation process in the route generation system of the 1st embodiment. It is a figure explaining the cancellation process of the moving body in the route generation system of the 1st embodiment. It is the block diagram which showed schematic structure of the route generation system by the 2nd Embodiment of this invention. It is a figure explaining the example of the determination method of the mobile body in the mobile body position extraction part of the 2nd embodiment.

<First Embodiment>
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram schematically showing a moving region in which the route generation system according to the first embodiment of the present invention is used. In FIG. 1, a plurality of laser range finders (laser rangefinders, hereinafter referred to as “LRF”) are installed on the ceiling and the side. In the following description, the LRF installed on the ceiling is referred to as a ceiling LRF10. Further, the LRF installed on the side is referred to as a side LRF 20. Further, when the ceiling LRF 10 and the side LRF 20 are not specified, they are simply referred to as “LRF”. In FIG. 1, a table, a chair, or the like is detected as an obstacle that is fixed in the movement area of the autonomous mobile robot 100, and is detected as an obstacle (moving body) in which a person and the autonomous mobile robot 100 are moving. Represents the case.

  FIG. 2 is a block diagram showing a schematic configuration of the route generation system according to the first embodiment. In FIG. 2, the route generation system 1 includes a ceiling LRF 10, a side LRF 20, a movable region extraction unit 30, and a movement route generation unit 40. The movable region extraction unit 30 includes a distance information acquisition unit 310 corresponding to the ceiling LRF 10, a distance information acquisition unit 320 corresponding to the side LRF 20, an obstacle position extraction unit 330, and a moving body position extraction unit 340. ing. The obstacle position extraction unit 330 includes a grid conversion unit 331, a moving body position cancellation unit 332, and a moving body shadow cancellation unit 333.

  The path generation system 1 detects an obstacle in a moving area where the autonomous mobile robot moves by the ceiling LRF 10 and moves a moving object (for example, a person or another autonomous mobile robot) existing in the moving area by the side LRF 20. Body and autonomous mobile robot 100 itself). Then, the route generation system 1 determines the area of the autonomous mobile robot 100 by removing the area of the moving body (moving obstacle) detected by the side LRF 20 from the area of all obstacles detected by the ceiling LRF 10. Extracted as a movable area, and a movement path of the autonomous mobile robot 100 is generated based on the extracted information on the movable area.

  The ceiling LRF 10 is a three-dimensional distance measuring sensor that detects a distance from an object (obstacle) existing in the observation range. The ceiling LRF 10 is installed at a high place in the moving area of the autonomous mobile robot 100, and emits a plurality of laser beams in a predetermined observation range in accordance with control from a control unit (not shown). And the reflected wave which the laser beam with which ceiling LRF10 irradiated was reflected by the object is observed. The ceiling LRF 10 outputs the observed reflected wave information to the movable region extraction unit 30.

  The side LRF 20 is a two-dimensional distance measuring sensor that detects a distance from an object (obstacle) existing within the observation range. The side LRF 20 is installed at a position of a predetermined height (for example, a height at which a person can be detected) in the movement area of the autonomous mobile robot 100, and in accordance with control from a control unit (not shown), A plurality of laser beams are irradiated within a predetermined observation range. Then, the reflected wave reflected by the object is observed by the laser beam irradiated by the side LRF 20. The side LRF 20 outputs the information of the observed reflected wave to the movable area extraction unit 30.

  Note that the ceiling LRF 10 in the first embodiment can detect the height of a surface (for example, a floor surface in a building) on which the autonomous mobile robot 100 moves within the movement area of the autonomous mobile robot 100. The description will be made assuming that the device is installed at a predetermined position with a certain posture and a certain height (for example, a ceiling in a building). The ceiling LRF 10 irradiates a conical range with the ceiling LRF 10 as an apex, for example, to detect an object existing within the observation range. Further, the side LRF 20 in the first embodiment has a predetermined height from a surface (for example, a floor surface in a building) on which the autonomous mobile robot 100 moves within the movement area of the autonomous mobile robot 100. In the following description, it is assumed that the robot is installed at a predetermined position in a certain posture capable of detecting the horizontal plane. Then, for example, the side LRF 20 irradiates a laser beam onto a fan-shaped range on a horizontal plane where the side LRF 20 is installed, and detects an object existing within the observation range.

  Further, the installation height and installation posture of the ceiling LRF 10 and the side LRF 20 are calibrated in advance so that the coordinates of the observation range of the ceiling LRF 10 and the side LRF 20 coincide with the coordinates of the movable area of the autonomous mobile robot 100. (Adjusted). Accordingly, an object of a certain height that can be observed by the side LRF 20 within a moving range of the autonomous mobile robot 100 is fixed at the same position by the ceiling LRF 10 and the side LRF 20. Observed as an object.

  As for the installation height and installation posture of the ceiling LRF 10, when the moving area of the autonomous mobile robot 100 is represented by a grid map, at least one point of the laser beam emitted from the ceiling LRF 10 is applied to each grid in the grid map. It is desirable to install in such a height and posture. Note that the size of the grid in the grid map representing the moving area of the autonomous mobile robot 100 is determined according to the size of the object to be detected by the ceiling LRF 10. For example, when trying to detect a small object such as a mobile phone falling on the floor, the grid is made smaller according to the assumed object, and when trying to detect a person, Increase the grid according to the width.

  Further, the installation posture of the side LRF 20 is such that when there is an object to be detected within the observation range of the side LRF 20, the laser beam emitted from the side LRF 20 is emitted to at least one point on the object. It is desirable to install in.

  In addition, when it is not possible to observe all the moving areas in which the autonomous mobile robot 100 moves by one ceiling LRF10 or side LRF20, a new ceiling LRF10 or side LRF20 is installed at another position, By merging the areas detected by the ceiling LRF 10 or the side LRF 20, it is possible to observe all the moving areas in which the autonomous mobile robot 100 moves.

  The movable area extraction unit 30 extracts obstacles in the movement area of the autonomous mobile robot 100 based on the information on the reflected wave input from the ceiling LRF 10 and the information on the reflected wave input from the side LRF 20. Then, the position information of the obstacle excluding the obstacle determined to be a moving body from all the extracted obstacles, that is, the position information of the obstacle fixed in the movement area of the autonomous mobile robot 100, The data is output to the movement route generation unit 40.

  The distance information acquisition unit 310 calculates the distance between the ceiling LRF 10 and the object that has reflected the laser beam, based on the information of the reflected wave input from the ceiling LRF 10. Then, the calculated distance information and the scan angle information of the laser beam emitted from the ceiling LRF 10 are output to the obstacle position extraction unit 330 as distance data.

  The distance information acquisition unit 320 calculates the distance between the side LRF 20 and the object that has reflected the laser beam based on the information of the reflected wave input from the side LRF 20. Then, the calculated distance information and the scan angle information of the laser beam irradiated by the side LRF 20 are output as distance data to the moving body position extracting unit 340.

The moving body position extraction unit 340 observes the side LRF 20 based on the information on the scan angle of the laser beam included in the distance data input from the distance information acquisition unit 320 and the information on the distance between the side LRF 20 and the object. A two-dimensional coordinate (x, y) representing the center position of the object is calculated. For example, the coordinates of a point at which the distance information between the side LRF 20 and the object included in the distance data input from the distance information acquisition unit 320 changes by a predetermined value or more are used as the scan angle of the laser beam included in the distance data. It calculates based on the information of each. Then, the coordinates of the center when the calculated two coordinates are connected by a straight line are set as two-dimensional coordinates (x, y) representing the center position of the object.
In addition, the moving body position extraction unit 340 outputs data representing the coordinates of the calculated center position of the object (hereinafter referred to as “moving body candidate data”) to the obstacle position extraction unit 330.

  The obstacle position extraction unit 330 is configured to detect the object observed by the ceiling LRF 10 based on the information on the scan angle of the laser beam included in the distance data input from the distance information acquisition unit 310 and the information on the distance between the ceiling LRF 10 and the object. The three-dimensional (x, y, z) coordinates representing the position and height information are calculated. For example, a horizontal plane that represents the observation range of the ceiling LRF 10 around the ceiling LRF 10 is assumed. Then, the position of the object observed by the ceiling LRF 10 in the horizontal plane is set as coordinates (x, y), and the height of the object observed by the ceiling LRF 10 is set as coordinates (z), so that the three-dimensional coordinates (x, y, z) is calculated. Hereinafter, data representing three-dimensional coordinates is referred to as “ranging point data”.

  In the following description, the observation range on the floor surface in the building where the ceiling LRF 10 is installed is assumed to be a horizontal plane, and the height (z) of the object observed by the ceiling LRF 10 is from the floor surface in the building. In the following description, it is assumed that the value is a height value. Further, as described above, the coordinates of the observation range of the ceiling LRF 10 and the coordinates of the observation range of the side LRF 20 are calibrated (adjusted) in advance so that the coordinates coincide. Therefore, the value of the two-dimensional coordinate (x, y) included in the moving object candidate data and the value of the coordinate (x, y) representing the horizontal plane included in the distance measuring point data are the movement of the autonomous mobile robot 100. The coordinate value is based on the coordinates in the area.

  In addition, the grid conversion unit 331 in the obstacle position extraction unit 330 converts the movement area of the autonomous mobile robot 100 into a grid map, and registers information representing the object represented by the distance measurement point data on the grid map. . In addition, the moving object position canceling unit 332 in the obstacle position extracting unit 330 determines whether or not the object registered in the grid map is a moving object according to the distance measurement point data. Whether or not the object is a moving object by the moving object position canceling unit 332 is determined based on the moving object candidate data input from the moving object position extracting unit 340 and the previously input moving object candidate data. Is called. If it is determined that the object is a moving object, the registration of the object in the grid map registered by the grid conversion unit 331 is deleted (cancelled). Further, the moving object shadow canceling unit 333 in the obstacle position extracting unit 330 deletes (cancels) information representing the shadow of the moving object. As described above, the obstacle position extraction unit 330 uses the grid information (hereinafter referred to as “movement”) of the movable area of the autonomous mobile robot 100 extracted by the movable area extraction unit 30 from the grid map in which objects other than the moving body are registered. Is output to the movement route generation unit 40 as “possible area information”. The detailed processing contents by each component in the obstacle position extraction unit 330 will be described later.

  The movement path generation unit 40 generates a movement path of the autonomous mobile robot 100 based on the movable area information input from the movable area extraction unit 30.

  Next, the movable area information extracted by the movable area extraction unit 30 of the first embodiment will be described. 3 to 9 are diagrams illustrating specific examples of processing performed in each component in the movable area extracting unit 30 according to the first embodiment.

  First, an object observed by the ceiling LRF 10 and an object registered in the grid map by the grid conversion unit 331 in the obstacle position extraction unit 330 will be described. FIG. 3 is a diagram for explaining the types of grids registered in the movable area extraction unit 30 of the first embodiment. As described above, the ceiling LRF 10 detects an object existing in the observation range by irradiating the conical range with the ceiling LRF 10 as the apex, thereby detecting an object in the observation range. In order to facilitate, a square range as shown in FIG. 3 is set as the observation range of the ceiling LRF 10.

  As shown in FIG. 3, the ceiling LRF 10 observes an object in the moving area of the autonomous mobile robot 100, and the distance information acquisition unit 310 obtains distance data based on the reflected wave information observed by the ceiling LRF 10 as an obstacle position. The data is output to the extraction unit 330. Based on the distance data input from the distance information acquisition unit 310, the grid conversion unit 331 adds the floor area, the obstacle area, and the outside sensing area illustrated in FIG. 3 to the grid map of the movement area of the autonomous mobile robot 100. Register each. 3 represents a region where the floor surface is observed by the ceiling LRF10, an obstacle region represents a region where the floor surface other than the floor surface is observed by the ceiling LRF10, and an outside sensing region represents the ceiling This represents a region where nothing was observed by the LRF 10. This non-sensing region is a region behind a so-called obstacle in which no distance measurement point data is present because the laser beam from the ceiling LRF 10 is not irradiated.

  Here, an example of a registration method of each area by the grid conversion unit 331 will be described. The distance information acquisition unit 310 obtains a plurality of distance measurement point data within the observation range by the laser beam emitted from the ceiling LRF 10 and outputs the distance measurement point data to the grid conversion unit 331. The grid conversion unit 331 first sets the height (z) information included in each ranging point data to a height within a range set in advance with respect to the height of the floor on which the autonomous mobile robot 100 moves. It is determined whether or not the distance measuring point data has the information (z). Then, ranging point data having information on the height (z) within the range set in advance with respect to the height of the floor surface is used as the floor surface data, and is outside the range set in advance with respect to the height of the floor surface. Ranging point data having information on the height (z) of the non-floor surface data.

  Subsequently, the grid conversion unit 331 sets the area of the horizontal plane coordinates (x, y) of the floor data as the floor area (floor area), and the horizontal plane coordinates (x, y) of the non-floor data. The area is defined as an obstacle position (obstacle area). And the information showing that it is a floor surface is mapped to the grid corresponding to the floor area in the grid map which converted the movement area of the autonomous mobile robot 100. Also, information representing an obstacle is mapped to the grid corresponding to the obstacle area.

  Subsequently, the grid conversion unit 331 uses an indeterminate area (non-sensing area) on a grid in which neither the information indicating the floor surface nor the information indicating the obstacle is mapped in the grid map. Map information that represents something. In the following description, in the grid map, a grid to which information representing a floor surface is mapped is a “floor grid”, and a grid to which information representing an obstacle is mapped is a “non-floor grid”. A grid on which information representing an area outside the sensing is mapped is called an “unknown grid”.

  Next, a method for determining a non-movable area by the grid conversion unit 331 in the obstacle position extraction unit 330 will be described. FIG. 4 is a diagram illustrating an example of a method for determining a non-movable area in the movable area extracting unit 30 according to the first embodiment. In FIG. 4A, a “desk” that is a fixed obstacle and a “person” that is a moving object are observed by the ceiling LRF 10, and the “desk” that is observed by the grid conversion unit 331 on the ceiling LRF 10. And “person” are mapped. As shown in FIG. 4A, the positions where “desk” and “person” exist in the grid map are non-floor grids (obstacle regions), respectively, ”Is an unknown grid (outside sensing area).

  First, the grid conversion unit 331 performs binarization processing that represents the presence or absence of an obstacle according to the information of each mapped grid. More specifically, as shown in FIG. 4B, the value of the non-floor grid with an obstacle is “1”, and the value of the floor grid without an obstacle is “0”. In addition, the grid conversion unit 331 sets the value of the unknown grid to “1”, which is the same as that of the non-floor grid, since the unknown grid is also an area with an obstacle. Thereby, an area where there is an obstacle or a possibility that there is an obstacle becomes an area of the obstacle.

  Subsequently, the grid conversion unit 331 checks the values of the respective grids, groups the grids having the same value, and performs a labeling process for assigning a label that is information for identifying each group. More specifically, for example, the values of eight grids positioned around a certain grid are confirmed, and the grids having the same value are set as the same group. Confirmation of the values of grids located in the vicinity is performed for all the grids in the grid map. A label for identifying the group is assigned to each group. In FIG. 4C, the label of the part where “desk” is located in FIG. 4A is “1”, and the label of the part where “person” is located is “2”. Then, the grid conversion unit 331 determines, for each labeled group, the minimum area rectangle that circumscribes the group as a non-movable region.

  Next, a method for calculating moving object candidate data by the moving object position extraction unit 340 will be described. FIG. 5 is a diagram for explaining an example of a method for calculating the moving object position in the movable region extracting unit 30 according to the first embodiment.

  As shown in FIG. 5A, the side LRF 20 observes an object existing in the observation range by irradiating a fan-shaped range on a horizontal plane having the height where the side LRF 20 is installed with a laser beam. . And the distance information acquisition part 320 calculates several distance data based on the information of the observation point of the side LRF20.

  When the difference between the distance data is greater than or equal to a preset value, the moving body position extraction unit 340 determines that the distance data is the distance data of the corner (edge) portion of the object. Then, the moving body position extraction unit 340 performs two-dimensional observation of the two edge portions based on the information on the scan angle of the laser beam included in each distance data and the information on the distance between the side LRF 20 and the object. The coordinates (x, y) of are calculated. In FIG. 5B, when the observation point A and the observation point B are determined as the distance data of the edge portion of the object, two-dimensional coordinates (xA, yA) and coordinates (xB, yB) are calculated. Is shown.

  Then, the moving body position extraction unit 340 calculates two-dimensional coordinates (x, y) representing the center position when the calculated two coordinates are connected by a straight line. FIG. 5B shows a case where the point C is calculated as the center position of the object and the two-dimensional coordinates are coordinates (xC, yC). This coordinate (xC, yC) becomes the moving object candidate data calculated by the moving object position extraction unit 340.

  When the same object is observed by a plurality of side LRFs 20, the calculated center is based on a plurality of two-dimensional coordinates (x, y) representing the center position calculated by the distance data of each side LRF 20. The new two-dimensional coordinates (x, y) representing the position can be used as the moving object candidate data (see point D in FIG. 5C) calculated by the moving object position extraction unit 340.

  Here, before explaining the method of canceling the moving body position, the relationship between the observation times of the objects by the ceiling LRF 10 and the side LRF 20 in the route generation system 1 according to the first embodiment will be explained. Normally, control for synchronizing the timing at which each LRF observes an obstacle is a complicated process. Therefore, even in the route generation system 1 according to the first embodiment, observation of objects of the ceiling LRF 10 and the side LRF 20 is performed. Timing is asynchronous. In addition, the observation time when the side LRF 20 observes the observation range once is different from the observation time when the ceiling LRF 10 observes the observation range once. This is because the difference in performance between the ceiling LRF 10 and the side LRF 20 and the difference in the number of laser beams applied to the observation range (the ceiling LRF 10 observes an object in the observation range in three dimensions, and the side LRF 20 in two dimensions. It depends on factors such as observing objects within the observation range. In the following description, for example, the time when the ceiling LRF 10 observes the observation range once is 250 [ms], and the time when the side LRF 20 observes the observation range once is 100 [ms]. The observation time (cycle) by the LRF 20 will be described as being shorter than the observation time (cycle) by the ceiling LRF 10. Therefore, when the same obstacle is observed on both the ceiling LRF 10 and the side LRF 20, the side LRF 20 observes the same object a plurality of times while the ceiling LRF 10 observes the object once.

  Moreover, in the following description, the case where a person is detected as a moving obstacle (moving body) will be described as an example. Then, assuming that the width of a person is 50 [cm], the ceiling LRF 10 and the side LRF 20 observe objects at intervals of 10 [cm]. In addition, the moving body position extracting unit 340 extracts only the position of the person, that is, the moving body position extracting unit 340 moves the object whose width is equal to the preset width (50 [cm]) of the person. The body candidate data will be described as being output to the obstacle position extracting unit 330 as human candidate data (hereinafter referred to as “person candidate position”).

<First moving body position canceling method>
Next, the moving body position canceling unit 332 in the obstacle position extracting unit 330 and the moving body position canceling method will be described. FIG. 6 is a diagram illustrating a first example of a method for canceling a moving body position in the movable region extraction unit 30 according to the first embodiment. In the first moving body position canceling method shown in FIG. 6, based on the information observed by the side LRF 20 as to whether or not the area determined as the non-movable area by the observation by the ceiling LRF 10 is due to the moving body. The grid of the non-movable area is canceled based on the label of the area determined to be due to the moving body. In FIG. 6A, a “person” that is a moving object is observed by the ceiling LRF 10, and a position where “person” exists in the grid map and a region that is shaded by “person” are moved by the grid conversion unit 331. The case where it is determined as an impossible area is shown. Moreover, based on the information observed by the side LRF 20, the moving body position extraction unit 340 extracts the candidate positions.

  In the first moving body position canceling method, the moving body position canceling unit 332 is based on the current moving body candidate data and the past moving body candidate data input from the moving body position extracting unit 340, so that the unmovable region Is determined to be due to a moving object. For example, as shown in FIG. 6A, the mobile body position canceling unit 332 includes the current human candidate position and the past human candidate position (one past human candidate position and In the case where the position of two past human candidate positions) changes by a predetermined distance or more, it is determined that the immovable region including the current human candidate position is due to the moving object. Then, the information mapped to all grids having the same label as the non-movable area determined to be due to the moving object is set to “0”, which is the same as the floor grid without an obstacle. Thereby, the non-movable area by the observed mobile body is canceled, and it can be set as the movable area where the autonomous mobile robot 100 can move.

  In FIG. 6A, it is determined whether the immovable region is due to a moving object based on the three candidate positions. The number of candidate positions used to determine this immovable region is determined according to the difference in observation time between the ceiling LRF 10 and the side LRF 20. In the first embodiment, since the observation period of the ceiling LRF 10 is 250 [ms] and the observation period of the side LRF 20 is 100 [ms], the observation timing of the ceiling LRF 10 and the side LRF 20 is, for example, FIG. The relationship is as shown in (b). From this, it is considered that there are three observation timings of the side LRF 20 within the observation period of the ceiling LRF 10, and the number of candidate positions used for determining whether the immovable region is due to the moving object is determined. decide.

  In addition, when canceling the immovable area by the first moving body position canceling method, not only the grid map based on the current distance measurement point data observed by the ceiling LRF 10 but also the movement of the past grid map is performed. It is desirable to cancel the non-movable area due to the observed moving object in consideration of information on the non-movable area. For example, if a grid that had a different label in the past grid map has the same label in the current grid map, it is determined that the grid is due to a moving object in the past grid map. Cancel only the non-movable area. Thereby, when the fixed obstacle and the person are approaching, it can prevent canceling even the fixed obstacle which should not be canceled originally.

  In the first moving body position canceling method, the moving body position extracting unit 340 extracts a candidate position based on the information observed by the side LRF 20, and the extracted candidate position is equal to or greater than a predetermined distance. If it has changed, it is determined that the immovable region including the candidate position is due to the moving object. The method of determining whether or not the immovable region is due to a moving object is not limited to the determination performed based on the information observed by the side LRF 20, and the method of determining the immovable region observed by the ceiling LRF 10 is not limited. It is also possible to determine whether the immovable region is due to a moving object using the information. For example, the information on the non-movable area of the current grid map determined as the non-movable area by the observation with the ceiling LRF 10 is compared with the information on the non-movable area of the past grid map, and it is determined that the obstacle is the same. When the non-movable area to be moved is changed by a predetermined grid (distance) or more, it can be determined that the non-movable area is due to the moving body.

<Second moving body position canceling method>
Next, another method for canceling the moving object position by the moving object position canceling unit 332 in the obstacle position extracting unit 330 will be described. FIG. 7 is a diagram for explaining a second example of the method for canceling the moving object position in the movable region extracting unit 30 according to the first embodiment. In the second moving body position canceling method shown in FIG. 7, the movement is determined to be due to the moving body based on the information observed by the side LRF 20 by being determined as the non-movable area by observation with the ceiling LRF 10. Among the impossible areas, the grid of the non-movable area having the same label within the range of the movable body assumed in advance is canceled.

  In the second moving body position canceling method, for example, as shown in FIG. 7A, a “desk” that is a fixed obstacle and a “person” that is a moving body are observed by the ceiling LRF 10. This is an effective method when the observed “desk” and “person” are close to each other. That is, when the observed “desk” and “person” are close to each other, as shown in FIG. 7B, the grid conversion unit 331 causes the “table” and “person” to move in the same group. It will be determined as a possible area. In this state, if all grids having the same label are simply canceled by the above-described first moving body position canceling method, up to the “desk” which is a fixed obstacle that should not be canceled originally. Will be canceled.

  Therefore, in the second moving body position canceling method, the moving body position canceling unit 332 assumes the size of the moving body assumed in advance, and in the first embodiment, the width of the “human” that is the assumed moving body. Based on (50 [cm]), the size (hereinafter referred to as “moving body cancellation range”) for canceling the grid in the immovable region shown in FIG. 7C is determined. In FIG. 7C, a radius half of the person's width (distance) is set as the radius r, and the range of the circle with the set radius r is centered on the person candidate position input from the moving body position extracting unit 340. The moving object cancellation range. And the example which cancels the grid in the mobile body cancellation range is shown. More specifically, the moving object position canceling unit 332 includes two-dimensional coordinates p1 (x1, y1) that are the center positions of the objects included in the moving object candidate data of the human candidate positions input from the moving object position extracting unit 340. ). Then, the grid in the immovable region where the relationship of the following expression (1) holds (in FIG. 7C, the grid of coordinates pn (xn, yn)) is determined to be a grid representing a moving body.

  In the second moving body position canceling method, the moving body position canceling unit 332 first determines the current human candidate position input from the moving body position extracting unit 340, as in the first moving body position canceling method described above. It is determined that a non-movable region in which past human candidate positions (one past human candidate position and two past human candidate positions) have changed by a predetermined distance or more is due to a moving object. Subsequently, the moving object position canceling unit 332 maps to all the grids within the moving object cancellation range determined to be a grid representing the moving object in the non-movable area determined to be due to the moving object. The information is set to “0”, which is the same as the floor grid without an obstacle (see FIG. 7D). As a result, as shown in FIG. 7E, the grid area representing the “person” that is the moving body is canceled and can be made a movable area in which the autonomous mobile robot 100 can move. The area of the grid representing the “desk”, which is a fixed obstacle, remains a non-movable area without being canceled.

  The moving body cancellation range determined by the moving body position canceling unit 332 based on the size of the moving body assumed in advance is the position detection accuracy when the side LRF 20 observes the object, and the ceiling LRF 10 observes the object. The radius r to be set can be changed according to the position detection accuracy at the time, the average shoulder width of the person in consideration of the thickness of the person's clothes. In addition, the radius r to be set can be dynamically changed according to the moving speed of the moving body. For example, when the moving speed of the moving body is high, the moving body cancellation range is widened by increasing the set radius r.

  In addition, the moving body canceling range is not a range of a circle with a radius r, but the shape of the moving body canceling range can be changed depending on the direction in which the moving body is moving. For example, when the moving body is moving upward in FIG. 7 (c), the range of the ellipse in which the distance in the upward direction in which the moving body is moving and the opposite downward direction is shortened is moved. It can also be a body cancellation range. In this way, by changing the moving body cancellation range according to the moving body, it is possible to prevent even a range that should not be canceled from being canceled.

<Movement shade cancellation method>
Next, a method for canceling the shadow of the moving object by the moving object shadow canceling unit 333 in the obstacle position extracting unit 330 will be described. FIG. 8 is a diagram illustrating an example of a moving object detected in the route generation system 1 according to the first embodiment. In the route generation system 1, an obstacle in a moving area where the autonomous mobile robot 100 moves is detected by a plurality of ceiling LRFs 10 and a plurality of side LRFs 20 (not shown). For example, as shown in FIG. 8A, the ceiling LRFs 10 are installed such that the observation ranges of the ceiling LRFs 10 overlap each other. Thereby, the area | region which is shadowed by the obstacle etc. which exists in the movement area | region of the autonomous mobile robot 100 is decreased. However, as shown in FIG. 8A, for example, when an object (“person” in FIG. 8A) is present toward the end in the moving area, “ Even the ceiling LRF 10 closest to “people” cannot eliminate all the shadow areas.

  That is, the area behind the moving object may remain even after the non-movable area by the moving object is canceled by the second moving object position canceling method described above. For example, as shown in FIG. 8B-1, consider a case where a “person” that is a moving object and the “shadow of a person” are observed by the ceiling LRF 10. As described above, the “human shadow” is an area that is an unknown grid outside the sensing area. Then, as illustrated in FIG. 8B-2, the grid conversion unit 331 determines that “person” and “person's shadow” are immovable areas of the same group. In this state, the grid area representing the “person” that is the moving body is canceled by the second moving body position canceling method described above. However, if the “human shadow” area is larger than the moving object cancellation range, as shown in FIG. 8B-3, the “human shadow” area outside the moving object cancellation range is It remains as a non-movable area.

  Therefore, in this moving body shadow canceling method, the shadow area of the moving body remaining after the non-movable area by the moving body is canceled by the second moving body position canceling method is cancelled. FIG. 9 is a diagram illustrating an example of a method for canceling the shadow area of the moving object in the movable area extracting unit 30 according to the first embodiment. In the moving object shadow canceling method, the height of the moving object is calculated according to the distance data calculated based on the information of the reflected wave input from the ceiling LRF 10, and based on the calculated height of the moving object. The length of the shadow of the moving object is calculated. Then, the center position of the mobile object cancellation range, that is, the position of the candidate person input from the mobile object position extraction unit 340, which is determined to have a mobile object (person) in the second mobile object position canceling method. While changing (hereinafter referred to as “person position”), the non-movable area is additionally canceled. As a result, the area behind the moving body (person) remaining as the immovable area can be canceled to be a movable area where the autonomous mobile robot 100 can move.

In the moving body shadow canceling method, the moving body shadow canceling unit 333 cancels the shadow area of the moving body (person) remaining as the unmovable area in the following procedure (see FIG. 9).
(Procedure 1): The moving object shadow canceling unit 333 has the coordinates of the installation position of the ceiling LRF 10 closest to the person position (the coordinates (x2, y2) shown in FIG. 9B) (shown in FIG. 9B). The coordinates (x1, y1)) are searched.
(Procedure 2): The moving object shadow canceling unit 333 calculates the distance Lh between the human position and the position of the ceiling LRF10. The distance Lh can be calculated using, for example, the following formula (2).

(Procedure 3): The moving object shadow canceling unit 333 calculates the length Ls of the shadow part based on the height of the “person” existing at the person position. Note that the length Ls of the shadow can be calculated using, for example, the following expression (3).

  In the above equation (3), Hs indicates the height at which the ceiling LRF 10 closest to the person position is installed, and Hh indicates information on the height (z) included in the distance measurement point data at the person position. The height Hh is information on the height (z) included in the distance measurement point data of a grid (for example, a grid within the range of the human width (50 [cm])) located around the grid of the human position. It can also be set to the highest value (peak value).

(Procedure 4): The moving body shadow canceling unit 333 calculates a unit vector E (ex, ey) between the ceiling LRF 10 closest to the person position and the person position. The unit vector E (ex, ey) can be calculated using, for example, the following equation (4).

(Procedure 5): The moving object shadow canceling unit 333 determines the coordinates of the person position in the direction of the unit vector E (ex, ey) from the current person position (initially coordinates (x2, y2)). The coordinates of the moved position are set as a new person position (for example, coordinates ps (xs, ys) shown in FIG. 9B). The movement of the coordinates ps (xs, ys) of the new person position can be calculated using, for example, the following equation (5).

  Then, the information mapped to the unknown grid (in FIG. 9C, the grid of coordinates pn (xn, yn)) existing within the moving object cancellation range centered on the new person position is free of obstacles. Set to “0”, the same as the floor grid. As a result, the autonomous mobile robot 100 can move around the circle of the set radius r around the new person position, that is, the grid in the non-movable area where the relationship of the following equation (6) holds. An area can be used (see FIG. 9C).

  Thereafter, the above-described procedure 5 is repeated, and the unknown grid existing up to the shadow length Ls is sequentially changed to the floor grid according to the coordinates ps (xs, ys) of the new person position. In the repetition of the procedure 5, for example, first, the distance Lh between the position of the person and the position of the ceiling LRF 10 before performing the moving body shadow cancellation is set as the distance L1. When the distance L is, for example, 100 [mm], the distance L1 is updated with the distance L2 represented by the following equation (8) while sequentially updating the distance L1 as the following equation (7). Repeat step 5 until

  In the route generation system 1 according to the first embodiment, as described above, the installation position of the ceiling LRF 10 is calibrated (adjusted) in advance within the movement area of the autonomous mobile robot 100, and thus the moving body. In the shadow cancellation method, for example, values such as the height Hs at which the ceiling LRF 10 closest to the person position is installed and the coordinates (x1, y1) of the installation position are all coordinates in the movement area of the autonomous mobile robot 100. This is the standard value.

  In the first moving object position canceling method described above, all moving objects having the same label determined to be due to the moving object are also taken into consideration in consideration of information on the non-movable area of the past grid map. When the possible area is canceled, it is also possible not to cancel the shadow portion of the moving body by the moving body shadow canceling method.

<Move route generation process>
Next, a method for generating a movement route of the autonomous mobile robot 100 in the route generation system 1 according to the first embodiment will be described. FIG. 10 is a flowchart showing a processing procedure of the movement route generation process in the route generation system 1 of the first embodiment. Note that the ceiling LRF 10 is installed at a predetermined height in advance when the route generation system 1 is operated, and the installation height and the installation posture of the ceiling LRF 10 are described as being calibrated (adjusted) in advance. Further, in the moving route generation processing procedure shown in FIG. 10, as shown in FIG. 8, the grid conversion unit 331 merges the ranging point data input from the plurality of ceiling LRFs 10, and the autonomous mobile robot 100. A processing procedure in the case where all of the moving areas are converted into a grid map will be described.

  When the route generation system 1 operates, first, in step S100, the distance information acquisition unit 310 acquires information obtained by the ceiling LRF 10 irradiating the laser beam to scan the obstacle in the observation range, and the obstacle distance data is obtained. Output to the obstacle position extraction unit 330. Further, the distance information acquisition unit 320 acquires information obtained by the side LRF 20 irradiating the laser beam to scan the obstacle within the observation range, and outputs the obstacle distance data to the moving body position extraction unit 340.

  Subsequently, in step S200, the obstacle position extraction unit 330 calculates distance measurement point data of the obstacle based on the distance data input from the distance information acquisition unit 310. Then, the grid conversion unit 331 in the obstacle position extraction unit 330 registers information on the position of the object (obstacle) represented by the distance measurement point data in the grid map of the movement area of the autonomous mobile robot 100.

  Subsequently, in step S <b> 300, the moving object position extraction unit 340 calculates moving object candidate data based on the distance data input from the distance information acquisition unit 320, and outputs it to the obstacle position extraction unit 330. In step S400, the moving object position canceling unit 332 in the obstacle position extracting unit 330 determines a moving moving object based on the moving object candidate data input from the moving object position extracting unit 340.

  Subsequently, in step S500, the moving body position canceling unit 332 cancels the grid of the non-movable area by the determined moving body. Then, in step S600, the moving object shadow canceling unit 333 in the obstacle position extracting unit 330 cancels the grid of the determined non-movable area in the shadow of the moving object.

  Subsequently, in step S700, the obstacle position extraction unit 330 determines whether or not cancellation of the immovable area by all the moving bodies has been completed. When the cancellation of the immovable area by all the moving objects is completed, the process proceeds to step S800. If cancellation of the immovable area by all the moving objects has not been completed, the process returns to step S300 to calculate and move the moving object candidate data based on the distance data input from the distance information acquisition unit 320. The determination of the moving body and the cancellation of the non-movable area by the moving body are repeated.

  Subsequently, in step S800, the obstacle position extraction unit 330 generates movable region information (grid map) in which the grid of the non-movable region by the moving body is canceled, and the generated movable region information is used as the movement route generation unit. Output to 40.

  Subsequently, in step S900, the movement path generation unit 40 generates a movement path of the autonomous mobile robot 100 based on the movable area information input from the obstacle position extraction unit 330. In addition, as a method for generating the movement route of the autonomous mobile robot 100 in the route generation system 1 of the first embodiment, a conventionally used movement route generation method such as RRT-connect is used. . Further, for example, a route connecting the current position of the autonomous mobile robot 100 and the destination of the movement destination with a straight line is set as a movement route. Here, when there are obstacles on the moving route, calculate a safe point where the obstacle can be avoided before the obstacle, and from the calculated point, the obstacle can be moved in a wider direction. Generate a route that avoids Then, a route that connects the point where the obstacle avoidance has been completed and the destination again with a straight line is defined as a new travel route. Furthermore, when an obstacle exists on the new movement route, a route that can avoid the obstacle is generated as described above. Such a movement route generation is repeated to generate a final movement route.

  As described above, according to the route generation system 1 of the first embodiment, the movable region extraction unit 30 is based on the information of the object at the position of the predetermined height observed by the side LRF 20. However, it is possible to detect an obstacle (moving body) moving within the movement range of the autonomous mobile robot 100. Then, the information on the non-movable area by the moving body can be canceled. Thereby, when the movement path | route for the autonomous mobile robot 100 to move is produced | generated, the movement path | route which can move efficiently to the destination can be produced | generated.

  There may be a moving body such as a person around the generated movement route. In such a case, it is desirable to generate the movable area information for generating the movement route by predicting the moving speed and moving direction of the moving body. For example, the moving speed of the moving body is calculated based on the current moving body candidate data and the past moving body candidate data input from the moving body position extracting unit 340, and the calculated moving speed v is determined in advance. If it is faster than the movement speed va, it is determined that the autonomous mobile robot 100 leaves the movement path generated while moving to the position of the moving body, and the information on the non-movable area by the moving body is canceled. Further, when the calculated moving speed v is slower than the predetermined moving speed va, it is determined that the autonomous mobile robot 100 stays (fixes) around the moving path generated while moving to the position of the moving body. The information on the non-movable area is not canceled by the moving body.

  FIG. 11 is a diagram illustrating a canceling process for a moving object in the route generation system 1 according to the first embodiment. In FIG. 11, a “person” that is a moving object exists around the movement route of the generated autonomous mobile robot 100. The moving speed of each “person” is moving speeds v1 to v4. The movable area extraction unit 30 compares the moving speeds v1 to v4 of the respective moving bodies with a predetermined moving speed va, and determines whether or not to cancel the information on the non-movable areas by the moving bodies. In FIG. 11, the moving speed v3 (v3> Va) is not canceled without canceling the non-movable area by the moving body having the moving speed v1 (v1 <Va) and the moving body having the moving speed v2 (v2 <Va). This represents a case where a non-movable area is canceled by a moving body and a moving body having a moving speed v4 (v4> Va).

  The predetermined moving speed va can be determined based on the speed of the autonomous mobile robot 100 until the mobile body 100 moves to the position of the mobile body and the distance between the autonomous mobile robot 100 and the mobile body. In addition, when it is determined that the moving body is staying (fixed) around the movement path of the autonomous mobile robot 100, the time during which the moving body stays is measured, and the mobile body cannot move based on the measured staying time. It may be possible to determine whether or not to cancel the area information.

<Second Embodiment>
Next, another embodiment of the present invention will be described. FIG. 12 is a block diagram showing a schematic configuration of a route generation system according to the second exemplary embodiment of the present invention. In FIG. 12, the route generation system 2 includes a ceiling LRF 10, a movable region extraction unit 31, and a movement route generation unit 40. The movable region extraction unit 31 includes a distance information acquisition unit 310 corresponding to the ceiling LRF10, a light amount information acquisition unit 370 corresponding to the ceiling LRF10, an obstacle position extraction unit 330, and a moving body position extraction unit 380. Yes. The obstacle position extraction unit 330 includes a grid conversion unit 331, a moving body position cancellation unit 332, and a moving body shadow cancellation unit 333.

  In the route generation system 2 shown in FIG. 12, the side LRF 20 provided in the route generation system 1 shown in FIG. 1 is deleted. Further, the movable area extraction unit 30 provided in the route generation system 1 shown in FIG. 2 has been changed to the movable area extraction unit 31, and more specifically, the distance information acquisition unit 320 and the moving body position extraction unit. 340 is only changed to a light amount information acquisition unit 370 and a moving body position extraction unit 380. Therefore, in the following description, the same components as those of the route generation system 1 shown in FIG.

  The route generation system 2 includes an obstacle in a moving area where the autonomous mobile robot moves by the ceiling LRF 10, a moving body (for example, a moving body such as a person or another autonomous mobile robot, and an autonomous mobile robot) 100 itself). Then, the route generation system 2 sets, as the movable region of the autonomous mobile robot 100, the region obtained by removing the region of the detected moving body (moving obstacle) from the region of all obstacles detected by the ceiling LRF 10. Based on the extracted information on the movable area, the movement path of the autonomous mobile robot 100 is generated.

  The movable area extraction unit 31 extracts an obstacle in the movement area of the autonomous mobile robot 100 based on the information of the reflected wave input from the ceiling LRF 10. Then, the position information of the obstacle excluding the obstacle determined to be a moving body from all the extracted obstacles, that is, the position information of the obstacle fixed in the movement area of the autonomous mobile robot 100, The data is output to the movement route generation unit 40.

  The light amount information acquisition unit 370 calculates the light amount of the reflected wave by the object that reflects the laser beam based on the reflected wave information input from the ceiling LRF 10. Then, the calculated light amount information of the reflected wave, the information on the distance between the ceiling LRF 10 and the object, and the information on the scan angle of the laser beam irradiated by the ceiling LRF 10 are output as light amount data to the moving body position extracting unit 380. .

  The moving body position extraction unit 380 is configured such that the shape of the object observed by the ceiling LRF 10 is based on the information on the reflected light amount and the laser beam scan angle information included in the light amount data input from the light amount information acquisition unit 370. It is determined whether or not the shape of a moving body assumed in advance moving in the movement area of the autonomous mobile robot 100 is determined.

  Here, an example of a method for determining a moving object when the assumed moving object is “person” will be described. FIG. 13 is a diagram illustrating an example of a moving object determination method in the moving object position extraction unit 380 according to the second embodiment. For example, the shape of a moving body (person) as shown in FIG. 13A can be represented by a grayscale image based on the amount of reflected waves input from the ceiling LRF 10. The moving body position extraction unit 380 determines whether or not the light amount data (grayscale image) input from the light amount information acquisition unit 370 is light amount data of a moving body (person) assumed in advance. In addition, as a method for determining whether or not a mobile object (person) is assumed in advance in the route generation system 2 of the second embodiment, a conventionally used pattern matching or template matching method is used. . The template at this time may be created according to the horizontal plane coordinates (x, y) and height of the position of the moving body (person) serving as the template, and the angle of the laser beam scan. Alternatively, instead of the light amount data input from the light amount information acquisition unit 370, the moving object (person) may be detected using distance data output from the distance information acquisition unit 310. In that case, for example, as shown in FIG. 13B, the shape from the vicinity of the shoulder of the “person” to the tip of the head is recognized. The template in this case is created not by the light amount data but by the distance data.

  The moving body position extraction unit 380 indicates the center position of the moving body when it is determined that the light amount data input from the light amount information acquisition unit 370 is light amount data of a moving body (person) assumed in advance. Dimensional coordinates (x, y) are calculated, and moving object candidate data representing the coordinates of the calculated center position of the moving object is output to the obstacle position extracting unit 330. For example, in FIG. 13B, data representing the coordinates of the position of the tip of the head is output to the obstacle position extraction unit 330 as moving object candidate data.

  The obstacle position extraction unit 330 calculates distance measurement point data based on the distance data input from the distance information acquisition unit 310, and the moving body position cancellation unit 332 in the obstacle position extraction unit 330 Based on the moving object candidate data input from the extraction unit 380 and the previously input moving object candidate data, it is determined whether or not the object registered in the grid map is a moving object. When it is determined that the object is a moving object, the obstacle position extraction unit 330 deletes (cancels) registration of the object in the grid map registered by the grid conversion unit 331 and cancels the moving object shadow. The movable area information in which the information representing the shadow of the moving object is deleted (cancelled) by the unit 333 is output to the movement route generation unit 40 as the movable area information extracted by the movable area extraction unit 31.

  Then, the movement path generation unit 40 generates a movement path of the autonomous mobile robot 100 based on the movable area information input from the movable area extraction unit 31.

  As described above, according to the route generation system 2 of the second embodiment, the movable region extraction unit 31 is configured to be an autonomous mobile robot based on the information on the assumed moving body observed by the ceiling LRF 10. Obstacles (moving bodies) that are moving within the 100 moving range can be detected. Then, the information on the non-movable area by the moving body can be canceled. Thereby, when the movement path | route for the autonomous mobile robot 100 to move is produced | generated, the movement path | route which can move efficiently to the destination can be produced | generated.

  As described above, according to the mode for carrying out the present invention, by installing the distance measuring sensor in the movement area of the autonomous mobile robot 100, the autonomous mobile robot 100 such as a person or another autonomous mobile robot can be used. An obstacle that becomes an obstacle in the movement of the autonomous mobile robot 100 can be detected without extracting a moving body moving on the movement route as an obstacle. Thereby, the movement path | route which the autonomous mobile robot 100 can move to a destination efficiently can be produced | generated.

  In addition, it is possible to generate a movement route only from the movable area information extracted by the movable area extraction unit 30 or the movable area extraction unit 31 of the present invention. Accordingly, the autonomous mobile robot 100 can safely move without attaching an obstacle detection device to the autonomous mobile robot 100 itself.

  In this embodiment, the case where “person” is detected as a moving obstacle (moving body) has been described as an example. However, the moving body is limited to a mode for carrying out the present invention. Instead, various mobile objects can be detected within a range not departing from the gist of the present invention.

  Further, in the present embodiment, the case where the side LRF 20 is installed at a predetermined height position has been described. However, the height at which the side LRF 20 is installed is limited to the mode for carrying out the present invention. Instead, by installing a plurality of side LRFs 20 at the same location, it is possible to detect moving objects of various heights by detecting obstacles having a plurality of heights. Moreover, in this embodiment, although the case where "person" is the mobile body assumed beforehand is demonstrated, the mobile body assumed beforehand is not limited to the form for implementing this invention, Various moving bodies can be detected by setting a plurality of shapes of the moving bodies assumed in advance corresponding to the various moving bodies.

  In the present embodiment, the autonomous mobile robot 100 itself is also recognized as an obstacle, and the obstacle position extraction unit 330 makes a determination in the same manner as a moving body other than the autonomous mobile robot 100, whereby the autonomous mobile robot 100 itself Unmovable areas are also deleted (cancelled). The method of deleting (cancelling) the non-movable area by the autonomous mobile robot 100 itself is not limited to the mode for carrying out the present invention, but by directly determining the autonomous mobile robot 100 itself, A method of easily deleting (cancelling) a non-movable area by the autonomous mobile robot 100 itself is also conceivable. For example, by notifying the route generation system of self-position (odometry) information of the autonomous mobile robot 100 by various sensors such as a gyro sensor provided in the autonomous mobile robot 100 itself, GPS (Global Positioning System) information, etc. The route generation system can calculate the position information of the autonomous mobile robot 100 itself, and the non-movable area can be deleted (cancelled) based on the position information of the autonomous mobile robot 100 itself calculated by the route generation system. In addition, for example, by attaching a reflection sheet or the like that reflects the laser beam with high efficiency to the tip of the autonomous mobile robot 100, the autonomous mobile robot 100 itself and other moving objects can be determined from the amount of reflected waves of the laser beam. And the non-movable area by the autonomous mobile robot 100 itself identified by the route generation system can be deleted (cancelled).

  In the present embodiment, the movable region information extracted by the movable region extraction unit 30 or the movable region extraction unit 31 is output to the movement route generation unit 40 every time the ceiling LRF 10 observes the observation range. However, it is also possible to output information in which the movable area information is changed, that is, difference information of the movable area information. In this case, for example, the movable area extraction unit 30 holds the movable area information output to the movement path generation unit 40 last time, and holds the movable area information and the movement to be output to the current movement path generation unit 40. Only the difference from the possible area information is output to the movement route generation unit 40. Thereby, for example, when the movable region extraction unit 30 and the movement route generation unit 40 are transferring the movable region information by communication, the transfer between the movable region extraction unit 30 and the movement route generation unit 40 is performed. Communication efficiency can be improved.

  In the present embodiment, the case where the distance measuring sensor is a laser range finder has been described. However, a similar movable region can be extracted using other distance measuring sensors. For example, the movable region information can be output to the movement path generation unit 40 by detecting an object from image data such as a stereo camera.

  The processing by each component of the movable area extraction unit 30 shown in FIG. 2, each processing step of the route generation system 1 shown in FIG. 10, or each configuration of the movable area extraction unit 31 shown in FIG. A program for realizing processing by elements is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read into the computer system and executed to extract a movable area in the route generation system 1 The above-described various processes relating to the movable region extraction unit 31 in the unit 30 or the route generation system 2 may be performed. Here, the “computer system” may include an OS and hardware such as peripheral devices. Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used. The “computer-readable recording medium” means a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, etc. This is a storage device.

  Further, the “computer-readable recording medium” refers to a volatile memory (for example, DRAM (Dynamic) in a computer system serving as a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. Random Access Memory)) that holds a program for a certain period of time is also included. The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, what is called a difference file (difference program) may be sufficient.

  The embodiment of the present invention has been described above with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes various modifications within the scope of the present invention. It is.

1, 2 ... Route generation system 10 ... Ceiling LRF (first distance sensor, distance sensor)
20 ... Side LRF (second distance sensor)
30... Movable area extraction unit (movable area extraction device)
310 ... Distance information acquisition unit (first obstacle detection means)
320 ... Distance information acquisition unit (second obstacle detection means)
330: Obstacle position extraction unit (first obstacle detection means, obstacle area extraction means)
331 ... Grid converter (first obstacle detection means)
332... Moving body position canceling unit (obstacle area extracting means)
333 ... Moving object shadow canceling unit (obstacle region extracting means)
340 ... Moving object position extraction unit (second obstacle detection means)
31... Movable area extraction unit (movable area extraction device)
370 ... Light quantity information acquisition unit (shape detection means)
380 ... Moving body position extraction unit (shape detection means)
40... Travel route generation unit (route generation means)
100: Autonomous mobile robot

Claims (11)

A first distance sensor installed so as to observe a lower portion of the space from a predetermined first height in the space in which the autonomously moving mobile body moves; and a predetermined in the space An obstacle present in the space in which the movable body moves is detected by a second distance sensor installed so as to observe the side of the space from a second height lower than the first height. In the movable area extracting device that extracts information on the movable area where the moving body can move based on the detected area information of the obstacle,
First obstacle detection means for detecting a first obstacle region representing the obstacle region based on information on a distance observed by the first distance sensor;
Second obstacle detection means for detecting a second obstacle region representing the obstacle region based on information on a distance observed by the second distance sensor;
The obstacle that is continuously moving based on the second obstacle area is detected as a moving object, and the first obstacle that represents the area of the detected moving object is detected from the first obstacle area. Obstacle area extraction means for extracting a movable area where the moving body can move by determining an area excluding the obstacle area as an immovable area ;
Comprising
A movable region extraction apparatus characterized by the above. The second obstacle detection means includes
Based on the information on the distance when the second distance sensor observes the obstacle, the obstacle position representing the position of the obstacle is extracted,
The obstacle region extracting means includes
The moving object is detected based on the obstacle position extracted by the second obstacle detecting means and the past obstacle position previously extracted by the second obstacle detecting means, and the detected movement is detected. An area excluding the first obstacle area including the obstacle position representing the position of an object is defined as the movable area.
The movable area extracting device according to claim 1, wherein The second obstacle detection means includes
Based on the information on the distance when the second distance sensor observes the obstacle, the obstacle position representing the position of the obstacle is extracted,
The obstacle region extracting means includes
The moving object is detected based on the obstacle position extracted by the second obstacle detecting means and the past obstacle position previously extracted by the second obstacle detecting means, and the detected movement is detected. An area excluding the first obstacle area in a predetermined range including the obstacle position representing the position of an object is set as the movable area.
The movable area extracting device according to claim 1, wherein The obstacle region extracting means includes
Search for the first distance sensor installed at a position closest to the obstacle position representing the position of the moving object;
The distance between the obstacle position of the moving object and the searched first distance sensor, and the moving object included in the distance information when the searched first distance sensor observes the moving object. On the basis of the height information, the length of the shadow area that the first distance sensor cannot observe by the moving object is calculated,
The predetermined distance between the obstacle position of the moving object and a position separated by the length of the calculated shadow area is determined as the obstacle position of the moving object and the first distance sensor searched. of moving at an extension of connecting the installation position, the excluding predetermined range from the first obstacle area region and the movable region,
The movable area extracting device according to claim 3, wherein The obstacle region extracting means includes
When the obstacle position extracted by the second obstacle detection means has changed by more than a predetermined threshold from the previous obstacle position previously extracted by the second obstacle detection means, Determining that the obstacle corresponding to the obstacle position is the moving object;
The movable area extracting device according to any one of claims 2 to 4, wherein the movable area extracting device is provided. When the space in which the moving body moves is in the building,
The first distance sensor is
A plurality are installed at predetermined intervals on the ceiling in the building,
The first obstacle detection means includes
Detecting the first obstacle region based on distance information in a space obtained by merging observation ranges of the first distance sensors installed in a plurality;
The movable area extracting device according to any one of claims 1 to 5, wherein The second distance sensor is
At a predetermined height between the ceiling and the floor, a plurality of sets are installed at predetermined intervals so as to observe an observation range parallel to the floor in the building,
The second obstacle detection means includes
Detecting the second obstacle region based on information on a distance in a plane obtained by merging observation ranges of a plurality of the second distance sensors installed,
The movable area extracting apparatus according to any one of claims 1 to 6, wherein the movable area extracting apparatus is provided. The obstacle region extracting means includes
A route generating means for generating a route along which the moving body moves based on the extracted movable region;
Further comprising
The movable area extracting device according to any one of claims 1 to 7, wherein the movable area extracting device is provided. An obstacle present in the space in which the moving body moves by a distance sensor installed so as to observe the lower part of the space from a predetermined height in the space in which the moving body moves autonomously In the movable area extracting device that detects an object and extracts information on the movable area where the moving body can move based on the detected area information of the obstacle,
Obstacle detection means for detecting an obstacle area representing the area of the obstacle based on information on the distance observed by the distance sensor;
Shape detection means for detecting the shape of the obstacle based on information on the amount of reflected wave light when the distance sensor observes the obstacle;
Comparing the shape of the obstacle detected by the shape detection means with the shape of the predetermined object, and when the shape of the obstacle is equivalent to the shape of the predetermined object, the distance sensor Based on the distance information when the obstacle is observed, the continuous movement of the obstacle is detected, and the detected area of the obstacle that is continuously moving is represented from the obstacle area. The movable body can move by determining the area that is the obstacle area , excluding the obstacle area set according to the speed of the obstacle, as the non-movable area. Obstacle area extracting means for extracting an area;
Comprising
A movable region extraction apparatus characterized by the above. A first distance sensor installed so as to observe a lower portion of the space from a predetermined first height in the space in which the autonomously moving mobile body moves; and a predetermined in the space An obstacle present in the space in which the movable body moves is detected by a second distance sensor installed so as to observe the side of the space from a second height lower than the first height. In the movable area extracting method for extracting information on the movable area where the moving body can move based on the detected area information of the obstacle,
A first obstacle detection procedure for detecting a first obstacle area representing the obstacle area based on information on a distance observed by the first distance sensor;
A second obstacle detection procedure for detecting a second obstacle region representing the obstacle region based on information on a distance observed by the second distance sensor;
The obstacle that is continuously moving based on the second obstacle area is detected as a moving object, and the first obstacle that represents the area of the detected moving object is detected from the first obstacle area. An obstacle area extraction procedure for extracting a movable area in which the moving body can move by determining an area excluding the obstacle area as an immovable area ;
including,
A movable region extraction method characterized by the above. An obstacle present in the space in which the moving body moves by a distance sensor installed so as to observe the lower part of the space from a predetermined height in the space in which the moving body moves autonomously In the movable region extraction method for detecting an object and extracting information on a movable region in which the moving body can move based on information on the detected obstacle region,
An obstacle detection procedure for detecting an obstacle region representing the obstacle region based on information on a distance observed by the distance sensor;
A shape detection procedure for detecting the shape of the obstacle based on information on the amount of reflected wave light when the distance sensor observes the obstacle;
Comparing the shape of the obstacle detected by the shape detection procedure with the shape of the predetermined object, and when the shape of the obstacle is equivalent to the shape of the predetermined object, the distance sensor Based on the distance information when the obstacle is observed, the continuous movement of the obstacle is detected, and the detected area of the obstacle that is continuously moving is represented from the obstacle area. The movable body can move by determining the area that is the obstacle area , excluding the obstacle area set according to the speed of the obstacle, as the non-movable area. Obstacle area extraction procedure for extracting the area;
including,
A movable region extraction method characterized by the above.